<%BANNER%>

Human Use of Two Species of River Turtles (Podocnemis spp.) in Lowland Eastern Bolivia


PAGE 1

HUMAN USE OF TWO SPECIES OF RIVER TURTLES ( Podocnemis spp.) IN LOWLAND EASTERN BOLIVIA By KRISTEN MARIE CONWAY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Kristen Marie Conway

PAGE 3

In my life there have been some women w ho have modeled success for me in special ways. I wish to dedicate this work to them, beginning with my grandmothersRuth Colletta Michals and Irene Francis Doolwho demonstrated strength and the value of perseverance. Several new friends in Piso Firme (particularly Doa Lucida Paz Gmez, Doa Nilda Gmez Pea, Pura Gmez Pea, Pura Montera Pea, and Mariluz Cesar Gmez) opened their lives and homes to me unselfishly as I faced new challenges, each sharing her individual strengths and optim ism, and demonstrating true virtue.

PAGE 4

ACKNOWLEDGMENTS For financial support of my research I am thankful to the National Science Foundation (Grant # BCS-0082250) and The Chelonian Research Foundation. Funding for preliminary research was provided by the InterAmerican Foundation (Grant # F-019) and the Tropical Conservation and Development Program of the University of Florida. For his patience in teaching me and sharing with me his enthusiasm for reptiles as well as untiring editing assistance, I am greatly indebted to Dr. J. Parren Ross. For his guidance, strong mentoring and exigencia I am deeply indebted to Dr. Charles Wood. For her encouragement to reach new levels of inquiry and understanding of scientific investigation I am grateful to Dr. Marilyn Swisher who has taught me valuable lessons through two graduate degrees. Every young scholar who has the benefit of mentors such as I found in Dr. Ross, Dr. Wood and Dr. Swisher is truly gifted. Dr. Michael Binford and Dr. Nigel Smith offered thoughtful insights along the way. Two past committee membersDr. Richard Bodmer and Dr. Edward Maleckiwere extremely helpful in the intellectual development of my course of study and the design stage of my research. I benefited greatly from their involvement. Dr. Ken Portiers guidance on categorical data analysis was absolutely critical. I am deeply grateful to the community members of Piso Firme and Remanso, Bolivia, who allowed me to enter into their lives and ask some strange questions, all for the sake of my research. Their time, help, and input are appreciated. I look forward to continuing relationships with them in the years to come. For his help with transect setup and data logging I am indebted to wildlife biologist Stan iv

PAGE 5

Howarter. I am grateful to my family, Joyce Marie Michals Conway, my motherfrom the beginning the ceaseless supporter of my education who taught me that happiness comes in many forms; my father, Timothy John Conwaywho has always encouraged and supported my independence; and my sister, Julie Anne Beckleswho regularly teaches me the importance of taking time to enjoy life, because the little details do take care of themselves. Many dear friends have helped me with diverse tasks related to this project as well as life in general. Without friends as good as these I might have lost my way. For their generous help and support I thank Alisa Coffin, Amanda Stronza-Rojas, Andreana Ososki, Ann Angelheart, Beatriz Alvarado Heath, Carol Lee Hilewick, Claudia Stickler, Elli Sugita, Jocelyn Peskin Coln, Naomi Trenier Sosa, Noemi and Roberto Porro, Pete Gozar, Samantha Stone, Tracy VanHolt and Zulma Villegas. Ultimately, I am truly grateful for the supportspoken and unspokenlove, encouragement (especially in those tough moments), humor and wisdom of my husband, Luis Gmez Paz. v

PAGE 6

TABLE OF CONTENTS Page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES .............................................................................................................ix LIST OF FIGURES ...........................................................................................................xi ABSTRACT .....................................................................................................................xiii CHAPTER 1 INTRODUCTION........................................................................................................1 2 TURTLES IN AMAZONIA.........................................................................................5 History of Turtle Use....................................................................................................5 Turtles in the River Ecosystem.....................................................................................8 Implications of Large Turtle Life History for Conservation and Management..........10 Legislative Stopgaps...................................................................................................12 Value of Wildlife, Including Turtles..........................................................................12 Value of Podocnemis Turtles in Eastern Lowland Bolivia........................................13 Price of River Turtles in Lowland Bolivia..........................................................16 Markets as they Apply to Turtles................................................................................17 Influence of the Market on Neotropical Wildlife.......................................................17 Negative Impacts of Markets...............................................................................18 Positive Effects of Markets.................................................................................20 Market Integration and Hunting IntensityEffects on People...........................20 Influence of Income on Hunting.................................................................................21 Theoretical Derivation of Hypotheses........................................................................26 Implications of Findings.............................................................................................29 Note on my Research Style and Findings...................................................................30 3 RESEARCH AREA: EASTERN BOLIVIA.............................................................32 River and Floodplain Habitats of Parque Nacional Noel Kempff Mercado.......32 Parque Nacional Noel Kempff Mercado and Ecological Benefits it Provides....34 The Community of Piso Firme............................................................................36 Indigenous Links to Piso Firme...........................................................................38 The Community of Remanso...............................................................................42 vi

PAGE 7

Diverse Cultural Links to Remanso....................................................................43 Note on the History of Natural Resource Use in the Area from the 1940s to 1980s...............................................................................................................44 4 EFFECTS OF HUMAN SETTLEMENT ON THE NUMBER AND SIZE DISTRIBUTION OF TURTLES................................................................................47 Introduction.................................................................................................................47 Study Area..................................................................................................................47 Survey Procedure.................................................................................................49 Habitat Description..............................................................................................53 Results.........................................................................................................................54 Mean Numbers of Turtles at Different Distances from Transect Center............54 River Edge Habitat Characteristics.....................................................................61 Size Differences among Sites..............................................................................63 Comparison of Size of Turtles at Different Distances from Transect Center.....66 Discussion...................................................................................................................69 Differences in Turtle Abundance in Relationship to Humans.............................69 Differences in Sizes in Relationship to Humans.................................................71 Conclusion..................................................................................................................72 5 SOCIAL STRATIFICATION, MARKET INTEGRATION, AND HOUSEHOLD CONSUMPTION.......................................................................................................74 Objectives and Hypotheses.........................................................................................74 Methods......................................................................................................................76 Sample Size Calculation......................................................................................76 Interview Administration and Questions.............................................................77 Wealth Ranking...................................................................................................78 Definition of Market Goods and Market Integration..........................................79 Socioeconomic Characteristics and Their Implications for Livelihoods............80 Turtle Meat Consumption in Piso Firme and Remanso......................................84 Differences Between Subsistence and Market Consumption of Turtle in Piso Firme and Remanso.........................................................................................86 General Socioeconomic Characteristics by Wealth Groups................................88 Differences Between Subsistence and Market Consumption of Turtle Among Wealth Groups.................................................................................................92 Discussion...................................................................................................................94 Network and Income Differences between Piso Firme and Remanso................94 Subsistence and Market Use by Community.......................................................95 Differences in Turtle Consumption Among Wealth Groups...............................99 Limitations of Data Collected...........................................................................101 Relevance of this Work to Future Research......................................................104 Conclusion................................................................................................................105 vii

PAGE 8

6 COMMUNITY-BASED CONSERVATION AND MANAGEMENT...................107 Need for River Turtle Conservation.........................................................................107 Role of Protected Areas and Wildlife Use in the Neotropics...................................108 Influence of Social Norms.................................................................................110 Harvest Rules.....................................................................................................111 Common-Pool Resources and Community-Based Management.............................114 Working Towards Community-Based Wildlife Management and Conservation.....116 Structure for Development of Community-Based Conservation..............................119 Participatory Possibility in Bolivia...........................................................................121 Adaptive Management of Common-Pool Resources...............................................122 Framework for Community-Based Turtle Conservation and Management.............123 Functioning Internal Organization....................................................................124 Capacity.............................................................................................................125 Incentives for Participation and Conservation..................................................125 Transparency and Accountability......................................................................125 Points for Discussion Towards a Conservation and Management Plan...................126 7 CONTRIBUTIONS AND FURTHER RESEARCH...............................................128 Introduction...............................................................................................................128 Theoretical and Practical Contributions...................................................................129 Effects of Hunting Pressure on Turtle Abundance and Size Composition.......130 Intensity of Turtle Harvest as Related to Distance to Market...........................133 Wealth and its Effect on Turtle Consumption...................................................134 Practical Application of Consumption and Market Integration Findings.........135 Further Research.......................................................................................................136 APPENDIX A BRIEF HISTORY OF PARQUE NACIONAL NOEL KEMPFF MERCADO......137 B DESCRIPTION OF VEGETATION CLASSIFICATION SYSTEM.....................141 C INTERVIEW QUESTIONS.....................................................................................142 D CHARACTERISTICS OF THREE WEALTH GROUPS.......................................144 LIST OF REFERENCES.................................................................................................147 BIOGRAPHICAL SKETCH...........................................................................................162 viii

PAGE 9

LIST OF TABLES Table page 2-1 Average values of carapace length and body weight of female Podocnemis expansa turtles in the Orinoco and Amazon River Basins, an indication of their value as protein resources........................................................................................14 2-2 Research questions addressed in this research and the hypotheses used to test the questions...................................................................................................................29 3-1 Numbers of flora and fauna species known to exist in Parque Nacional Noel Kempff Mercado, indicating biodiversity present....................................................35 4-1 Descriptive statistics for the turtles counted along the three 40-km transects (Piso Firme, Remanso and Mangabalito)..........................................................................54 4-2 Kruskal-Wallis test statistics comparing median number of turtles in segments within (Close) and beyond (Far) 10 km of the center point between the three sites. These results indicate that the difference is not random..........................................55 4-3 Mann-Whitney test statistics for the comparison of close vs.far segments within sites. Results indicate a significantly higher median number of turtles farther from the center of the Remanso transect only..........................................................55 4-4 Frequencies of counts greater than or less than/equal to median values within 10 km (Close) or beyond (Far) transect center points of the three sites........................56 4-5 Test statistics for test of difference between numbers of turtles counted at close vs. far segments from center point of three transects...............................................58 4-6 Comparison of differences between sites in numbers of turtles counted at close vs. far segments from center points..........................................................................58 4-7 Proportions of turtle size classes seen within sites at far and close categories with averages over all three sites and at Mangabalito, the less-hunted site.....................66 4-8 Kruskal-Wallis test statistics comparing all four size classes at close vs. far segments from center point at all three sites. Results indicate significant differences between size classes in close vs. far segments overall..........................68 ix

PAGE 10

4-9 Frequencies of size classes counted in relationship to medians for far vs. close transect segments when all three sites are combined...............................................68 4-10 Gamma test for known size classes at different distances compared between hunted and less hunted sites indicating a negative association between turtle size at close vs. far distances in hunted and less-hunted areas........................................69 5-1 Motivation for trips taken out of communities to San Ignacio, Bolivia and Pimenteiras, Brazil by residents of Piso Firme and Remanso..................................82 5-2 Descriptive and test statistics of socioeconomic characteristics that directly and indirectly influence turtle consumption detailed by community..............................83 5-3 Distribution of households in three wealth groups, with relative percent and percent of number of households in each wealth group...........................................88 5-4 Descriptive statistics of socioeconomic characteristics that indirectly and directly influence turtle consumption detailed by wealth group within Piso Firme and Remanso...................................................................................................................90 5-5 Differences in hours turtling between wealth groups in Remanso. Mann-Whitney U test shows that the wealthiest group does not dedicate time to catching turtles, while households in groups two and three do................................93 6-1 Comparison of geopolitical areas, human populations and population densities in the ranges of P. unifilis and P. expansa.................................................................114 x

PAGE 11

LIST OF FIGURES Figure page 2-1 Conibo Indians along Ucayali in 1866 Turtle eggs being crushed to extract oil.......6 2-2 Turtle being cooked on Amazon River......................................................................7 3-1 Map indicating the communities where research was conducted, nearby cities that community members have links with, the Park and the Paragu and Itenz Rivers.............................................................................................................33 3-2 Aerial view of seasonally flooded savanna habitat, a scarce habitat type in the tropical Americas, pictured here near the Flor de Oro guard station in Parque Nacional Noel Kempff Mercado..............................................................................35 3-3 Arco Iris Waterfall, a popular ecotourism destination on the Paucerna River in Parque Nacional Noel Kempff Mercado..................................................................36 3-4 Piso Firme from the Paragu River in the dry season..............................................37 3-5 Distribution of jobs that provide the major source of income for households interviewed in the community of Piso Firme, located along the Paragu River......37 3-6 Example of a ceramic pot presumed of Guarasugwe origin, unearthed in Piso Firme in November 2001.........................................................................................39 3-7 Panero basket made from motac palm fronds (Attalea phalerata) and used to carry harvested products in Piso Firme....................................................................41 3-8 Distribution of jobs that provide the major source of income for households interviewed in the community of Remanso, located along the Itenz River............43 4-1 Podocnemis unifilis turtles basking along the Itenz River.....................................51 4-2 Schematic diagram of the 40-km transects illustrating location of Piso Firme and Remanso at the center of the 40-km transects and categorization of km into distance categories....................................................................................................52 4-3 Number of turtles counted at Piso Firme, along the Paragu River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect.....................................................................................................................57 xi

PAGE 12

4-4 Number of turtles counted at Remanso, along the Itenz River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect.........57 4-5 Number of turtles counted at Mangabalito, along the Itenz River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect.....................................................................................................................58 4-6 Average number of turtles seen in each km over eight samples at populated sites (Piso Firme and Remanso) with one standard deviation indicated by bars on each kilometer and the communities at KM 20................................................................59 4-7 Average number of turtles seen in each km over eight samples at the unpopulated site (Mangabalito) with one standard deviation indicated by bars on each kilometer...................................................................................................................60 4-8 Meters of each habitat type at the three sites, shown individually by site...............62 4-9 Number of turtles in each size class observed with one standard deviation of the number of turtles in each size class at each site and relative percent of the turtles at the site in each of the four size classes noted.......................................................64 4-10 Proportions of turtles of different size classes by distance categories at the three sites...........................................................................................................................67 5-1 Main source of household income, in increasing order of cash-generating potential in Piso Firme and Remanso.......................................................................81 5-2 Proportions of meat consumed in a purposive sample of households from July to August 2000.............................................................................................................86 6-1 Map of P. unifilis range (dots represent sightings) in the Amazon and Orinoco River Basins,..........................................................................................................112 6-2 Map of P. expansa range (dots represent sightings) in the Amazon and Orinoco River Basins...........................................................................................................113 xii

PAGE 13

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HUMAN USE OF TWO SPECIES OF RIVER TURTLES (Podocnemis spp.) IN LOWLAND EASTERN BOLIVIA By Kristen Marie Conway August 2004 Chair: Michael W. Binford Major Department: Geography Centuries of harvest (of adults and eggs) have left remaining populations of two species of Amazonian freshwater turtles, Podocnemis unifilis and P. expansa, seriously reduced throughout their ranges. Today, these turtles remain important wildlife resources for ribereo communities in and near National Park Noel Kempff Mercado, in the Bolivian Amazon for subsistence and market consumption. I quantified turtle abundance and investigated the impacts of socioeconomic factors driving hunting pressure to determine whether turtle abundance differs depending on whether sites are adjacent to human communities or not and if different socioeconomic characteristics affect turtle consumption. The communities where my study was conducted may ultimately use this information to inform a management plan for their turtle resources. Basking counts at three sitestwo adjacent to human communities and one adjacent to an abandoned community sitewere conducted to obtain information on the xiii

PAGE 14

impact of hunting on turtle abundance. Tests of difference applied to these data show differences in turtle abundance between sites with and without human communities, which may be a result of hunting pressure. Size classes of turtles among the three sites were negatively associated with proximity to humans. Analyses of socioeconomic data show differences between subsistence and market consumption patterns of turtle between the two communities. My findings suggest different levels of market integration, not influenced by distance to market as anticipated, do affect turtle consumption. Local phenomena including household wealth, subsistence-based livelihoods and indigenous origins are suggested as explanations for turtle consumption. The combined abundance, size class and socioeconomic data gathered in my study have implications for the effects of increased access to cash income on wildlife resources and community-based conservation and management of species important in subsistence and market economies. Poorer households are more involved in turtle harvest and sale. The more subsistence-based community, which has indigenous origins, reports more turtle consumption. Increased access to cash, through road improvement or greater market integration may actually reduce turtle consumption, positively affecting turtle abundance. Since the communities studied lie in the buffer zone of a protected area, the results have implications for protected area management. The impacts of hunting and information on some of the driving forces behind it, provide baseline information for a management plan in one eastern lowland Bolivia setting, which may provide insight for community-based wildlife conservation elsewhere in Amazonia. xiv

PAGE 15

CHAPTER 1 INTRODUCTION The concern for the preservation of biodiversity in fragile environments has been voiced by researchers and many others the world over, and has been reflected in the environmental policies adopted by numerous countries. The importance given to biodiversity has defined a research agenda that seeks to identify endangered species, and strives to understand the dynamics of human behavior that threaten the sustainable reproduction of particular animal populations. Critical dimensions of this larger research agenda are studies that address the ways in which particular rural communities in lowland tropical regions exploit the environment for their livelihood, and harvest animals for sale in local and regional markets. My dissertation focuses on the harvest for consumption of two species of turtles, Podocnemis unifilis and P. expansa in rural lowland tropical communities. The site for my research is eastern Bolivia, on the Paragu and Itenz/Guapor Rivers. The research design targets three locations: Piso Firmea smaller and comparatively subsistence-oriented settlement, Remansoa larger, wealthier, and more market-oriented settlement, and a third site, Mangabalitothe site of a human settlement that was abandoned in 1996. Data collection and analyses are organized in terms of two different yet related topics. The first is based on field observations of the abundance and size distribution of turtles at varying distances from the human settlements. The objective is to generate quantitative estimates of the effects of turtle harvesting on the abundance/size structure of 1

PAGE 16

2 the turtle populations. The second analytical priority is based on socioeconomic data obtained through random surveys of households in Piso Firme and Remanso, the rural lowland tropical communities. In this case, the objective was to identify community and household-level variables associated with the harvest of river turtles. Taken together, the findings provide insights into human behaviors that profoundly affect wildlife species that have long played a major role in the riverine environment of Eastern lowland Bolivia and across the Amazon Basin. The results provide the baseline data that can be used to promote community-based conservation and management programsa strategy that has the most promising potential to alter human behavior in a manner that promotes the sustainability of threatened species. To set the stage for the empirical analysis, my initial objective is to review the historical uses of P. unifilis and P. expansa, describe the ecology of turtles, and summarize research that documents the ecological effects of the harvesting and commercialization of animals in the lowland tropics. Chapter Two is therefore divided into four sections: historic use; turtle ecology; value of turtles; and, turtles and markets. The characteristics of the research area are described in Chapter Three. Chapter Four presents data on changes in the number and sizes of turtles as one travels along the river from the human settlement sites to more distant regions, upstream and downstream. The organizing theme of Chapter Four is derived from geographic theory that examines the spatial relationship between distance to market and value of agricultural goods by adapting this relationship to a wildlife resource. The specific hypotheses tested are: H1: turtles will be less abundant closer to human communities (where there is hunting); H2: there will be more turtles at Mangabalito (less hunting

PAGE 17

3 pressure) than at Remanso (more hunting pressure); and, H3: there will be fewer large turtles closer to human communities than farther away (because hunters prefer larger turtles). Chapter Five turns attention to characteristics of the communities of Piso Firme and Remanso. The analyses are based on effects of market integration and social stratification on household consumption of turtles. The following hypotheses are tested: H4: there is greater market consumption of turtle protein in Remanso than in Piso Firme and greater subsistence consumption of turtle in Piso Firme than in Remanso (because there is closer access to market, more cash and more traffic in and out of the latter); and, H5: there is a negative relationship between wealth and turtle protein sale/trade such that as personal wealth increases, the consumption of turtle protein decreases (based on initial field observations). The findings point to a number of important conclusions: the greatest threat to turtles is local subsistence consumption; and geographic location influences cash-generating possibilities, which in turn affect social stratification and market integration. The results of Chapters Four and Five can provide the initial basis for a community-based conservation and management initiative. The justification for endorsing this environmental strategy is summarized in Chapter Six, which is organized around five major themes: 1.) the need for turtle conservation; 2.) the role of protected areas and wildlife use; 3.) common property resource use and community-based management; 4.) participatory possibilities; and, 5.) a roadmap for community-based conservation and management in Piso Firme and Remanso.

PAGE 18

4 Chapter Seven synthesizes results of the different research topics and offers conclusions based on my findings. Future research is suggested.

PAGE 19

CHAPTER 2 TURTLES IN AMAZONIA History of Turtle Use For centuries, humans in the Amazon and Orinoco River Basins have exploited freshwater turtles of the genus Podocnemis. Records from both river basins relayed by early explorers (including Bates [1863] 1962; Bozo [1816] 1991; Herndon [1854] 1952; Humboldt 1852; Keller 1874; Le Cointe 1945; Maw [1831] 1989; Smyth and Lowe [1836] 1973; Spix and Martius 1968; and Spruce 1908) cite the abundance of P. unifilis and P. expansa and their extensive harvest by local inhabitants particularly during nesting season. Anthropological evidence for turtle use by humans dates to pre-Hispanic times (Castro de Len 1986; Licata 1992) with the earliest written records of human use of Podocnemis unifilis (Paez and Bock 1998) and P. expansa (Castro de Len 1986) dated prior to the 16 th century. Illustrations contributed by early explorers such as Marcoy (1875) support historic records that cite the utility of Podocnemis turtle eggs to indigenous, colonist and mestizo people throughout the Amazon Basin (Castro de Len 1986; Robinson and Redford 1991) (Figures 2-1 and 2-2). Prior to the introduction of kerosene and vegetable oil substitutes, P. expansa eggs provided the raw material for cooking and lighting oil production. An estimated 48 million eggs, likely to have represented the nesting efforts of thousands of turtles annually, were used to make oil in the 19 th century (Bates 1962; Smith 1974). Some of these early explorers quantified the harvests of turtles and their eggs, noting an apparent decline in their abundance over the course of time. One of these explorers, Bates, noted that 5

PAGE 20

6 At least 6,000 jars, holding each three gallons of the oil, are exported annually from the Upper Amazons and the Madeira to Par, where it is used for lighting, frying fish, and other purposes. It may be fairly estimated that 2000 more jars-full are consumed by the inhabitants of the villages on the river. Now, it takes at least twelve basketfull of eggs, or about 6000, by the wasteful process followed, to make one jar of oil. The total number of eggs annually destroyed amounts, therefore, to 48,000,000. As each turtle lays about 120, it follows that the yearly offspring of 400,000 turtles is thus annihilated. A vast number, nevertheless, remain undetected; and these would probably be sufficient to keep the turtle population of these rivers up to the mark, if the people did not follow the wasteful practice of lying in wait for the newly-hatched young, and collecting them by thousands for eating; their tender flesh and the remains of yolk in their entrails being considered a great delicacy. The universal opinion of the settlers on the Upper Amazons is, that the turtle has very greatly decreased in numbers, and is still annually decreasing. (1863: 364-365). Figure 2-1. Conibo Indians along Ucayali in 1866 Turtle eggs being crushed to extract oil (Marcoy. 1875. Travels in South America: From the Pacific Ocean to the Atlantic Ocean. Vol II. New York. Scribner, Armstrong and Co.: 35). Writing even earlier, Herndon noted [p]rolific as they are, I think the turtle is even now diminishing in number on the Amazon. Vast numbers of the young are eaten by the Indians, who take them by

PAGE 21

7 the time they are able to crawl and when they do not measure more than an inch in diameter, boil them, and eat them as a delicacy (1854: 157-8). These passages, among the earliest written records of Podocnemis turtles, mention their use and possible vulnerability in a setting shared with humans dating back at least two centuries. This pressure on turtles has likely increased as human populations in the Amazon Basin have grown (Allen and Barnes 1985; Darwin et al. 1996; Godoy 2001). As European influence spread in the Amazon and Orinoco River basins, so too did large-scale harvest of Podocnemis turtle eggs for oil and turtles for protein by ribereos, as depicted by Keller (1874) in Figure 2-2. Since the 16 th century, these turtles have exhibited declining population trends similar to those of other threatened and endangered wildlife, mainly believed to be the result of overhunting (Alho 1985; Ergueta & de Morales 1996; Escalona & Fa 1998; Johns 1987; Luxmoore et al. 1988; Salati et al. Figure 2-2. Turtle being cooked on Amazon River (Keller. 1874. The Amazon and Madeira Rivers: Sketches and Descriptions from the Notebook of an Explorer. London. Chapman and Hall: 74.).

PAGE 22

8 1990). Local residents and scientists have noted significant declines in populations of these turtles throughout their ranges (Alho 1985; Ergueta and de Morales 1996; Escalona and Fa 1998; Johns1987; Luxmoore et al. 1988). Turtles in the River Ecosystem River turtles, including P. unifilis and P. expansa, inhabit lowland higher-order sections of rivers characterized by broad, deep channels and rapid water velocity (Moll and Moll 2000). In the Bolivian Amazon, as in other areas of Amazonia, turtles share these habitats with humans. People are attracted to rivers for their abundant water supply, food resources, agricultural foundation and transportation (Brondizio et al.1994; Denevan 1996; Hiraoka 1985; Moll and Moll 2000; Moran 1991). River turtles are thought to contribute a significant proportion of biomass to river ecosystems and through their scavenging provide a valuable contribution to the cycling of nutrients (Moll and Moll 2000). Due to their sensitivity to environmental change, P. unifilis and P. expansa are also indicators of potentially negative ecosystem changes (e.g. pollutants that harm ecosystem function) (Oficina Regional de la FAO para Amrica Latina y el Caribe 1988). In the rainy season, P. expansa, a frugivore, enters the flooded forests where it finds abundant food in the form of fruits and seeds from trees (Moll and Jansen 1995; Ojasti 1971; Smith 1999). This behavior is significant because it contributes to the relationship of the tropical forest fauna on tree diversity (Wilcox 1995: 45). Multiple studies show Podocnemis species disperse seeds in the flooded forests of Amazonia (Bruno Coca 1999; Fachn Tern 1992; Kubitzki and Ziburski 1994; Ojasti 1971). Moll and Jansen (1995) specifically found that turtles play a role in the dispersal of plants, by acting as seed dispersers, secondary to birds and monkeys. In their study of

PAGE 23

9 aquatic black wood turtle (Rhinoclemmys funerea) and terrestrial brown wood turtle (R. annulata) turtles at Tortuguero, Costa Rica, Moll and Jansen (1995) found the majority of seeds in turtle stomachs to be from riparian, wetland and second-growth plant species commonly found in and adjacent to canals and streams. Germination of eaten versus uneaten seeds was not significantly differentevidence for the role of turtles in the dispersal of plants (Moll and Jansen 1995). [T]he potential for this [P. expansa] and other frugivorous turtle species (e.g., P. unifilis) to act as seed dispersers (with fish) in these cyclically flooded, forest ecosystems seems especially promising and warrants further investigation" (Moll and Jensen 1995: 126). A benefit to flooded forests of frugivorous turtles is that the low concentration of oxygen in the stagnant water does not negatively affect the turtles because they get their oxygen from the air. In fact, this characteristic confers an advantage to turtles, by reducing the concentrations of frugivorous fish that compete with them for fruit (Ojasti 1971). This biological attribute enhances the mutually beneficial relationship between turtles and tropical forest diversity. Frugivory in P. expansa has been reported throughout most of its range (e.g. Bruno Coca 1999 in Bolivia; Fachn Tern 1992 and Goulding 1980 in Brazil; Ojasti 1971 in Venezuela; Soini 1995b in Peru). Ojasti (1971) examined the stomach contents of ten adult P. expansa turtles from the Orinoco River and found that 86% of the ingested material was fruit from trees, 4% was leaves and stalks of diverse plants and the remainder was aquatic and terrestrial invertebrates, bones from cattle, fish and turtles in insignificant amounts. In Brazil, Fachn Tern et al. (1995) found that P. unifilis ate more seeds and fruits from the flooded forest than from rivers and lakes. In Bolivia, Bruno Coca (1999) showed that the diet of P. expansa is largely composed of fruits and

PAGE 24

10 seeds from the forest while P. unifilis ingests fruits and seeds more as a complement to its diet, which is mainly composed of leaves and stalks of terrestrial and aquatic plants. Bruno Cocas study (1999) suggests that P. unifilis and P. expansa in and around Parque Nacional Noel Kempff Mercado (the Park) act consistently with their conspecifics in other parts of the Amazon Basin as seed dispersers. These studies in Costa Rica, Brazil, Bolivia, Peru and Venezuela suggest that river turtles, including Podocnemis sp., depend upon access to the flooded forests and are important links in the life cycles of trees and other vegetation in the forest floodplains. Implications of Large Turtle Life History for Conservation and Management Some general life history traits of large turtles are thought to reduce their population resilience in the presence of hunting pressurethe main threat that accompanies sharing habitat with humans. Freshwater turtles like P. unifilis and P. expansa are long-lived species characterized by low adult mortality and high losses at both embryonic and hatchling stages (Gibbons 1968). Turtle populations are further threatened because nesting beaches overlap with human settlements, which has resulted in P. unifilis and P. expansa being particularly vulnerable to hunting and their populations reduced throughout most of their ranges (Cantarelli 1997; Ergueta and de Morales 1996; Johns 1987; Luxmoore et al. 1988; Mittermeier 1978; Pritchard and Trebbau 1984; Shaw 1991; Smith 1974). Because of the annual nesting season of Podocnemis turtles, populations are under heavy pressure for approximately 120 days each year, depending on location (June to September in my research area) (Pritchard and Trebbau 1984). The main threats to eggs are flooding and human predation (Escalona and Fa 1998; Luxmoore et al. 1988). If eggs survive their incubation period, upon the hatchlings

PAGE 25

11 emergence from the nest a host of avian, amphibian and reptilian predators await (Alho 1985). The most serious predators however, are humans. In a study of P. unifilis nest survival on the Nichare and Tawadu Rivers of Venezuela, Escalona and Fa (1998) reported that human collection of eggs affected 84.9% of nests (N = 351), with removal of all eggs from 70.6% of the human-disturbed nests (animals and flooding also destroyed nests, but to a much lesser extent). As a colonial nester (large numbers of turtles gather concurrently on particular beaches over several days to lay their eggs) P. expansa is even more susceptible to large-scale nest excavation, which can lead to higher percentages of nests being destroyed. In population ecology, five factors influence the rate of reproduction. These factors are clutch size; number of clutches per year; minimum and maximum breeding ages; sex ratios; and density effects. The harvest of nesting adult females is especially harmful because it poses a significant threat to the turtle populations survival by removing key reproducers (Congdon et al. 1993). P. expansas nesting-style vulnerability combined with the sizeable amount of meat they offer makes them particularly attractive prey for humans. Although P. unifilis exhibits different nesting site preferences and is not a colonial nester, humans have learned and pass on the necessary insight to detect their nests efficiently and these turtles are also threatened. The combination of a slow reproductive rate, ease of captureparticularly of reproductive females during nesting seasonand historic harvest pressure for subsistence and market consumption has decimated turtle populations throughout the Amazon Basin (Johns 1987; Mittermeier 1978; Pritchard and Trebbau 1984). These factors negatively influence the reproductive efforts of freshwater turtles of the genus Podocnemis found in eastern lowland Bolivia in

PAGE 26

12 the rivers adjacent to the communities of Piso Firme and Remanso. Greater market integration and increased demand for turtle meat may make populations even more susceptible (Campbell 1998) and are considered in my study. Legislative Stopgaps Multiple forms of legislation have been drafted and passed in attempts to reduce hunting pressure on these turtles. Both P. unifilis and P. expansa are listed on the Convention on International Traded in Endangered Species (CITES) Appendix II, which legislates regulation and control of international trade for their protection. Additionally, both species are rated one on the Tortoise and Freshwater Turtle Specialist Group action plan rating, a priority conservation-rating plan indicating that they are in trouble through over-exploitation for eggs and meat, and some are particularly vulnerable at colonial nesting sites (Tortoise and Freshwater Turtle Specialist Group 1991: 10). Brazilian law prohibits hunting of any wildlife found in its territory for sale nationally or internationally (Article 1 Law 5,197 January 3, 1967 and Article 29 of Law 9,605 February 12, 1998). While these laws and treaties exist, use that threatens turtle existence continues. Value of Wildlife, Including Turtles The forests and rivers of Amazonia contain tremendous faunal and floral diversity, including many species that are important resources for humans (Redford and Padoch 1992; Smith 1999; Terborgh 1992; Wilson 1988). Among the important resources are river turtles (Aramayo Cuenca 1989; Moll and Moll 2000). Surveys assessing the consumption of wildlife have cited the importance of wildlife as food resources to humans (Bennett 2002; Edwards and Abivardi 1998; Fa et al. 1995; Funes and Novaro 1999; Robinson and Redford 1991; Santos Brito and Ferreira 1978).

PAGE 27

13 Santos Brito and Ferreira (1978) quantified Amazonian wildlife consumption in Manaus in 1976 and seven additional cities in 1977. Their study identified species preferred in restaurants, suggesting the pressure facing animals in the wild. Both P. unifilis and P. expansa were mentioned by participants in these studies as being desirable (Santos Brito and Ferreira 1978). Sale of bushmeat, which was outlawed in Brazil in 1967 and is restricted within and near the Park, continues today (e.g. Johns 1987 and Conway 1999-2001 personal observation). This translates into hunting pressure that is greater than just subsistence hunting pressure alone. In his work on Amazon River turtles in Tef Johns (1987) noted that during the 1985 season, the average market price per turtle for a P. unifilis was US$11-14 but dropped to US$7 if several boats arrived in port simultaneously. Podocnemis expansa was reputed to have a less desired taste among ribereos but was more highly prized in the city where individual turtles sold for US$60 (range $40-$130) in the market while a portion in a restaurant was US$13-$20 (Johns 1987: 26). The giant South American river turtle (P. expansa) is unusual among bushmeat because it tends to be more expensive than domestic meat (Ojasti 1996). In 1978, the market price for P. unifilis and P. expansa meat in seven Brazilian cities was US$0.91 per kilogram while servings in a restaurant were US$7.83 and US$9.56, respectively in 1978 US$ (Chinese University of Hong Kong 2000 for the currency conversion; Santos Brito and Ferreira 1978 data). Value of Podocnemis Turtles in Eastern Lowland Bolivia Podocnemis turtles are large, widely distributed and provide a sizeable amount of high quality protein (Pritchard and Trebbau 1984). The largest recorded length and weight of a P. expansa specimen is 89 cm and 45 kg (Luxmoore et al. 1988; Pritchard

PAGE 28

14 and Trebbau 1984). Turtles of this size can provide 10 to 12 kg of high-protein meat (Table 2-1). In terms of protein content, Cantarelli (1997) states that turtle meat contains 85 to 88% protein vs. 43 to 70% for domestic meats (pig, chicken, and beef). Ribereos (floodplain residents) in eastern lowland Bolivia state a preference for turtle meat over all domestic meats and claim it offers more nutritional value (various residents personal comm. 1999, 2000) Preliminary fieldwork in three Bolivian communities (Piso Firme on the Paragu River and Remanso and Bella Vista on the Itenz River) in August 1999 confirmed that ribereos there consume P. unifilis and P. expansa meat and eggs. Observation of middens and interviews with 36 individuals in the two larger communities (Piso Firme and Remanso) indicated that turtle meat and eggs are sources of both meat and cash/trade value. The meat and cash/trade values of P. unifilis and P. expansa appeared to influence livelihood strategies in this area, which may have long-term effects on the humans, the turtles and the natural systems. Table 2-1. Average values of carapace length and body weight of female Podocnemis expansa turtles in the Orinoco and Amazon River Basins, an indication of their value as protein resources Source Year Location Average Carapace Length (cm) Average Body Weight (kg) N Ojasti (1971) Early 1960s Orinoco River, Venezuela 23.3 100 Ojasti (Pritchard and Trebbau 1984) 1962 Playa del Medio, Estado Apure, Venezuela 63.1 Unknown Ojasti (Pritchard and Trebbau 1984) 1964 & 1966 Playa del Medio, Estado Apure, Venezuela 64.1 & 64.0 Unknown Vanzolini (1967) 1967 Trombetas River, Brazil 70.1 38 Pritchard (Pritchard and Trebbau 1984) 1977 Playa del Medio, Estado Apure, Venezuela 65.1 20

PAGE 29

15 Pritchard (Pritchard and Trebbau 1984) 1977 Playa Pararuma, Estado Apure, Venezuela 66.2 20 Paollilo (Pritchard and Trebbau 1984) 1981 Playa del Medio, Estado Apure, Venezuela 66.8 26.0 302 Conway Present work Itenz/Guapor and Paragu Rivers, Bolivia 59 (sd=10.67) 22 (sd=8.77) 8 Included in Table 2-1 are the results of a sample of eight (for length) and six (for weight) P. expansa turtles captured on the Itenz and Paragu rivers in eastern Bolivia. I was able to weigh the components of one ~28 kg specimen of P. expansa as it was slaughtered for consumption. A 28-kg female P. expansa caught and weighed as it was slaughtered was constituted as follows: carapace (dorsal shell) ~ 5 kg; plastron (ventral shell) ~ 11 kg; organs ~ 1.5 kg (including eggs yolks in the uterus); and, meat ~ 10 kg. Although the turtles I measured were smaller on average they do, nonetheless, provide a meat resource for local people, who highly value it. Podocnemis unifilis, the yellow-headed sideneck turtle, shares much of the same range with P. expansa, extending somewhat further south (Iverson 1992; Pritchard and Trebbau 1984). It is smaller with a mean carapace length that ranges from 34.7 to 46.3 cm (Pritchard and Trebbau 1984). Reported weights of two female P. unifilis specimens weighed by Pritchard were 8.1 and 9.0 kg (Pritchard and Trebbau 1984). A sample of ten P. unifilis measured during my fieldwork in 2000 yielded an average carapace length (measured longitudinally/in a straight line from anterior to posterior end) of 38 cm (sd=7.90) and an average weight of 7 kg (sd=1.53). Because P. unifilis is much more locally abundant than P. expansa (Pritchard and Trebbau 1984) the yellow-headed sideneck turtle is a more frequent source of protein for ribereo communities in lowland eastern Bolivia today.

PAGE 30

16 Price of River Turtles in Lowland Bolivia In 2000-01 whole P. unifilis in rural eastern lowland Bolivia averaged US$3.25. In 2000-01 the price for a female P. expansa in the same area ranged from US$8.13 to US$16.26 (50 bolivianos) for a 20 to 30 kg turtle, which is similar to Ojasti (1996) and Johns (1987). Since ribereos in general also have limited access to cash-generating income, their preference for turtles may be a result of economic scarcity as much as historic and cultural preferences. Based on the measurements of a female P. expansa in Bolivia, a yield of 41% comestible product is possible. Extrapolating from a sample of ten female P. unifilis captured in my study area for local consumption that weighed an average of 7 kg (sd=1.45), a yield of 3 kg of meat each may be expected. If a 7-kg P. unifilis turtle yielded 3 kg of edible meat, each kilogram would cost roughly the equivalent of US$1. In 2000-01 the prices of meat in my study area were: beef US$1/kg; chicken US$1/kg; and fish US$ 0.75/kg. Daily minimum wage during this time was US$5.00. Thus, P. unifilis turtle meat is valued roughly equivalent to the prices of domestic meat (beef and chicken). The average daily meat consumption based on a purposive sample taken from November 2000 to May 2001 was 1.64 kg (sd=1.206 n=12) in Piso Firme and 1.23 kg (sd=5.157 n=17) in Remanso. These figures indicate a daily range of US$1.23 to US$1.64 in Piso Firme and US$0.92 to US$1.23 in Remanso potentially spent on meat. If households buy the meat consumed they will spend anywhere from 18 to 33% of daily wages on this commodity. Relatively speaking this is a high percentage, which suggests the economic and cultural worth of meat in these communities.

PAGE 31

17 Markets as they Apply to Turtles Von Thnen ([1826] 1966) described the correlation between distance to market and intensification of agricultural ecosystems. All other things being equal, the intensity of production depends upon the price a farmer gets for a product, which is directly dependent on transport costsa function of market distance (Von Thnen [1826] 1966). Von Thnens intensity theory posits that sites nearer the market, which have higher prices, should be cultivated/harvested more intensely (Von Thnen [1826] 1966). There is a distance at which diminishing returns dictate that intensive production is no longer cost effective. In 1999 in the northeastern Brazilian Amazon, residents in a community on the Ro Iriri, a tributary of the Xingu River in the Brazilian state of Par, earned US$5 per P. expansa turtle from middlemen, who then sold the turtles in Manaus for a profit (R. Porro personal communication 1999). Reports of market prices for P. expansa in urban Amazon markets range from $100 per turtle (Smith 1999) to $150 per turtle in Manaus (Luxmoore et al. 1988). These figures suggest that the $5/turtle earned by ribereos may have been worth more to fishers than the meat provided, which suggests turtles are a cash-crop in emerging rural markets. It may also indicate market pressure on the turtles if each ribereo was capturing and selling multiple turtles. Most importantly, it points out the ongoing cash and market value of Podocnemis turtles. Influence of the Market on Neotropical Wildlife In addition to their monetary value, Podocnemis turtles are attractive from a market perspective because they stay alive for days without food or water. This is important in rural communities without refrigeration, for fishers on multi-day fishing trips and for shipping turtles to cities on the bus. Also, turtles are a luxury item for emigrants to urban

PAGE 32

18 areas who live more than a days journey away from their rural homes but maintain cultural and dietary preferences from home. Both positive and negative benefits for wildlife conservation and income-earning power result from increased value of species. Wildlife with high market value may face greater hunting pressure, which would in turn cause greater threats to their existence. Conversely, increased awareness, when accompanied by management or captive breeding of prized species, are positive outcomes for conservation that may result from market incentives (Bodmer and Puertas 2001; Bodmer and Pezo 2001). When done in a sustainable fashion (i.e. not heavily targeting both reproductive females and new recruits), harvest of wild species can offer benefits for human and wildlife species (Freese 1997b). Freese notes "[f]reshwater species that have been depleted during the twentieth century include several species of crocodilians (Messel et al. 1992), valued for their hides, and Amazonian river turtles (Podocnemis spp.) (Alho 1985), valued for their meat" (1998: 26). In another study, Ojasti (1996) cites turtles (specifically Podocnemis spp.) as being the reptiles most heavily affected by subsistence hunting in the neotropics. Assessing the market value of and impact on turtles as natural resources is important in analyses preceding conservation and management planning because of the history of use and the evidence that market access can have both negative and positive ramifications on natural resources (Freese 1997a, Godoy 2001, Johns 1987, Klemens and Thorbjarnarson 1995). Negative Impacts of Markets Market access has often been associated with increased harvest of goods ranging from agricultural to nontimber forest products (Godoy 2001; Godoy et al. 1998a; Redford

PAGE 33

19 and Robinson 1985; Sierra et al. 1999; Von Thnen [1826] 1966). Multiple studies around the world have pointed out negative effects of hunting practices on wildlife (Alvard 1994; Campbell 1998; Martin 1978; Ministerio de Agricultura y Cria 1972; Murray 2003; Redford and Robinson 1991; Terborgh and Van Schaik 2002). Redford and Robinson (1985) suggest that market integration reduces the sustainability of hunting by indigenous groups because of the introduction of goods and technologies that accompany integration. These changes can interrupt ways of life through demands for "the trappings of modern civilization [that] can cause or increase the over hunting of certain species" (Redford and Robinson 1985: 43, see also Fa et al. 1995). Hunting wildlife to extirpation negatively effects human populations that depend on it and the resource must be replaced if people are to maintain familiar lifestyles. Of course, removal of a faunal or floral component from an ecosystem can disrupt the functioning of natural systems as well (Gunderson and Holling 2002; Odum 1993). This is not to say that markets, technology or development should not be introduced, but it is a cautionary note on the potential effect of such factors on cultural practices that have ramifications for wildlife. For, as Bebbington (1996) points out in reference to the introduction of new tools and technologies for indigenous development, important components for improving lives without compromising social and political systems include restructuring social relationships and putting power into the hands of the people so they have the ability to increase their incomes. So, it is possible that shifts in income sources and empowerment can reduce pressure on wildlife and therefore be desirable ends of markets, development and technology.

PAGE 34

20 Positive Effects of Markets Markets have traditionally been observed as negative forces on natural resources, but research now indicates that there are certain conditions under which markets do not pose greater negative consequences and are even beneficial to natural resources (Agrawal 1995; Agrawal 2001; Angelsen 1999; Dolak and Ostrom 2003). Agrawal and Yadama (1997), working in forests in India, surmise that with the presence of paved roads, the market is brought closer because transportation costs are reduced, which then increases villagers awareness of the forest value and in turn their desire to conserve. Increased awareness and active conservation of resources can occur with greater market access (Agrawal and Yadama 1997; Anderson 1990; Godoy 2001; Schweik et al. 2003; Tang and Tang 2001). Market Integration and Hunting IntensityEffects on People Bodmer and Pezo (2001) note a disjunction between rural development schemes and sustainable wildlife use. In their view, rural people (as opposed to urban dwellers or international market players) suffer the greatest economic costs by converting unsustainable wildlife hunting practices into more sustainable practices. This is in direct conflict with a main objective of most rural development schemes, which strive to increase economic development (Coomes 1996; Yapa 1998). For rural people to hunt more sustainably, they must be able and willing to absorb the costs of change. This means that rural development projects must increase short-term economic income potential for the gain of long-term sustainable wildlife harvest. In cases where systems of use are demonstrated to be unsustainable, conservation plans must incorporate concessions or alternatives for the people in the short-term, and wildlife survival in the long-term.

PAGE 35

21 Bodmer and Pezo (2001) suggest that economic benefits may be gained in the form of compensation payments in lieu of hunting. To realistically establish payment plans that will offer some form of long-term relief from wildlife exploitation pressure, many factorssuch as origin and handling of funds and potential social effects of introduction of these fundsmust be taken into consideration. Additional payment options may include greater service employment opportunities (e.g. ecotourism) or alternative subsistence and market crops. A discussion of the amounts and effects of payments is beyond the scope of this work, it is mentioned here as one alternative for conservation. Influence of Income on Hunting Scholars of non-timber forest product extraction in Central America conclude that as market integration continues, bringing with it higher income-earning capacity, foraging for animals in these rainforests increases (Godoy et al. 1995). Hunting can and does cause wildlife populations to decline or disappear from areas (Martin 1978; Murray 2003; Peres and Lake 2003; Robinson and Redford 1994; Shaw 1991). Podocnemis turtles used by ribereos in the Bolivian Amazon are no exception to this phenomenon (Aramayo Cuenca 1989; Quiroga Vera 2000; Urea Aranda 2000). Godoy et al. (1998) found that in some cases involving sale of non-timber forest products, rural dwellers knowledge of animals natural history increases as commercial values rise, an offshoot that may benefit wildlife. In his work in Bolivian and Honduran tropical rainforests, Godoy (2001) expands on the consequences of increased market integration on wildlife use noting that as income increases, economic development will increase consumption of most game and generally will lead to increases in human populations. These two factorseconomic development and human population growthcan work synergistically to greatly increase pressure on

PAGE 36

22 wildlife through market availability of game meat. Redford and Robinson (1985) and Robinson and Redford (1994) found that market integration acted as a deteriorating factor in the sustainability of hunting by indigenous groups with the giant river turtle specifically noted. However, Godoy (2001) also notes that in some cases economic development will lower the price of game meat substitutes through commercial production, thereby reducing demand for game and thus hunting pressure. In his examination of market prices of wild-caught meat Godoy applies three categories. These categories are defined as: inferior goodstheir consumption falls as income rises, especially in households whose income is below the median; necessitiestheir consumption goes up less than 1% with each 1% rise in income in bottom-income categories and becomes and inferior good in the top income category; and, superior goodstheir consumption increases more than 1% with each percent rise in income (Godoy 2001). According to Godoys definitions and observations, fish is an inferior good and game meat is a necessity (Godoy 2001). "The evidence suggests that growth in income probably discourages the consumption of fish and increases the consumption of game meat, but at low rates"(Godoy 2001: 94). Whether Podocnemis turtle meat is analogous to fish or game meat is unknownthey are aquatic and terrestrial prey that provide greater catch per unit effort than most of the common fish in the area, but their capture requires some specific knowledge (which areas of the river/flooded forest to look in and best times for capture) and techniques (how and where to place baited hooks, how to use harpoons, where to wait on nesting beaches and when to approach). The specific knowledge and techniques that facilitate harvest add to the cost, which could serve to raise the price of this commodity. Higher

PAGE 37

23 catch per unit effort combined with the demand for the turtle meat and evidence for its consumption in San Ignacio (the nearest Bolivian city), at prices greater than in the communities, suggest turtle is a superior good. If turtle meat is a superior good, current consumption threats may be exacerbated with augmented market integration. If turtle meat is an inferior good, which implies it is less expensive, the stress on turtle populations from hunting will be reduced if incomes increase and other meats are more heavily consumed. This could protect turtles, barring other stresses. But, if turtle meat is a necessity and consumption increases at rates just below income increases and turtle population growth, it will continue to be threatened because consumption will continue to rise as incomes rise with growing availability of cash through more market integration. My assumption is that either through improved road access to San Ignacio or more involvement in cash-generating employment, these communities, especially Piso Firme, are bound to be more market integrated in the future. My observations during preliminary fieldwork suggested that turtle meat was an inferior good for households with greater access to economic resources. However, increased market integration could make it a necessity or a superior good. Hypotheses four and five were set up based on the relationships between wealth, market access and turtle consumption. Piso Firme and Remanso have ties to the markets of San Ignacio and Santa Cruz de la Sierra, Bolivia and Pimenteiras, Brazil. In general, residents of Piso Firme rely on goods coming from San Ignacio while Remanseos rely more heavily on goods purchased in Pimenteiras, Brazil. Since at least 1999 there has been talk of improving the road that connects the communities to San Ignacio, which would reduce transportation

PAGE 38

24 costs (currently it is a 12 to 24 hour bus trip, depending on the season). Reduced transport costs may affect a range of phenomenafrom market access to migration patterns to access to cash. Such changes may influence turtle consumption patterns, which is why I emphasize the impact of the market here, using distance as a proxy indicating market integration. There are presently several cash-generating activities in my research areafishing with Brazilian commercial fleets, mining for gold, and cocaine running and other illicit goodsthat boost the facility of purchasing goods. These activities in all likelihood will continue to exist until the resources are depleted, as in the case of the fisheries and mining, or the activity is deemed unworthy of the risk, as in the case of cocaine and other illicit goods. In the case of cocaine, Castells and Laserna point out that short of massive chemical defoliation, which could trigger an ecological catastrophe and a cultural war, coca production cannot be eradicated (1994: 76), which has direct effects on the continuance of cocaine running that is a source of income for several households in each community. In the case of other illicit goods, rural frontier settlements like Remanso are prime locations for the transport and resulting financial rewards of such practices. The impact of disposable cash on turtles could be increased or sustained levels of consumption if turtle is a superior good or a necessity. Conversely, increased cash availability could have a positive effect on turtle populations if it is an inferior good and consumption declined with growing incomes. Because road improvement is a community and Park priority, the effect of the market on P. unifilis and P. expansa in the Bolivian Amazon must be considered because

PAGE 39

25 a better road will change the lifestyles of these ribereos through an improved linkage with the city and its markets. Ayres et al. (1991) had an opportunity to observe wildlife harvest rates in a Brazilian Amazon village pre(1978) and post-construction (1980) of a road. In the case of Dardanelos, hunting yields and number of hunters decreased after the construction of the road, suggesting reduced hunting pressure (Ayres et al. 1991). Furthermore, the road into Dardanelos was associated with reduced consumption of game meat and increased consumption of domestic meat (Ayres et al. 1991), which suggests that game meat was an inferior good whose consumption declined with greater incomes. These are positive impacts of a road for wildlife. But a road also represents greater market access and other changes. So, while there can be positive effects for wildlife, at least as suggested in the two years immediately following construction, the longer term effects associated with road construction and their impacts on wildlife must be considered. As Ayres et al. (1991) pointed out, the road brought with it considerable sociocultural change because of integration into the national economy. Despite the fact that hunting became less important for human subsistence, the integration of Dardanelos into the national market economy by the road brought other threats to the regional wildlife, including logging, cattle ranching, mining, and large-scale agriculture (Ayres et al. 1991: 92). Particularly because some of these activities already exist on a small-scale in Piso Firme and Remanso this note is cause for concern. Contemporary examples of P. unifilis and P. expansa use indicate that their meat and eggs hold market value for the people in the Amazon floodplain, both at the local scale observed in this research as well as at the larger Amazon Basin scale (as noted in Cavalcanti 1999). Figures from the greater Amazon Basin and my data from Bolivia, are

PAGE 40

26 evidence that turtles remain a sought-after resource at the local scale. Demand for turtles in the Amazon Basin (Cavalcanti 1999; Johns 1987; Luxmoore et al. 1988; Mittermeier 1978) coupled with growing human populations foretell continued and increased stress on the species. If international trade demands rise or the level of market integration in rural communities such as Piso Firme and Remanso grows, the negative effects on the turtles are sure to grow with them. Theoretical Derivation of Hypotheses My research questions in eastern lowland Bolivia were derived from two areas of inquiryecological and social (Table 2-2). First, I tested spatial relationships between humans and wildlife abundance and population structure based on wildlife ecology. Previous studies on wildlife use have shown contrasting impacts of hunting on wildlife. In many cases, hunting is viewed as detrimental to wildlife populations (Hames 1987; Martin 1978; Murray 2003; Peres and Lake 2003; Redford and Robinson 1985 and 1991; Robinson and Redford 1994). Yet, other studies have found that hunting does not always lead to negative impacts on wildlife species (Bodmer 1995; Bodmer et al. 1997; Chernela 1994; Fitzgerald 1994; Francis 1997; Vickers 1991). Furthermore, some studies indicate that differences in wildlife population structures may indicate hunting pressure and unsustainable use (Congdon et al. 1993; Robinson and Bodmer 1999). I wanted to see if similar patterns could be detected in river turtle populations in eastern lowland Bolivia and examine socioeconomic factors that might impact the effects of market integration on turtle use. Multiple studies cite evidence that suggests market integration contributes to decreased sustainability of hunting (Alvard 1994; Campbell 1998; Martin 1978; Ministerio de Agricultura y Cria 1972; Murray 2003; Redford and Robinson 1994;

PAGE 41

27 Terborgh and Van Schaik 2002; Robinson and Redford 1985). Measuring abundance is one relatively simple and quantifiable means of assessing sustainability when a comparison is made between hunted and less-hunted sites. The comparative assessment of suspected hunting impacts on turtles can provide data that corroborate market impacts. If population abundance is observed to be low around human settlements, this is preliminary evidence of potentially negative impacts of hunting. Like assessment of abundance, observation of turtle size in hunted vs. less hunted sites can suggest impacts of hunting pressure on the wildlife resource. Quantification of reptile population structure offers information about the maturity of the population, which gives an indication of recruitment and longevity (Caughley 1977). Noting the value of age structure for populations, Cole (1954) stated [w]hen the mortality factors affecting a population are altered either through natural environmental changes or through human exploitation or attempts at control there will in general result a change in the age structure of their population, and this may be observable even before changes in population size or in birth rates provide evidence of the consequences of the changed mortality factors (28-9). The second component of my research examined relationships between socioeconomic factors affecting market integration and its impact on wildlife use. I argue that the relationship between distance to market and intensity of turtle harvest is similar to the one Von Thnen described for agricultural goodsincreasing intensity of turtle harvest will coincide with closer market access (as suggested by Godoy et al. 1995). This relationship may be caused directly by increased value of meat or indirectly through

PAGE 42

28 increased demand for cash acquired through the sale of meat. Multiple socioeconomic factors contribute to the actual impact of the market on wildlife. Market integration can lead to increased harvest of natural resources, including wildlife (Godoy 2001; Godoy et al. 1995 and 1998a; Santos Brito and Ferreira 1978; Sierra et al. 1999; Von Thnen [1826] 1966). Specifically I wanted to examine the relationship between market integration and river turtle use in an eastern Bolivian context where there are different levels of market integration. Quantifying turtle consumption at sites with different ties to markets is one means of assessing the market impact on a wildlife commodity. The comparison of turtle consumption between Piso Firme and Remanso offers a chance to determine what the market impact on turtles is in these settings. Comparing several additional socioeconomic characteristics provides an opportunity to add further insight into turtle consumption. This additional information may offer insight into the driving forces of turtle consumption if market influence does not prove a main force affecting consumption. Income has varying effects on wildlife meat consumption (Godoy 2001). Perception of wealth, as a proxy for income, is one measure of potential impact on turtle harvest. Quantifying this relationship offers an opportunity to better understand factors driving turtle consumption and ultimately informs conservation planning. The socioeconomic characteristics examined were chosen based on the preceding review of literature on markets and natural resources, prior knowledge of the communities from a preliminary site visit and observations during my field research that suggested characteristics that directly or indirectly influence turtle consumption. A

PAGE 43

29 summary of the questions that drove my research and the specific hypotheses to test these questions are stated below. Table 2-2. Research questions addressed in this research and the hypotheses used to test the questions. Question Addressed Hypothesis Tested What is the spatial relationship between human communities and Podocnemis unifilis and P. expansa turtle abundance? H1: turtles will be less abundant closer to human communities H2: there will be more turtles at Mangabalito (less hunting pressure) than at Remanso (more hunting pressure) What is the spatial relationship between location of human communities and size classes noted of Podocnemis unifilis and P. expansa turtles? H3: there will be fewer large turtles closer to human communities than farther away How does distance to market affect subsistence vs. market consumption of turtle? H4: there is greater market consumption of turtle protein in Remanso than Piso Firme and greater subsistence consumption of turtle protein in Piso Firme than in Remanso How does household wealth affect turtle consumption? H5: there is a negative correlation between wealth and turtle protein sale/trade such that as personal wealth increases, the consumption of turtle protein decreases Implications of Findings Fewer turtles near human communities is one indication of hunting pressure as a depressing element on turtle populations. The implication here is that this effect can continue spreading out away from communities and threatening even greater portions of turtle populations, as has been observed in other parts of the Amazon Basin. Extinction is the ultimate threat. Rejection of null hypotheses one and two are indicators of negative impacts of hunting pressure. Fewer large turtles closer to humans imply negative effects of hunting on population maturity. Populations that are not able to maintain sufficient recruitment as measured by mature females, will eventually disappear. Rejection of null hypothesis

PAGE 44

30 three is another potential indicator of negative hunting pressure on turtles in my research area. If there is more market consumption (sale of turtles, sending turtles and eggs out of the community, preference for cash value of turtles) in Remanso, which is closer to its market links, then the market may be a negative influence on turtles. If there is more subsistence consumption (hours turtling, eating turtle and preference for turtle over its cash equivalent) in Piso Firme, which is farther from its market link, this will indicate that turtles face threats other than market integration. Rejection of null hypothesis four will suggest the role of the market and distance to market on turtle populations. Analyses of further socioeconomic characteristics (hectares of cultivated land, trips out of the community, years of education, hours fishing), indicate supplementary factors that impact turtle populations. As wealth increases, consumption of inferior meat decreases. If turtle is an inferior meat, greater wealth will be associated with decreased consumption. Increased wealth could have a positive effect on turtle populations. Because of potential changes coming to these communities through improved road accessincreased access to markets and cashquantification of turtle consumption under different wealth conditions has important implications and can inform potential conservation strategies. Rejection of null hypothesis five will indicate that turtle meat is an inferior good. Note on my Research Style and Findings It is worth noting that this research is largely top-down in spite of the participatory approach I describe and support. Local participation was included as much as possible in this research, mainly through training and participation of field assistants. Data have been returned to the communities to help illustrate portions of the outcome. And, most

PAGE 45

31 importantly, the research was designed to be accessible and it may be easily transferred to future community-based research efforts where community members collect similar data. Because of the influence of a national park on my research sites my efforts offer an opportunity to contribute to applied conservation through return of the data to the communities. This may be a starting point for discussion of conservation and management of the turtle resource. Insight gained into how household wealth and distance to market affect turtle consumption will further understanding of the interplay between socioeconomic systems and natural resources. This information is valuable in the design of conservation and management plans for natural resources by user groups.

PAGE 46

CHAPTER 3 RESEARCH AREA: EASTERN BOLIVIA River and Floodplain Habitats of Parque Nacional Noel Kempff Mercado The river turtles P. unifilis and P. expansa are found in eastern lowland Bolivia, inhabiting rivers that lie within the boundaries of the Park (Figure 3-1) (Iverson 1992; Ministerio de Desarollo Sostenible y Medio Ambiente 1996). There are five rivers within the Park, two of them included in my studythe Paragu and the Itenz. The Itenz River is the largest river associated with the Park, forming the northern and eastern limits as well as a portion of the frontier between Bolivia and Brazil. The Itenz River is a tributary of the Mamor River, which empties into the Madera River, a tributary of the Amazon River. Tributaries of the Itenz River include the Paragu, Verde and Paucerna Rivers. The fifth river in the Park, the Tarvo, is a tributary of the Paragu (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The Paragu and Tarvo Rivers form the western limit of the Park. The watershed of the Paragu River covers 5,500 km 2 and dominates the hydrology of the Park west of the Huanchaca Plateauthe source area for the Paragu and Itenz Rivers and a landscape feature that defines the eastern boundary of the Park (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The Huanchaca Plateau has eroded over the past 20 million years, exposing Precambrian material of the Brazilian shield and leaving a gently sloping landscape composed mainly of quartz with some granite (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The seasonal flooding of 32

PAGE 47

33 the Paragu River influences the area that lies to the west of the Huanchaca Plateau and east of Piso Firme. Huanchaca Plateau Figure 3-1. Map indicating the communities where research was conducted, nearby cities that community members have links with, the Park and the Paragu and Itenz Rivers. From their dark color, the Paragu and Itenz Rivers resemble blackwater rivers. However their properties are not strictly consistent with blackwater rivers, which are characterized as oligotrophic with strongly acidic water and rising in white podsolic soils in the Amazon Basin (Morley 2000). In contrast, the waters of the Paragu and Itenz Rivers have a nearly neutral pH and do not contain large amounts of suspended organic podsolic materialcharacteristics of clearwater rivers (Ministerio de Desarollo Sostenible y Medio Ambiente 1996; Navarro and Maldonado 2002). Clearwater rivers originate in Precambrian shield areas and are almost neutral but lack the dark coloration of blackwater rivers (Junk and Furch 1993; Morley 2000; Salati et al. 1987).

PAGE 48

34 A reflection of the absence of organic material in the Paragu and Itenz Rivers is the poor quality of nearby soils for agricultural use. Generally, soils in the Park have low nutrient, cation exchange, and fertility levels and a highly acidic pH. They have been classified as appropriate for forests, not agricultural activity (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). As is typical of this type of soils in the neotropics, the thin layer of leaves and organic material that lies on top provides most of the nutrients, requiring the presence of forest for maintenance of much biomass (Terborgh 1992). To date there is not extensive information on the Paragu and Itenz Rivers other than notation of their mixed limnological characteristics (Navarro and Maldonado 2002). Parque Nacional Noel Kempff Mercado and Ecological Benefits it Provides The Park includes 1.5 million hectares of undeveloped habitat in an ecological transition zone in the northeastern tip of the department of Santa Cruz, contiguous with Brazil (See Appendix A for a detailed history of the Park.). Within Bolivia, the Park is one of the most ecologically heterogeneous protected areas (El Deber November 30, 2000; Ministerio de Desarrollo Sostenible 1996), home to a great amount of biodiversity (see Table 3-1). The high level of biological diversity in the Park is related to the diversity of habitats within ittropical rainforest, gallery forest, subtropical scrub forest, seasonally flooded savanna (Figure 3-2), and seasonal dry tropical forest and savannathe only place in South America where these five ecoregions converge and are protected. Human occupation and use of these habitat types in other locations is substantial, increasing the value of such large, scarcely or wholly unpopulated contiguous areas (Ministerio de Desarrollo Sostenible 1996). In addition to its biodiversity assets potential, the Park offers ecotourism opportunities for national and international travelers. As a combined result of its area,

PAGE 49

35 biological characteristics, scenic and ecotourism opportunities, the Park was named a Patrimony of Humanity by the United Nations Education, Science and Environment Organization (UNESCO) in November 2000 (El Deber 2000). The Arco Iris waterfall, on the Paucerna River (Figure 3-3), is one of the impressive sites in the Park that tourists come specifically to see. Because images of Arco Iris are used in much of advertising for the Park it is often associated with the Park. Table 3-1. Numbers of flora and fauna species known to exist in Parque Nacional Noel Kempff Mercado, indicating biodiversity present. Flora or Fauna Estimated Number of Species Vascular Plants 4,000 Mammals 139 Birds 700 Reptiles 74 Amphibians 62 Fish 250 Invertebrates 347 Source: Ministerio de Desarollo Sostenible y Medio Ambiente 1996 and El Deber November 30, 2000. Figure 3-2. Aerial view of seasonally flooded savanna habitat, a scarce habitat type in the tropical Americas, pictured here near the Flor de Oro guard station in Parque Nacional Noel Kempff Mercado (K. Conway).

PAGE 50

36 Figure 3-3. Arco Iris Waterfall, a popular ecotourism destination on the Paucerna River in Parque Nacional Noel Kempff Mercado (K. Conway). The Community of Piso Firme Piso Firme, one of the two communities where this research was conducted, is located alongside the Paragu River, at 13.62 S latitude and 61.73 W longitude, which lies in the buffer zone of the Park. Piso Firme is 365 km N by road from San Ignacio, the provincial capital, and 770 km NNE of Santa Cruz de la Sierra, the departmental capital. Located on the Paragu River Piso Firme was an access point to former rubber plantations. This combination of access to work and a river for transport and resources made it an attractive site for settlement (de Mesa et al. 1999). Many aspects of life in the community center around the river. With the rise and fall of the Paragu River each year come respective scarcity and abundance of aquatic prey, including Podocnemis turtles. The photograph in Figure 3-4 shows Piso Firme from the Paragu River in the dry season, a time of abundant aquatic resources for predators.

PAGE 51

37 Figure 3-4. Piso Firme from the Paragu River in the dry season (K. Conway). According to the census I conducted in Piso Firme from June to July 2000 there are 473 permanent residents (present in the community for at least six months of the year) in 70 households. The main source of household income for Piso Firmeos was agriculture followed by salaried and day labor (Figure 3-5). Many families have some livestock Agriculture46%Sawmill7%Salaried11%Mining0%Fishing9%Day Labor11%Store9%Cattle0%Park7% Figure 3-5. Distribution of jobs that provide the major source of income for households interviewed in the community of Piso Firme, located along the Paragu River (N=55 households).

PAGE 52

38 commonly chickens, ducks and a few pigs. Some families have several head of cattle, but only six families have more than 15 head of cattle, with one household claiming 160 head. This is largely a subsistence agriculture community. Indigenous Links to Piso Firme Culture being a factor that influences daily life I describe here cultural antecedents of Piso Firme. I believe this background, which differs from that of Remanso, is worth pointing out because it influences use of natural resources. As such, this information is meant as relevant background that further informs the livelihood systems observed today, although an in-depth comparative analysis of Piso Firme and Remanso on these terms is beyond the scope of this work. The residents of Piso Firme are classified by the government and identify themselves as members of the Chiquitano indigenous group. In the 18 th century, floodplain residents in the area between the Paragu, Itenz and Paucerna Rivers were presumed to be from the indigenous group known as Guarasugwe and were referred to as Paucerna by mestizos in the area (Riester 1977). The Guarasugwe rejected Christian evangelization efforts as well as attempts to involve them in rubber collection. By the mid 20 th century, the Guarasugwe population was greatly reducedan estimated 50 individuals comprised the population in 1965. Remaining members had moved away from the banks of the Itenz in 1954, their original place of residence, and with the death of the last Captain, Tarekuv Miguel Frey, the group was considered culturally extinct by anthropologist Riester (1977). Archeological evidence linked to the Guarasugwe culture includes ceramic pottery found in the communities of Piso Firme, Bella Vista, Remanso, Cafetal, Porvenir, Puerto Rico and Florida, which lie in or near the Park. Pottery uncovered to date has contained

PAGE 53

39 human bones, tools and ceramic pieces (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). Figure 3-6 is a photograph of one of these pots unearthed in 2001 while residents were digging canals for water pipes in Piso Firme. Today a woman living in the community of Bella Vista, which lies 45 km NE of Piso Firme via the Paragu and Itenz Rivers inside Park boundaries, claims to be a Guarasugwe (interview with resident of Bella Vista, name withheld, August 1999). There are residents in Piso Firme who claim Guarasugwe descendence (interview with community leader, name withheld, April 2000). So, although at least one anthropologist considers the Guarasugwe culturally extinct, direct or indirect descendents of this indigenous group persist in small numbers in this area today, and maintain their cultural identification alongside the more common Chiquitano group. Figure 3-6. Example of a ceramic pot presumed of Guarasugwe origin, unearthed in Piso Firme in November 2001 (P. Torico). Overlapping the area previously occupied by the Guarasugwe and that of the newer cultural group known as Chiquitano, which is partially descended from the

PAGE 54

40 Guarasugwe. The area occupied by Chiquitanos is delimited by the Itenz River to the north, the Grande River to the west, the frontier with Brazil to the east and the Santa Cruz Corumb rail line to the south (Riester 1977). The Chiquitano cultural group is the result of the work of the Society of Jesus, through whose efforts in the 17 th century approximately 30 different ethnic groups were brought together in reducciones (villages created by Jesuit missionaries and populated by indigenous people of multiple groups) and acculturated over the following century (Riester 1985). Among the 30 indigenous groups were those whose language groups included Chiquita, Otuke and Arauak (Riester 1977). With the expulsion of the Spanish and Portuguese Jesuits in the late 18 th century this acculturated group established itself in the Amazon and Chaco regions of northeastern Bolivia, its settlements characterized by poverty and marginal conditions (Riester 1990 in Ministerio de Desarollo Sostenible y Medio Ambiente 1996). Several artifacts specific to the Chiquitano culture still used in Piso Firme today include the tac (a wooden bowl carved from cuchi, tajibo or chonta (Astrocaryum spp) and used for crushing, mashing or grinding foods), baskets made from motac (Attalea phalerata) and cus (Casearia gossypiosperma) and planting sticks made from chonta (Riester 1977). See Figure 3-7 for an example of a panero basket made of motac palm fronds and used in Piso Firme today. Residents of Piso Firme commented to me about the loss of additional utilitarian elements of the Chiquitano culture and other items, including music and the language. A small number of Piso Firmeos I spoke with about indigenous identity identify themselves as Chiquitano. But based on five conversations specifically about identity, it was apparent that there are conflicting sentiments about what it means to be indigenous.

PAGE 55

41 One prominent elder member of the community indicated to me that to be indigenous is to express a status that is shameful or low, while another prominent elder indicated that being indigenous served an end as far as the national government was concernedreferring to efforts to get land title as an indigenous community. A middle-aged man of modest means was quite proud of his Chiquitano heritage and lamented the loss of the culture, especially the music. Interestingly, one of the people most proud of being indigenous was a young man, approximately 30 years old, whose fathers opinion of the indigenous culture was that it was low. Figure 3-7. Panero basket made from motac palm fronds (Attalea phalerata) and used to carry harvested products in Piso Firme (K. Conway).

PAGE 56

42 Several older residents expressed a desire to teach the younger generation the Chiquitano language and other elements of the culture, stating that it is important to keep it alive in spite of all the migration and acculturation. In Piso Firme, only one older woman and a regular visitor from Santa Cruz, an older man, spoke Chiquitano. The man, a retired schoolteacher, visited annually and there was talk of the community contracting him to teach schoolchildren the Chiquitano language. The different views and level of cultural integration illustrate that this is clearly a community with one foot in the indigenous world and another foot in the mestizo world and possibly on the brink of making a choice to be in just one of these worlds. My impression of this community is that it is more likely to become further integrated with western culture and economy than to maintain its indigenous cultural ties and turtle consumption is likely to be affected by ensuing changes that result from this transition. The Community of Remanso The census I conducted in the community of Remanso from June to July 2000 indicated that there were 97 households with 591 permanent residents (present in the community for at least six months of the year). Remanso lies along the Itenz/Guapore River at 13.52 S latitude and 61.87 W longitude, on the border with the Department of Rondnia, Brazil. Like their neighbors in Piso Firme, the lives of Remanseos are dominated by the Itenz River that runs parallel to one edge of the community, but apart from this commonality, the communities have quite distinct demeanors. The distribution of income sources in Remanso is more diverse than that of Piso Firme and while the percentage of people who claim agriculture as their first source of income is lower, it is still the most prevalent form of household income. Many residents have small livestock and a few families have cattle, but few of these families have large herds and only two

PAGE 57

43 claim cattle as their primary source of income. In Remanso the majority of households obtain their income from agriculture, but unlike in Piso Firme, a substantial proportion of households%obtain most of their income through day labor or storekeeping (versus 20% in Piso Firme) (Figure 3-8). Remanso is truly a frontier townBrazil lies across the Itenz River. The heavier volume of river traffic on the Itenz River, which connects Remanso to Pimenteiras, Brazil, contributes to the stronger identification with this Brazilian city and the culture. Piso Firme, in contrast, has more of a rural agricultural settlement with stronger ties to Bolivia. Sawmill3%Agriculture33%Store17%Cattle1%Fishing8%Salaried13%Mining7%Day Labor18%Park0% Figure 3-8. Distribution of jobs that provide the major source of income for households interviewed in the community of Remanso, located along the Itenz River (N=71 households). Diverse Cultural Links to Remanso Remanso, which was established in 1906, a year after Piso Firme, does not share the same indigenous roots. It is a colonist settlement, begun in conjunction with the rubber boom in the area. This history leads to a different sense of community in this case from Piso Firme. I have several thoughts on how the colonist influences the

PAGE 58

44 communitys identity. Possibly because residents hail from more varied backgrounds, in this community the varied backgrounds impede a sense of common identity, or people want to fit in and therefore shy from pointing out their different cultural backgrounds. Whatever the reason(s) for this, I do not report on the diverse cultural antecedents of Remanso. Note on the History of Natural Resource Use in the Area from the 1940s to 1980s Communities in this region have experienced several shifts in employment sources in the 20 th century. These shifts have affected lifestyles and economies in the region through gradually increasing market integration. There were four stages of exploitation of natural resources, beginning with rubber in the early 20 th century. Rubber barons came to the region and had a strong economic influence in the Departments of Santa Cruz and Beni (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). Residents of the area, including Piso Firmeos and Remanseos, collected rubber, which transformed the labor structure by introducing a cash economy (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The Bolivian market collapsed when the rubber trade dropped off in the early 20 th century due to the fall of prices on the international market, the growth of improved rubber (Hevea brasiliensis) plantations in Malaysia and the introduction of synthetics. Harvest of a medicinal plant used to cure headaches, locally known as ipecacuana or poalla (Cephaelis spp.) was initiated following the rubber boom, although this plant did not develop into a large-scale trade item (Caballero 1994). The next stage of natural resource exploitation, which spanned from the 1940s to s, involved several types of wildlife, killed for their skins and feathers, for sale on the international market. The black caiman (Melanosuchus niger) and caiman yacar (Caiman yacare) were harvested

PAGE 59

45 from the 1940s through the 1970s in Bolivia (B. Godschalk personal communication January 2002). Spotted catsjaguar (Panthera onca), ocelot (Leopardus pardalis) and margay (Felis wiedii)were killed for their skins until the 1980s (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). In the 1980s psittacine birds, most notably the hyacinth macaws (Ara hyacinthinus) and scarlet macaws (Ara macao), were captured for the pet trade in North America and Europe (Luxmoore et al. 1988; Ministerio de Desarollo Sostenible y Medio Ambiente 1996). With the imposition of stricter laws on the trade and exploitation of wildlife through the Convention on International Trade in Endangered Species (CITES) in the 1980s, trade of these species has largely been reduced (Luxmoore et al. 1988). In most cases today, felines are killed to protect cattle or when they are thought to pose a threat to humans. During the year this research took place, at least two large cats were killed, including an adult male jaguar killed in a cattle pasture by a resident of Piso Firme who was en route to the forest to hunt. The hunter reported that the skin and skull would obtain a good price when sold to a buyer in San Ignacio, indicating that commerce in these items continues. More common today is resident involvement in the exploitation of forests for timber and other nontimber forest productsthere is a functioning palm heart processing center in Porvenir, another community in the zone of influence of the Park that lies 50 km south of Piso Firme, and a defunct one in Piso Firme. A natural resource whose exploitation is of concern today is the river turtle. A wildlife resource important at the regional scale, Podocnemis turtle populations have been diminished by trade for food and as pets since the 17 th century (Johns 1987; Luxmoore et al. 1988; Pritchard and Trebbau 1984; Cavalcanti 1999; Santos Brito and

PAGE 60

46 Ferreira 1978; Licata 1992). After the expulsion of the Jesuits in the late 18 th century and the introduction of substitutes for turtle oil, hunting pressure on P. unifilis and P. expansa dropped dramatically (Castro de Len 1986). While these turtles are not exploited as much for oil today as they were in the 18 th century, they still face hunting pressure at local and regional levels as subsistence and market food resources (Johns 1987; Klemens and Thorbjarnarson 1995; Ministerio de Agricultura y Cria 1972; Mittermeier 1975; Santos Brito and Ferreira 1978). The continued use of river turtles in Piso Firme and Remanso in the shadow of a protected area provided an opportunity to compare local socioeconomic and cultural contexts surrounding wildlife resource use near a conservation area. Because markets have been implicated in affecting wildlife use globally and because people in Piso Firme and Remanso have access to and are involved in markets, I have chosen to emphasize markets and factors affecting natural resource use as a result of market integration in my research. Deeper understandings of socioeconomic factors that directly and indirectly influence wildlife use go a long way in explaining communities needs, which are ultimately necessary for the development of community based conservation and management programs, an applied objective of my research. Thus, by collecting general socioeconomic data (e.g. household size, education level of head of household, primary income source, number of annual trips to the city, connections to markets through sales/trade, level of agricultural practice and wealth) in addition to specific turtle consumption data I hope to provide a detailed picture of two communities (see Appendix C for specific questions).

PAGE 61

CHAPTER 4 EFFECTS OF HUMAN SETTLEMENT ON THE NUMBER AND SIZE DISTRIBUTION OF TURTLES Introduction Assessment of turtle abundance and population structure has implications for the future of these wildlife resources and the human communities that depend on them. In areas where wildlife is harvested, monitoring harvest impacts is critical. In eastern lowland Bolivia such impact monitoring is necessary to determine the effects of harvest on abundance and population structure. This information can help determine whether hunting pressure poses a threat to turtle populations and it is an important precursor to conservation and management programs. This information was collected at three sitesPiso Firme, Remanso and Mangabalitoto compare hunted and less-hunted sites. Study Area Three 40-km transects were established along two rivers in lowland eastern Bolivia. Two of these transects, Piso Firme and Remanso, straddle human communities. The third transect, Mangabalito, is located adjacent to an abandoned human community. Piso Firme and Remanso are located in the buffer zone of the Park. Piso Firme lies on the Paragu River in the Department of Santa Cruz while Remanso lies on the Itenz River in the Department of Beni, approximately 30 km northwest of Piso Firme. Mangabalito, also on the Itenz River, is approximately 150 km upriver from Remanso in the Department of Beni. The central docks in Piso Firme and Remanso served as the middle points in the 40-km stretch of river used to collect the turtle abundance data. 47

PAGE 62

48 Piso Firme and Remanso were selected because: 1) both are on rivers within the ranges of P. unifilis and P. expansa; 2) P. expansa is presently found in the Itenz River and, fishers report that the species was formerly abundant in the Paragu River; 3) I determined in preliminary fieldwork in August 1999 that residents in these communities eat turtle meat; and, 4) the leaders of each community were willing to allow me to conduct my research in their respective community. Mangabalito, with no village within 50 km, was selected as the less-hunted site in the Park. When the Park was expanded to its present extent in 1996 the community living at Mangabalito abandoned the site (Juan de Dios Pea, pers. comm. 2000). While the community of Mangabalito was smaller than that of either Piso Firme or Remanso, approximately 12 families lived here (Juan de Dos Pea, personal communication 2001), the fact that people lived there four years prior to my study make this an interesting site for observing the potential response of a turtle population to reduced hunting pressure. Mangabalito is upriver from Remanso, on the Itenz River, at 13.78 S latitude and 60.55 W longitude. The only remaining signs of the community that disbanded in 1996 are four dilapidated houses and a few hectares of fruit orchards. No livestock was seen in the area during the time of the surveys. All three sites have some level of institutional presence dedicated to the conservation of wildlife. There is a Park camp located in Piso Firme, which means there is almost constant contact with the Park via Park guards who live there. Three of the Park guards are from Piso Firme, although they are not always based in this camp. Mangabalito is also the site of a Park camp, but it is not occupied on a regular basis anymore. In addition to being occasionally patrolled by Park guards, the Brazilian

PAGE 63

49 Institute for the Environment and Renewable Natural Resources (IBAMA) occasionally patrols at Mangabalito and even less frequently at Remanso. IBAMA oversees the sport fishers and local residents who base their activities out of Cabixe, Brasil, which is 10 km downriver from the endpoint of the Mangabalito transect. Remanso lies on the Itenz/Guapor River, which forms part of the natural border between Bolivia and Brazil, and is also patrolled by IBAMA as well as Park guards. During my study, relations between residents of Remanso and the Park were not as well developed nor accepted as in Piso Firme. There are continued efforts on the part of the Park to involve residents of Remanso in the implementation of the Park. Both Bolivian and Brazilian fishermen report that IBAMA agents issue harsh sanctions to people found with turtles or turtle eggs in their boats. Because the Itenz/Guapor River is an international body of water, Brazilian and Bolivian agents share authority. At all three sites there is patrolling by Park guards, although it was highly erratic during the year of my study due to various factorslack of guards (the Park employs approximately 25 guards for its 1.5 million hectare extent), gasoline, or working boats. Survey Procedure Turtles were counted using a basking-observation technique. This technique is non-invasive, can yield large samples, and produces less variability in data compared to hand capture and hoop net capture techniques (Koper and Brooks 1998). Among these three techniques, results from basking surveys provide more accurate population estimates although with more variability (from underestimation by 32% to overestimation by 13%) in relation to true population size (Koper and Brooks 1998). Basking surveys also result in lower heterogeneity of capture for the sexes than hand capture, but more

PAGE 64

50 accurate estimates of female populations than male, suggesting that basking surveys may offer a relatively unbiased form of estimating population size and obtaining large samples over a short time (Koper and Brooks 1998). The basking technique is limited by curves in the river and angles of basking logs that prevent all turtles from being spotted before they dispersed as the boat approached. I conducted all surveys and alternated the start side of the river on each run, so bias should be low or consistent and the abundance counts for the three areas comparable to each other (as suggested in Caughley 1977). Basking counts were chosen as the most appropriate means of collecting abundance counts. The technique as applied here is simple, requiring a boat, a global positioning system (GPS) unit, observation and notation. With these tools and techniques, visible turtles were counted along 40-km stretches of the Paragu and Itenz Rivers. Transects were set up along 40 km of river length traveling down the middle of the river. The 40-km length was used because fishermen indicated they would travel more than 10 km for fishing/hunting trips of more than one day and 40 km was the maximum physical distance that could be covered in the four and one half-hour period during which surveys were conducted. The window of survey time (11:00 a.m. to 3:30 p.m.) was delimited by the peak hours of sunshine, a significant factor influencing turtle basking (Koper and Brooks 1998). See Figure 4-1 for a photograph of basking turtles. I conducted the basking surveys during the dry season, between July and October 2000. During the dry season there are more basking spots available along river edges because the river is low and turtles are seen along the main river channel. During the wet season turtles more commonly frequent bays and flooded forests adjacent to rivers (Soini and Cppula 1995) and are rarely seen along the main river channels. The two-way surveys

PAGE 65

51 were conducted from an aluminum johnboat with a 25-horsepower outboard motor, traveling at an average speed of 18 to 24 km/hour. An assistant drove the boat approximately five meters from the riverbank. Waiting periods of at least 30 minutes were observed a few kilometers beyond the end point of each survey to minimize disturbance in the count on the opposite side of the river. Turtles were observed basking in areas we had passed by on the first run after thirty minutes, so this was presumed a sufficient waiting period. The side of the river and direction of travel were alternated between each journey. Side of the river was alternated for surveys conducted at the less-hunted site, Mangabalito, since all surveys were initiated from the furthest point upstream. Figure 4-2 shows a schematic drawing of transects. Figure 4-1. Podocnemis unifilis turtles basking along the Itenz River (K. Conway). All basking turtles and turtles seen in the water were counted. For each turtle observed the one-kilometer segment in which it was seen, according to the GPS, and its size class were recorded.

PAGE 66

52 km 31-40 km 21-30 km 11-20 km 1-10 Far Close community Close Far Downstream Upstream Figure 4-2. Schematic diagram of the 40-km transects illustrating location of Piso Firme and Remanso at the center of the 40-km transects and categorization of km into distance categories. I divided transects into two distance categories in relation to the communitiesfar and close. Far segments are those more than 10 km upstream or downstream from the center point while close segments are those within 10 km of the transect center. This distance represents a cutoff point for one-day fishing trips according to local fishers and was the main criterion in making this delimitation. In an Amazon Basin-wide study of hunting pressure, Peres and Lake (2003) found a 9-km threshold for hunting activity around villages, which further suggests this approximate distance as a division. I conducted 30 surveys in the period from August 14 to October 3, 2000. The sum of these surveys provided eight complete counts of turtles in 40-km transects at each site. It was not always possible to survey the full 40-km transect on each river within each assigned survey day due to weather, logistic constraints, equipment failure, or assistant availability. Therefore survey coverage was not identical in each transect but section results were combined among surveys conducted on successive days to provide effectively equal effort in all sections over the course of the study. Counts of turtles observed in each kilometer section were summed over 10-km intervals and the resultant turtle counts were compared among the different rivers and between the close (within 10 km of community/center point) and far (>10 km from community/center point) distance categories.

PAGE 67

53 Size classes for my study were determined based on a combination of research on both P. unifilis and P. expansa done on the Pacaya River in Peru by Soini (1995a and 1995b); the Maniqui River in Bolivia by Quiroga Vera (2000) and Knothe and Mhlteich (1996); near Iquitos, Peru by Fachn Tern et al. (1992); and, life history data from Pritchard and Trebbau (1984). I set the size classes for my study to reflect the larger measurements of P. unifilis cited by Soini (1995a) and Pritchard and Trebbau (1984) because I anticipated seeing more of these turtles due to their greater abundance compared to P. expansa (Pritchard and Trebbau 1984). Turtles in my study were classified into three size classes. These classes are: < 20 cm (small), 20 to 35 cm (medium), and > 35 cm (large). The turtles for which size could not be distinguished were classified as unknown. For P. unifilis, the size classes assigned in my study apply as follows: small includes juveniles of both sexes; medium includes mature males and immature females; and, large includes mature females. For P. expansa, the size classes apply as follows: small includes juveniles of both sexes; medium includes immature males and females; and, large includes adult males, immature and mature females. In addition to the size class data collected on turtles in transects, there were several opportunities to take measurements of captured specimens. These measurements combined with information from the literature helped in determining the size classification used here. Habitat Description In addition to the abundance and size class data collected, each km of transect was described using a 13-category classification developed in the field (Appendix B). This classification system was determined based on visual observations made over the course of abundance counts and quantified near the end of the basking season (September 2000).

PAGE 68

54 Once the categories were defined, I traveled along transects in my boat, driven by my assistant, and quantified the meters of shoreline of the categories in each km. Results Basic descriptive statistics of the numbers of turtles counted at the three study sites are presented in Table 4-1 and Figures 4-3 to 4-5. The numbers indicate a disparity in numbers of turtles observed along the 40-km transects, with fewer turtles seen at the two sites where there are human communities, Piso Firme and Remanso. This disparity is echoed visually in the graphical presentation of this data, Figures 4-6 and 4-7, which illustrate the average number of turtles seen in each kilometer at each of the three sites over the sampling period with one standard deviation bars indicated. There is a clear decline in the number of turtles around the center point of the two human-inhabited transects. Table 4-1. Descriptive statistics for the turtles counted along the three 40-km transects (Piso Firme, Remanso and Mangabalito). Site N Minimum # turtles/transect Maximum # turtles/transect Mean Turtles/km Std Dev Piso Firme 387 12 77 1 21.08 Remanso 469 18 98 2 21.93 Mangabalito 2889 181 501 9 102.45 Mean Numbers of Turtles at Different Distances from Transect Center I compared the mean number of turtles observed in 10-km segments between and within sites. The 10 km immediately upstream and downstream of the center point of the transects were designated close while the 10 km furthest upstream and furthest downstream from the communities were designated far based on local turtling and fishing practices as well as data from other studies suggesting this is a threshold near which hunting intensity changes. Turtle counts in the 10-km segments used for analyses

PAGE 69

55 are graphed in Figures 4-3 to 4.5. Descriptive statistics and results of the analyses are found in Tables 4-2 to 4-5 below. Table 4-2. Kruskal-Wallis test statistics comparing median number of turtles in segments within (Close) and beyond (Far) 10 km of the center point between the three sites. These results indicate that the difference is not random. Far Close N 24 24 Median 47.50 27.0 2 16.99 16.90 df 2 2 p-value <<0.001 <<0.001 The results in Table 4-2 indicate that the larger median number of turtles seen in far km is not due to chance. Further examination of the data included looking at each site individually. The results of the Mann-Whitney comparison of medians within sites (Table 4-3) indicate that the differences seen in close vs. far segments are significant at Remanso only. The median values indicate that more turtles were observed at distances beyond 10 km from the center. On a site by site basis, the difference in number of turtles seen in close km is significantly smaller at Remanso only. This finding suggests that hunting pressure is stronger at Remanso than Piso Firme. Mangabalito, which has no constant human presence, has more turtles overall and does not exhibit a pattern like Remanso. This may be indicative of a more robust turtle population. Table 4-3. Mann-Whitney test statistics for the comparison of close vs.far segments within sites. Results indicate a significantly higher median number of turtles farther from the center of the Remanso transect only. Site Distance Mean Std Dev U p-value Piso Firme Far 47.75 Close 49.00 11.75 30 0.88 Remanso Far 95.50 Close 21.75 86.00 5 0.003 Mangabalito Far 375.00 Close 347.25 52.21 28 0.72

PAGE 70

56 Table 4-4. Frequencies of counts greater than or less than/equal to median values within 10 km (Close) or beyond (Far) transect center points of the three sites. Median # Turtles Piso Firme Remanso Mangabalito Far 47.50 > Median 1 3 8 <= Median 7 5 0 Close 27.00 > Median 3 0 8 <= Median 5 8 0 The results of the median frequency test show patterns at the three sites. At Piso Firme only one median count was above the median value of 47.50 for far kilometers while seven were below it and three of eight sample counts were greater than the median calculated value of 27.00 turtles for close segments (Table 4-4). These values indicate that in 12 of 16 cases the number of turtles observed was less than expected, suggesting a small population. From the graph of turtles counted at Piso Firme it is evident that there were more turtles downstream than upstream (Figure 4-3). This is possibly a combined result of hunting pressure and ecological factors. The observations are similar at Remanso, where the number of turtles counted in km farther from the community was above the median value in three of the five samples but never at the close segments. Again, this relationship is illustrated in the graph of turtles counted, where there is a clear dip in the numbers around the center of the transect (Figure 4-4). In stark contrast, at all distances for all samples at Mangabalito the mean number of turtles observed was above the median. The graphs of numbers of turtles observed (Figure 4-5) indicate that more turtles were observed downstream at this site. Clearly the turtle population at Mangabalito is greater than those at Piso Firme and Remanso. Finally, to observe differences between hunted and less hunted sites, Mann-Whitney tests were used (Tables 4-5 and 4-6). It is worth noting again that Mangabalito

PAGE 71

57 and Remanso are on the same river, the Itenz, which makes a stronger comparison between these two sites and the basis for testing research hypothesis two. 05101520253035404550KM 1-10KM 11-20KM 21-30KM 31-40KM Section# Turtles Counted Piso Firme Figure 4-3. Number of turtles counted at Piso Firme, along the Paragu River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect. 01020304050KM 1-10KM 11-20KM 21-30KM 31-40KM Section# Turtles Counted Remanso Figure 4-4. Number of turtles counted at Remanso, along the Itenz River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect. The results indicate that the difference in mean numbers of turtles counted at Piso Firme is less than that at Mangabalito and the difference is not due to chance alone. The

PAGE 72

58 difference between Remanso and Mangabalito also indicate that the difference in mean numbers of turtles counted at Remanso is less than that at Mangabalito. Hypothesis two states that there will be more turtles at Mangabalito (suggesting less hunting pressure) than at Remanso (suggesting more hunting pressure). Based on the results of the Mann-Whitney test of difference, this is the case. 020406080100120140160180200KM 1-10KM 11-20KM 21-30KM 31-40Km Section# Turtles Counted center Figure 4-5. Number of turtles counted at Mangabalito, along the Itenz River, shown in 10-km segments. Each line represents a complete survey along the 40-km transect. Table 4-5. Test statistics for test of difference between numbers of turtles counted at close vs. far segments from center point of three transects. Far Close Mann-Whitney U 0.00 0.00 p-value <<0.001 <<0.001 Table 4-6. Comparison of differences between sites in numbers of turtles counted at close vs. far segments from center points. Far Close Comparison Mann-Whitney U p-value Mann-Whitney U p-value Piso Firme vs. Mangabalito 0.00 <<0.001 0.00 <<0.001 Remanso vs. Mangabalito 1.0 <<0.001 0.00 <<0.001 Piso Firme vs. Remanso 12 0.04 14.5 0.07

PAGE 73

-3-11357911131516111621263136KM# of Turtles Piso Firme Remanso Piso Firme & Remanso 59 Figure 4-6. Average number of turtles seen in each km over eight samples at populated sites (Piso Firme and Remanso) with one standard deviation indicated by bars on each kilometer and the communities at KM 20. Note the different Y-axis scale from Figure 4-5, which illustrates the unpopulated site.

PAGE 74

-551525354555657516111621263136KM # of Turtles 60 Figure 4-7. Average number of turtles seen in each km over eight samples at the unpopulated site (Mangabalito) with one standard deviation indicated by bars on each kilometer. Note the different Y-axis scale on Figure 4-4, which illustrates the populated sites.

PAGE 75

61 River Edge Habitat Characteristics I describe the similarities and differences among the three sites based on the habitat type classification I developed because it provides additional comparative information between sites (Figure 4-8). I provide this information in addition to the comparison of turtle numbers between sites based mainly on human presence as an important consideration in factors leading to differences seen among the sites. There are more steep barren cliffs at Remanso and Mangabalito than at Piso Firme. This is likely due to the higher order of these rivers as well as geological differences. These physical characteristics may cause a concentration of turtles in the main river channel, which may contribute to the larger number of turtles seen in the Itenz vs. the Paragu Rivers. A related characteristic is depth. In some parts of the Paragu River the bottom at center is less than 1 m, while on the Itenz River the bottom was rarely to never less than 1 m deep. This characteristic may also contribute to the Itenz River being more suitable habitat for the turtles and could partially explain greater numbers seen on this river. There are more steep cliffs with aquatic vegetation (that constitutes some of the turtles diets (Bruno Coca 1999)) and sticks (important basking platforms) at Mangabalito than either Remanso or Piso Firme. This characteristic may partially explain greater abundance at Mangabalito due to the concentration of turtles in the main channel, more foraging opportunities and basking platforms. There are a few more beaches at Remanso and Mangabalito than at Piso Firme. One possibility is that this may reflect more nesting habitat, particularly for P. expansa, which nests exclusively on sandy beaches. This characteristic may have less effect

PAGE 76

62 A050010001500200025003000Steep barren cliffSteep cliff w/ ...Steep cliff w/ s...TaropeCanuelaMixed aquatic vegHigh sandy beachHigh veg beachLow veg beachMuddy flatBayLow sandy beachRocksHabitat TypeTotal meters in Transect B050010001500200025003000Steep barren cliffSteep cliff w/ aqu...Steep cliff w/ sticksTaropeCanuelaMixed aquatic vegHigh sandy beachHigh veg beachLow veg beachMuddy flatBayLow sandy beachRocksHabitat TypeTotal meters in Transect C050010001500200025003000Steep barren cliffSteep cliff w/ aqu...Steep cliff w/ sticksTaropeCanuelaMixed aquatic vegHigh sandy beachHigh veg beachLow veg beachMuddy flatBayLow sandy beachRocksHabitat TypeTotal meters in Transect Figure 4-8. Meters of each habitat type at the three sites, shown individually by site. A) Piso Firme. B) Remanso. C) Mangabalito.

PAGE 77

63 on the abundance of P. unifilis. The possibility of more nesting habitat may contribute to Mangabalito being a preferable site in comparison to Piso Firme or Remanso. In summary, while there are subtle differences in river edge habitat between sites, nothing stands out as dramatically different. The proportions of each vegetation type are relatively similar across transects. This lends further strength to the suggestion that the main factor leading to difference in abundance between the sites is human presence. Size Differences among Sites The mean carapace length, measured along the straight line from anterior to posterior edges, of 10 P. unifilis females measured from the Paragu and Itenz Rivers was 38 cm, with a range of 19 to 46 cm (sd=7.5 cm). Of these females, at least four had eggs (the smallest egg-bearing turtle measured 36 cm length), which indicates that at least four were reproductively mature. The mean carapace length of P. expansa females measured was 59 cm (sd=10.7 cm, n=8) and 39 cm for the males (sd=4.9 cm, n=7). Of the P. expansa specimens measured, one (length=70 cm) had egg yolks. Soini (1995) gives both the straight line and curved measurements of one female P. unifilis carapace as 47.7 cm and 50.5 cm, respectively, which yields a ratio of 1.05. Extrapolating from this datum, the range of curved length measurements for the turtles measured in my study may be calculated as 20 cm to 48 cm, which suggests that the start of the large size class category at > 35 cm is a conservative estimate. Based on measurements taken in the field that included a 36-cm long P. unifilis female with eggs, combined with data from the literature (Pritchard and Trebbau 1984; Soini 1995a), the large size class should consist solely of reproductively mature females and males. This may negatively affect the number of nests and the number of eggs laid since smaller size in P. unifilis is correlated with smaller clutch size (Vanzolini 1977). Figure 4-9 illustrates

PAGE 78

64 ASmall <20 cmMedium 20-35cmLarge >35cmUnknown050100150200250300Number of Turtles 37% 52%7% 4% BSmall<20 cmMedium20-35 cmLarge>35 cmUnknown050100150200250300350400Number of Turtles 16% 63% 19%2% CSmall<20 cmMedium20-35 cmLarge>35 cmUnknown02004006008001000120014001600Number of Turtles .1% 48% 32%20% Figure 4-9. Number of turtles in each size class observed with one standard deviation of the number of turtles in each size class at each site and relative percent of the turtles at the site in each of the four size classes noted. Note the different scales between sites. A) Piso Firme B) Remanso C) Mangabalito

PAGE 79

65 that the medium size class is the largest of the four size classes, at all three sites. Also, the large size class is considerably greater at Mangabalito and is a greater proportion of the overall sample at this site in comparison with the others. The small class is very abundant at Piso Firme. The standard deviation for the medium size class is the greatest (sd=74.45) (the other standard deviations are: small (sd=34.02), large (sd=58.8) and unknown (sd=1.67)). There is a trend towards smaller turtles near Piso Firme, but this does not exclude the presence of P. expansa in the Paragu River, because fishermen reported catching two individuals, and I observed at least one adult female in transect counts. Two possible explanations for the lack of sightings of more P. expansa are: 1) they are not as commonly seen basking (local fishermen and R. Vogt pers. comm. 2000); or, 2) they are not necessarily resident in this river but are there seasonally in search of nesting beaches. Of these explanations, the former has been cited in the literature (Pritchard and Trebbau 1984) and noted by biologists working in the Amazon (R. Vogt pers. comm. 2000). According to local residents, in the past P. expansa was commonly found on the Paragu River. Again, this information suggests reduced populations near humans. The turtles observed at Remanso are on the whole larger than those on the Paragu, with a 16:63:19 ratio of percentages of small:medium:large. Based on the large size class, which includes reproductively mature P. unifilis females, there is a larger portion of potentially reproducing females at Remanso than Piso Firme. The large turtles observed at Mangabalito (32% n=926) probably include reproductively mature females of both species and is by far the largest percentage among the three sites. Assuming the large category is the only one that contains reproductively mature females, less than seven

PAGE 80

66 percent of the turtles seen at Piso Firme, less than 19% at Remanso and less than 32% of the turtles at Mangabalito are reproductively mature. For long-lived species with delayed sexual maturity, the ratio of mature to immature members is low, and these numbers may be representative of this. But further counts must be conducted and compared over time to discern trends. Due to differences in sizes of the two species and the size class categories applied, the majority of the P. expansa specimens seen are immature, so discussion of population structure will be restricted to P. unifilis except when stated otherwise. Comparison of Size of Turtles at Different Distances from Transect Center Table 4-7 indicates the proportions of turtles of all sizes observed closer and farther from transect center points. These numbers suggest trends towards larger turtles farther from transect center points. Table 4-7. Proportions of turtle size classes seen within sites at far and close categories with averages over all three sites and at Mangabalito, the less-hunted site. Proportions of Size Classes Distance Small Medium Large Unknown Piso Firme far 37% 47% 10% 6% close 38% 56% 3% 3% Remanso far 15% 63% 19% 2% close 17% 62% 18% 2% Mangabalito far 17% 44% 39% 0% close 23% 52% 25% 0% Size Class Average 25% 54% 19% 2% Mangabalito Average 20% 48% 32% 0% Visual inspection of the proportions of the size categories observed at each site (Figure 4-10) confirms the differences in size class distributions among the three sites shown in Figure 4-9. Examining the size class proportions between close and far sections at each site, it appears that there is a difference in the number of large turtles at close

PAGE 81

67 categories at both Piso and Mangabalito, but not at Remanso. The proportion of small individuals also differs somewhat between close and far categories at Mangabalito. 10% 18% 19% 25% 39% 56% 47% 62% 63% 52% 44% 38% 37% 17% 15% 23% 17%6%2%2%0%0% 3%3%0%10%20%30%40%50%60%70%80%90%100%Piso Firme closePiso Firme farRemanso closeRemanso farMangabalito closeMangabalito farSite and DistanceProportion in Size Class unknown small medium large Figure 4-10. Proportions of turtles of different size classes by distance categories at the three sites. The median number of turtles was greater at further segments when the three sites were combined for analysis (Table 4-8). Analyses of these differences by Kruskal Wallis test (Tables 4-8 and 4-9) indicate that these differences in median size are significantly different (p=0.07 for far and p=0.09 for close). If the Mangabalito site, which has experienced less hunting pressure since 1996, can be considered as a recovered or partly recovered location with normal size class distributions, then it is apparent that turtle size class can be quite variable in the absence of human predation. While the pattern at Piso Firme conforms to expectations that sites nearer to the community would have fewer large specimens, results at Remanso are not as dramatically low. The very low proportion of large turtles near Piso Firme (3%) compared to the levels at Mangabalito (25 to 39%) suggests depletion of this size class while the proportion of

PAGE 82

68 large turtles near Remanso (18%) suggests moderate depletion. However additional data collection is needed to confirm whether these interpretations are robust. Table 4-8. Kruskal-Wallis test statistics comparing all four size classes at close vs. far segments from center point at all three sites. Results indicate significant differences between size classes in close vs. far segments overall. Far Close N 12 12 Median 72.50 35 2 7.21 6.57 df 3 3 p-value 0.07 0.09 Table 4-9. Frequencies of size classes counted in relationship to medians for far vs. close transect segments when all three sites are combined. Size Class Median # Turtles Small Medium Large Unknown Far 72.50 > Median 1 3 2 0 <= Median 2 0 1 3 Close 35.00 > Median 2 3 1 0 <= Median 1 0 2 3 Examining the median values (Table 4-9) reveals that: 1) small turtles are more abundant in closer km; 2) there are more medium turtles at both distances; 3) large turtles are more common in far km; and, 4) unknown size turtles are less frequent at both distances. These trends suggest that the medium size is the most common size class among all three sites. This size class is not indicative of sexually mature populations because the medium size class only includes mature males. My data suggest that hunting pressure is negatively affecting the longevity of the populations as posited in hypothesis three. A Gamma test was applied to compare the association between size and distance at hunted and less hunted sites for turtles of known size classes (Table 4-10). The greater

PAGE 83

69 number of discordant pairs indicates that there is a negative association between size and distance. Table 4-10. Gamma test for known size classes at different distances compared between hunted and less hunted sites indicating a negative association between turtle size at close vs. far distances in hunted and less-hunted areas. Size Class and Number of Turtles Site and Distance Small Medium Large Piso Firme & Remanso close 89 164 22 Mangabalito close 322 718 347 Piso Firme & Remanso close 130 331 93 Mangabalito far 258 662 579 Concordant Pairs 1,098,451 Discordant Pairs 1,781,691 Gamma -0.24 When a comparison of turtles is made on the basis of increasing hunting pressure and known size, there is a negative association (Table 4-10). There is a 76% probability that a decrease in turtle size is associated with proximity to humans. Humans appear to have a negative impact on the sizes of turtles seen. Discussion Differences in Turtle Abundance in Relationship to Humans The abundance data presented in this chapter suggest that human activities, including predation, may play roles in reduction of turtle numbers. For research hypothesis one (H1: turtles will be less abundant closer to human communities) there was statistically significant evidence that indicated differences in turtle abundance closer to humans vs. farther away from humans with fewer turtles counted closer to humans at the hunted sites. For research hypothesis two (H2: there will be more turtles at Mangabalito (less hunting pressure) than at Remanso (more hunting pressure)) there was also evidence for differences being due to proximity to humans at Remanso.

PAGE 84

70 If hunting is the main reason for the distance-abundance relationships, my data suggest that current practices of turtle hunting depress turtle abundance. The fewer turtles seen close to the community at Piso Firme and Remanso may be a result of hunting pressure. The lower turtle counts on the Paragu River, around Piso Firme, suggest that the effect of humans on turtle abundance may be greater at this site. This may be a result of the smaller river and hunting pressure. The contrast of the two human-inhabited sites with uninhabited Mangabalito, which has greater numbers of turtles at all distances in 100% of the samples, suggests even more clearly that the cause of the difference is human activity. A comparison of the two sites on the Itenz River, Remanso and Mangabalito, provides further support for the hypothesis that hunting pressure is the main difference between these sites that leads to fewer turtles close to humans because the confounding factor of different rivers is not present. Redford and Robinson (1985) showed that depletion of game species within the catchment area of villages led to decline of game. Peres and Lake (2003) quantified a correlation between proximity to humans and reduced prey abundance in the Amazon Basin, with 9 km being the threshold beyond which hunting did not negatively affect prey abundance. My use of a 10-km threshold from humans was based on information from fishers that they do not typically travel more than that distance on day-long fishing trips. Thus, my results concur with Peres and Lakes large-scale work and the conclusion of hunting pressure within a catchment area leading to degradation of natural resources closer to humans that agrees with Redford and Robinson. Like Bodmer et al. (1997) my results show that human activities, which may include hunting pressure, can negatively influence game species, in this case river turtles.

PAGE 85

71 Differences in Sizes in Relationship to Humans The third research hypothesis addressed the size of turtles expected nearer to humans vs. farther (H3: there will be fewer large turtles closer to human communities than farther away). Distance in relationship to humans may play a role in the size of turtles observed, most likely because hunters preferentially take larger turtles. My findings indicate no significant differences in sizes of turtles between sites only a trend towards smaller turtles more frequently observed closer to human communities and larger turtles farther away. The difference in population structure suggested by the analyses of all three sites indicates more mature populations farther from the communities than closer, similar to the cases reported by Robinson and Bodmer (1999). When the human-inhabited and presumably heavier hunted sites (Piso Firme and Remanso) are combined and compared to the uninhabited and less-hunted site (Mangabalito) there is a negative association (=-0.24). Larger turtles are less abundant closer to Piso Firme and Remanso than Mangabalito. This suggests that the larger turtles are the first to be removed nearer to humans as a result of hunting pressure. The number of different-size turtles observed at the three sites indicates many more small and medium than large turtles at Piso Firme and more medium than large turtles at Remanso and Mangabalito. The most mature population among the three sites is Mangabalito, where 32% of the turtles observed were large. According to Congdon et al. (1993) juvenile loss rates must be extremely low in long-lived organisms for populations to remain stable. If small turtles bask at the same rate as medium and large turtles, higher proportions of this size class than were noted would be expected in populations with high recruitment. In terms of juvenile loss rates, Piso Firme shows the lowest proportions (37% of the turtles seen were small vs. 16% at Remanso and 20% at Mangabalito), which

PAGE 86

72 may indicate that juveniles are surviving longer at this site or recruitment is greater. However, it is not certain whether small turtles have the same basking behavior as larger turtles, so no definite conclusions about survivorship may be reached based on this basking count data alone. Trends in the medium size class (which includes juvenile females) must be observed over time to determine what is happening with the juvenile portion of these populations, because this will have an effect on the viability of these populations of P. unifilis. My data show a high percentage of medium turtles at Remanso (63%) but not many of these appear to reach maturity since only 19% of those seen were large. Signs of hunting pressure are particularly strong at Remanso. Piso Firme also appears to have hunting pressure, but it is possible, considering that abundance and size class data indicate fewer differences and fishers report catching fewer turtles from this river, that hunting pressure is lower than at Remanso and turtles are beginning to recuperate. Noting the trend towards larger turtles seen at the Mangabalito site it is worth considering that humans may have a negative effect on the size of turtles in demographic terms. Conclusion My results show that there are fewer and smaller turtles close to human communities. This may be the result of hunting pressure. If this is the case, removal of turtles and their eggs is harmful and is leading to decreased river turtle populations in this area of Bolivia. The observation of more and larger turtles near the recently abandoned site of Mangabalito compared to both of the two currently hunted sites may be the result of turtle populations having recovered from hunting pressure here. If there has been a recovery of the population at Mangabalito in the absence of humans for five years,

PAGE 87

73 currently reduced hunting pressure may be sufficient to allow turtle recovery. The implication that recovery is possible is exciting for conservation efforts. Because turtle size is an indication of population maturity, my findings suggest negative potential impacts of turtle hunting. If a population consists mainly of small and medium turtles the implication is that large reproductively-mature turtles and long-term survival of the populations may be threatened. This is especially true at Remanso, where abundance appears to be strongly affected by hunting pressure. My data show abundance, and the time to reproductive maturity is uncertain (estimates range from 5-10 years (Pritchard and Trebbau 1984)), so the persistence of the turtle populations studied at these sites in lowland eastern Bolivia cannot be estimated. Both P. unifilis and P. expansa turtles are threatened and in danger of being hunted to local extirpation in much of their ranges as a result of hunting pressure. My data suggest that eastern lowland Bolivia is no exception to this trend. However, if hunting pressure on the turtles is reduced, there is reason to believe abundance can increase, restoring two valuable ecosystem and social system components and providing humans with continued access to this important food resource.

PAGE 88

CHAPTER 5 SOCIAL STRATIFICATION, MARKET INTEGRATION, AND HOUSEHOLD CONSUMPTION Objectives and Hypotheses The objective of this portion of my research was to collect data on selected socioeconomic characteristics and degree of market integration and analyze their effects on Podocnemis turtle use in lowland Bolivia. Data were collected on assumed measures of market integration and turtle consumption patterns (described below). Von Thnens intensity theory ([1826] 1966) described a relationship between distance to market and intensity of production. Building from this theory I made a comparison between two ribereo communities. My designation was that Remanso was more market integrated because it is 1) closer to a marketPimenteiras, Brazil; 2) lies on an international, highly traveled river; and, 3) offers greater cash-generating job possibilities for residents. Based on these qualities, the cost of participation in a market is less for residents of Remanso than Piso Firme, leading to its a priori classification as more market integrated. Various socioeconomic data collected served as dependent variables for the market integration analysis. The variables tested include: tripsthe number of times per year a household member traveled out of the community (an indication of outside ties, the majority are at least partially market-related); chacohow much land they cultivate (an indication of subsistence livelihood and therefore not market-related); educationyears of education of the head of household (indication of greater access to cash-earning 74

PAGE 89

75 jobs); sizenumber of people in the household (indicator of a greater need for access to market or cash); and, jobtype of job (certain employment offers greater cash-earning possibilities). The indirect characteristics reported were chosen because they appear to effect natural resource use in the context of market integration as demonstrated through linkages via trips out of the community. Increased harvest of natural resources attributed to market integration has been described by Godoy (2001), Redford and Robinson (1985), Sierra et al. (1999), and Von Thnen ([1826] 1966). In contrast, other studies cite positive relationships between market integration and natural resource conservation (Agrawal 2001; Agrawal and Yadama 1997; Godoy 2001; Schweik et al. 2003; Tang and Tang 2001). My objective in testing the variables I tested was to better understand the relationships between market and subsistence livelihoods and turtle consumption in Piso Firme and Remanso. The two communities were also compared in terms of household turtle consumption. Turtle consumption variables tested include: eatnumber of times the previous week turtle was consumed in the household; send turtleswhether someone in the household sends turtles out on the weekly bus; send eggswhether someone in the household sends turtle eggs out on the weekly bus; hours turtlinghours per household per week dedicated to turtling; and, preferpreference between turtle, other meat or cash. Eat and hours turtling were considered subsistence forms of consumption because they do not require cash. Prefer (turtle) was considered subsistence consumption because it suggests greater usefulness for the meat over the cash equivalent. Send turtles, send eggs were considered market consumption because

PAGE 90

76 they involve consumption outside of the community, which has been designated a non-subsistence use by the Park. Prefer (cash) was considered market consumption because it was thought to indicate greater importance of cash over turtle, which would be more likely in a more market-integrated household. The hypotheses tested are: H4: There is greater market consumption of turtle protein in Remanso than Piso Firme and greater subsistence consumption of turtle protein in Piso Firme than in Remanso. H5: There is a negative relationship between wealth and turtle protein sale/trade such that as personal wealth increases, the consumption of turtle protein decreases. Personal wealth affects decisions about consumption. A deeper understanding of the relationship of these potentially interactive factors has implications for the future of the turtles as food sources, as incorporated into attitudes and a management plan. Tests of difference on these data examine differences and similarities of effects of wealth on turtle consumption. If community members are interested in developing formal rules of use for the turtle resource, these socioeconomic analyses illustrate some confounding and unrelated characteristics they may chose to consider further. Methods Sample Size Calculation The unit of analysis is the household. A household is defined as the group of individuals who lives together, shares food resources and eats together. Individuals over the age of sixteen are considered adults because it is common for children to move out of their parents home and start their own families at this age. A systematic random sample of households was the basis for data collection and hypothesis testing.

PAGE 91

77 To determine the number of households per community I needed to sample, I began with censuses. In June and July 2000 I went door to door in each community, accompanied by a member of the community, and asked an adult of the household how many people lived in the house for more than six months of the year. I also asked for the breakdown of individuals living in the household under and over 16 years of age. The sample size was calculated based on these censuses. The following formula was used with the census information to determine the number of households required in the samples: 2 N P(1-P) C 2 (N-1) + 2 P(1-P) = Sample population Where 2 is the value for chi-square at the 95% confidence level (3.841) with one degree of freedom, N is the population size, P is the population parameter (0.5 was used in this calculation because it assumes the greatest amount of variability in a population and results in the largest sample population being calculated) and C is the confidence interval of five percent (Bernard 1995). The calculated required sample size of households was 56 for Piso Firme and 77 for Remanso. Interview Administration and Questions Sampling design was random and based on the censuses, with household head names alphabetized and interviews assigned using a random number chart. Data were collected using structured interviews administered from August 2000 through May 2001. The interview consisted of three sectionsdemographic, income and market access, and turtle-related questions (Appendix C). For the majority of interviews a community member accompanied me. In instances when an assistant was not available I conducted the interviews alone. The adult head of household of either sex who was available at the

PAGE 92

78 time of the visit was interviewed. Visits took place between 7:30 am and 6:30 pm, seven days of the week. In many cases both adult heads of household were interviewed. Wealth Ranking The wealth ranking of households in each community was conducted following Slocum et al. as a very effective way to determine socio-economic groupings in a community by measuring households (1995: 236-7). This exercise provided a valid indicator of relative wealth rankings in each community, from the perspective of knowledgeable informants (Adams et al. 1997; Takasaki et al. 2000) rather than based on my personal interpretation of wealth in Piso Firme and Remanso. The wealth-ranking exercise was completed by the recommended five people in each community (four individuals and one couple) who were chosen because they were long-time residents who knew all residents in the community, represented both sexes, had different educational backgrounds and were engaged in different livelihoods (Slocum et al. 1995). Responses were converted into one ranking scale for each community. Names of household heads from the censuses conducted in each community (June and July 2000) were written on cards. Participants classified all households in their community, relative to the others, in terms of personal wealth. Working with the participants, I read the names aloud and asked the participant to sort the cards into groups according to their criteria. After the first sort, names in each pile were read aloud again and participants made any changes they decided were necessary. Next, participants explained their criteria for placement in the various groups and elaborated common characteristics among group members (see Appendix D for a detailed description of the participants categorization criteria). Participants created three to six groups. The final score for each

PAGE 93

79 household was generated based on the number of groups each participant created and the total number of households in the community using the formula: households rank by participant n total # of groups assigned by participant n X total # of households in community = household score The five scores for each household were averaged by community. Ultimately, three wealth groups were deemed sufficient indicators of differences among wealth groups within each community. Each household was placed into a wealth group based on its average score. The wealth group assignments were the basis for testing hypothesis five. Definition of Market Goods and Market Integration For my study, market goods are those grown, hunted or fished and sold or used as cash. Goods that are grown, hunted or fished for household consumption or given to other households as gifts and not as payment for services are considered subsistence items. This definition is based mainly on the exchange of cash or goods with an established cash value in an attempt to address the conditions of rural communities where access to cash is limited and residents have developed systems of barter and exchange that do not require cash for acquisition of goods or services. Market influence, as it exists for residents of Piso Firme and Remanso, is based on growing access to cities and ensuing dependence on cash to acquire goods and services necessary to achieve a desired lifestyle coupled with more consistent and greater cash-generating activities outside the communities (e.g. work on Brazilian fishing boats, in the mine located south of Remanso, or skilled and unskilled labor in towns such as San Ignacio, Bolivia and Pimenteiras, Brazil). Access to cash and a need to leave the

PAGE 94

80 community to acquire it was noted by a number of people throughout my time in the field. This is particularly relevant to turtles, which are a local commodity that brings in a high-cash profit. Access to cash depends directly upon income source. Job categories used in the analyses are mutually exclusive except for State, which is a combination of teachers, doctors and nurses; and Park, which is a combination of Park guards and those who work for Fundacin Amigos de la Naturaleza, the nongovernmental organization that co-manages the Park with the Servicio Nacional de Areas Protegidas. These jobs were combined because of the overlap among entities that pay the salaries. The income-earning potential for jobs was determined through a combination of interviews with residents and personal observation of livelihood strategies that suggested greater outside linkages and cash generation as well as more frequent cash exchange afforded by certain jobs. Socioeconomic Characteristics and Their Implications for Livelihoods Some characteristics have built-in assumptions, which are described with the variable. The job that supplies the household with its primary source of livelihood has an influence on use of natural resources as well as access to cash. Jobs that are based on natural resource extraction may directly involve turtle harvest in the communities of Piso Firme and Remanso, e.g. fishers who dedicate a portion of their time to capturing turtles for household consumption or sale for cash to obtain other goods. This is especially important in Piso Firme, where job options that provide cash are limited but demands for cash are decidedly present. One measure of a households access to cash is suggested by its primary income source. When jobs are graphed in order of income-earning potential, it is evident that agriculture is the main source of income in both Piso

PAGE 95

81 Firme and Remanso (Figure 5-1). The mean for chaco (hectares of cultivated land) indicates that most households in both communities maintain a chaco ( x =0.80 ha sd=0.47 for Piso Firme and x =1.03 ha sd=3.17 for Remanso) (Table 5-2). It is also worth noting that a greater percent of the households in Remanso (48%) are supported by higher-income-earning jobs (day labor + park + state + store) than are in Piso Firme (38%). 051015202530AgricultureFishingCattleSawmillMiningDay LaborParkStateStoreNumber of Households Piso Firme Remanso Figure 5-1. Main source of household income, in increasing order of cash-generating potential in Piso Firme (N=55) and Remanso (N=71). Trips outside the communities is one measure of external linkages, which typically involve markets. First, leaving the communities requires cash to pay for transportation and lodging, although there are occasional opportunities for free logistics (travel in Park or private vehicles and lodging with family or friends). Secondly, I observed that most people returned to the community with goods from the city for household use or sale. Thus, the assumption with trips is that it is an indication of market integration due to the cash required for logistics and the opportunities presented to

PAGE 96

82 participate in the market. The motive for the majority of trips outside the communities was health-related (Table 5-1). Residents of Remanso travel more frequently and more commonly to Pimenteiras, Brazil, which is five hours upriver from Remanso and six hours upriver from Piso Firme. Piso Firmeos travel more often to San Ignacio, Bolivia, which is a 12 to 24 hour journey by road. These data suggest closer ties to Pimenteiras by residents of Remanso and to San Ignacio by residents of Piso Firme. The proximity of Pimenteiras to Remanso implies lower transport costs and greater likelihood of participation. Table 5-1. Motivation for trips taken out of communities to San Ignacio, Bolivia and Pimenteiras, Brazil by residents of Piso Firme and Remanso. Trips to San Ignacio Motivation for Trip Piso Firme Remanso Health 19 2 Work 2 3 Visit Someone 4 6 Make Purchases 3 0 Total 28 11 Trips to Pimenteiras Health 1 14 Work 0 2 Visit Someone 1 3 Make Purchases 0 3 Total 2 22 Mann-Whitney tests were used to compare socioeconomic characteristics between the two communities. Pearson Chi-square tests for difference between distributions of categorical characteristics that indirectly influence turtle consumption at the community level were computed (Table 5-2). Results are discussed below. Education exhibits a large variance among heads of households, with most residents having at least two years of education and head of household in Remanso averaging one additional year of education ( x =6.55) than in Piso Firme. Education indirectly influences turtle consumption because it is likely to effect employment, which

PAGE 97

83 Table 5-2. Descriptive and test statistics of socioeconomic characteristics that directly and indirectly influence turtle consumption detailed by community. Mann-Whitney test results for H4 indicate more turtle consumption in Piso Firme for all variables but prefer, for which Remanseos indicated greater preference for cash equivalent. Note: Definitions of variables, c refers to characteristics that have been categorized: Chaco refers to number of hectares of cultivated land a household maintains; Trips number of times someone from the household leaves the community for an urban destination during the year; Education is years of education of head of household; Hours fishing if someone in the household fishes, this refers to the hours per week that are dedicated to this activity; Size number of people in the household sharing food; Eat is the number of times the previous week that turtle was eaten in the household; Send turtle the number of times someone in the household sent turtle out on the bus in one year; Send eggs whether or not someone in the household sends turtle eggs out on the bus, 0=no, 1=yes; Hours turtling is the number of hours dedicated to catching turtles per week by someone in the household; Sell whether or not someone in the household sells turtles/eggs, 0=no, 1=yes; Prefer preference for turtle, or monetary equivalent of a P. unifilis (the sequence represents increasing gradations of market consumption) 1=turtle, 2=of equal value, 3=cash equivalent. s indicates subsistence consumption variable, m refers to market consumption variable. Significant results at 95% confidence level appear in boldface. PisoFirme Remanso Characteristic x Min Max StdDev n x Min Max Std Dev n TestStatistic p-value Indirect Influence Chaco 0.80 0 3 0.466 53 1.03 0 25 3.17 71 U=1385.50 0.008 Trips 2.13 0 6 1.77 52 4.10 0 24 2.61 70 U=1761.00 0.58 Education 5.55 0 13 3.48 51 6.55 0 17 4.34 66 U=1464.50 0.23 Hours fishing 9.17 0 10 12.32 55 6.42 0 18 8.06 69 U=1640.50 0.19 Size 6.05 2 15 2.67 55 5.8 1 15 2.82 71 U=1887 0.75 Direct Influence Eat(s) 1.38 0 7 1.52 50 0.87 0 4 0.88 71 U=1432.5 0.04 Send turtle (m) 0.65 0 6 1.36 48 0.13 0 4 0.62 69 U=1335.5 0.003 Send eggs (m) 8%=yes 34%=no 49 3%=yes 55%=no 69 2 =4.58 df=1 0.03 Hours turtling (s) 11.02 0 24 24.4 50 6.31 0 16 14.93 65 U=1321.50 0.03 Sell (m) 47%=yes 16%=no 20 16%=yes 22%=no 12 2 =3.54 df=1 0.06 Prefer (s, m) 20%=1 3%=2 18%=3 49 17%=1 6%=2 36%=3 70 2 =5.19 df=2 0.07

PAGE 98

84 in turn influences time spent directly producing food or cash, and the complimentary need for cash or food. Hours fishing indirectly influences turtle consumption because it increases the probability of dedicated or opportunistic turtle capture. Also, because most of the fishing in these communities is a subsistence activity, it may indicate less reliance on or reduced access to cash, prompting households to find other means of generating cash to meet their needs, if the latter. The mean for fishing is greater in Piso Firme (9.17 hours/week) than Remanso (6.42 hours/week). Size of household also has indirect effects on turtle consumption because more people to care for is likely to result in greater needs for food and cash or a higher probability that someone in the household will dedicate themselves to activities that indirectly or directly involve capture of turtles. On average, families in Piso Firme are larger than in Remanso. Turtle Meat Consumption in Piso Firme and Remanso Heads of several households in each community reported daily meat consumption in July and August (n=3 in Piso Firme, n=17 in Remanso). This data show that turtle constitutes a large proportion of the meat consumed during this period, which is the prime turtle consumption period because it coincides with nesting (Figure 5-2). Turtle meat constituted a quarter to a third of all meat consumed, which illustrates its dietary importance. Domestic meat (beef, chicken, pig and duck) also constituted large portions of meat consumed. All the socioeconomic characteristics directly related to turtle consumption (eat, send turtle, send eggs, hours turtling, sell and prefer) described in Table 5-2

PAGE 99

85 have higher means in Piso Firme than Remanso. This indicates both greater subsistence and market consumption of turtle in Piso Firme. The variable prefer was created to determine how much market integration was expressed by what was of more interest to household heads. The question asked was if you had the choice, which would you prefera P. unifilis turtle, the monetary equivalent in Bolivianos, or they are the same? The assumption made with this variable was that greater interest in having the turtle was an indication of a more subsistence-based household where turtles are worth more than their cash equivalent, while preference for the monetary equivalent indicated a more market-based household where cash held a higher value. Preference for neither the turtle nor its cash equivalent was an intermediate state and was interpreted as not suggestive of one livelihood system over another. In any case, turtle meat is an important part of peoples diet in both communitiesit was reported as consumed an average of once a week per household over the course of the year of data collection. With an average size of six in Remansos 101 households, this represents 31,500 meals, or 29% of the years meals if three meals a day are consumed. In Piso Firme, where average size is six among the 70 households and the average number of times turtle was eaten throughout the year was 1.25 times a week, turtle may represent the basis for 27,300 meals over the course of a year or 36% of a years meals. Based on the responses to the questions about prefer and eat, more turtle meat is consumed in Piso Firme.

PAGE 100

86 A fish31%tapir6%wild duck2%beef15% turtle24% peccary4%chicken4%paca6%pig6%deer0%duck2% B fish30% turtle34% duck3%deer1%chicken9%peccary1%pig6%paca0%tapir0%wild duck0%beef16% Figure 5-2. Proportions of meat consumed in a purposive sample of households from July to August 2000. Note that turtle meat is the largest proportion of meat consumed in these months. A) Sample of three households (47 meals) in Piso Firme B) Sample of 17 households (77 meals) in Remanso. Differences Between Subsistence and Market Consumption of Turtle in Piso Firme and Remanso To test hypothesis four (H4: there is more market consumption of turtle in Remanso than in Piso Firme and more subsistence consumption of turtle protein in Piso Firme than in Remanso) Mann-Whitney tests and Pearson chi-square tests of difference were used to compare the variables for turtle consumption between Piso Firme and Remanso (Table 5-2). Sell, prefer and send eggs were tested with a Pearson chi-square test because they are categorical response variables. The difference between communities for sell is significant (p=0.06) and indicates greater incidence of sale in Piso Firme. The test for

PAGE 101

87 prefer indicates a significant difference between households in Piso Firme and Remanso (p=0.07). Sixty-two percent of Remanseos prefer the monetary equivalent to turtle over turtle itself, which is interpreted to indicate greater market consumption, and there is greater preference for turtle in Piso Firme. The result for send eggs indicates that 97% of the time, sending eggs is a factor of community of residence with more turtles being shipped out of Piso Firme than Remanso (p=0.03). For send eggs the difference is significant (p=0.03), indicating that more households in Piso Firme send eggs out on the bus. Mann-Whitney tests for difference in subsistence and market consumption between Piso Firme and Remanso reveal that there are significant differences among these consumptive activities. For eat and hours turtling, which are taken to represent subsistence consumption, the tests reveal that residents of Piso Firme engage in more of both of these forms of subsistence consumption of turtle. Market consumption indicated by send turtle is also greater in Piso Firme. Thus, hypothesis four is not rejected for subsistence consumption in Piso Firme, there is more subsistence consumption of turtle in Piso Firme than in Remanso. Overall, three of the three variables indicating subsistence consumption and three of the four variables for market consumption suggest greater amounts of these activities in Piso Firme. The outcomes of tests of difference between communities for subsistence consumption variables (eat, hours turtling and prefer (turtle)) suggest that there is more subsistence consumption in Piso Firme. The outcomes of tests of difference between communities for market consumption variables (send turtle, send eggs, sell and prefer (money)) indicate more consumption in Piso Firme. Specifically,

PAGE 102

88 there is more market consumption in Piso Firme as demonstrated by more incidents of sell, send turtle and send eggs. There is a greater preference for the monetary equivalent in Remanso. The difference is a greater link of turtle with cash in Remanso than in Piso Firme. General Socioeconomic Characteristics by Wealth Groups Results of a chi-square test indicate that wealth is distributed differently in the two communities (Table 5-3). There is a greater percentage of wealthy households with more equal distribution of wealth in Remanso than Piso Firme. Because the distribution of wealth is different in the two communities, tests for difference among socioeconomic characteristics and turtle consumption were calculated by community. Table 5-3. Distribution of households in three wealth groups, with relative percent and percent of number of households in each wealth group. Maximum likelihood chi-square test results indicate a difference in distribution of wealth between Piso Firme and Remanso ( 2 =11.19 p <0.01 df=2). Wealth Group Piso Firme Remanso Number Households in Group 1 1 15 % within community 1.82% 21.13% % of total 0.79% 11.90% Number Households in Group 2 19 18 % within community 34.55% 25.35% % of total 15.08% 14.29% Number Households in Group 3 35 38 % within community 63.64% 53.52% % of total 27.78% 30.16% Wealth group one is wealthiest, three is least wealthy. General socioeconomic characteristics separated by the wealth groups determined by community members provide insight into the economic contexts of each community. These characteristics influence peoples access to cash, which in turn influences dependence on turtles as food or cash-generating resources (Table 5-4). Kruskal-Wallis and Pearsons chi square tests were used to examine differences among wealth groups. These results are shown in Table 5-4. Among the three wealth groups, the

PAGE 103

89 highest mean for chaco in Piso Firme was group two while in Remanso it was group one. While agriculture is the most common source of household income in the two communities, wealthier households tend to have greater land holdings. Because chacos of 1 ha or less are typically used for subsistence purposes, a higher proportion of households is involved in subsistence agriculture in Piso Firme although there is not a significant difference between the two communities in hectares cultivated. Hours fishing shows a negative relationship with wealth, with increasing amounts of time spent fishing by members of poorer households. The Kruskal-Wallis test indicates that this difference is highly significant among wealth groups in both communities. This is an important indirect variable since increased fishing time is the most likely indicator of increased opportunistic turtle catching. In both Piso Firme and Remanso the variables education and size suggest negative relationships with wealth. Hours fishing on average, increases as wealth decreases, which concurs with fishing being a subsistence activity and resulting in less access to wealth. Education is significantly different among wealth groups in both communities while the difference in household size is not significantly different among wealth groups in Piso Firme. A higher mean number of trips out of the community is attributed to those in the wealthiest group. In Remanso, the difference between wealth groups is significantly different, while in Piso Firme it is not. The pattern suggested by these numbers is that households with more monetary resources are more likely to come and go. In terms of the socioeconomic variables that directly influence turtle consumption, there is less variation among wealth groups in the two communities (Table 5-4). The

PAGE 104

Table 5-4. Descriptive statistics of socioeconomic characteristics that indirectly and directly influence turtle consumption detailed by wealth group within Piso Firme and Remanso. Test statistics for difference among wealth groups by community are included. Characteristic Wealth Group 1 Wealth Group 2 Wealth Group 3 Piso Firme x Std Dev n x Std Dev n x Std Dev n Test Statistic p-value Indirect Influence Chaco 1 0.88 0.73 17 0.78 0.62 35 H=1.86 0.39 Trips 1 2.44 1.92 18 1.83 1.55 33 H=3.9 0.14 Education 1 6.61 3.52 18 4.78 3.24 32 H=5.57 0.06 Hours fishing 1 10.63 9.40 19 46.6 51.68 35 H=15.97 0.0003 Size 1 5.79 2.68 19 6.31 2.63 35 H=3.07 0.22 Direct Influence Eat 1 1.72 1.84 18 1.23 1.31 31 H=3.55 0.17 Send turtle 1 1 1.51 15 0.5 1.30 32 H=2.63 0.27 Send eggs c 0% = yes 0% = no 1 8% = yes 25% = no 16 11% = yes 56% = no 32 2 =0.85 df=2 0.65 Hours turtling 1 4.23 8.98 19 15.27 29.54 33 H=3.35 0.19 Prefer c 0% = 1 0% = 2% = 3 1 16% = 1 2% = 2 16% = 3 17 33% = 1 4% = 2 27% = 3 31 2 =1.37 df=4 0.84 Sell c 0% = yes 5% = no 1 20% = yes 15% = no 7 55% = yes 5% = no 12 2 =5.97 df=2 0.05 Remanso Indirect Influence Chaco 2.2 6.50 15 0.24 0.39 18 0.93 1.45 38 H=4.38 0.11 Trips 7.57 8.12 14 1.22 0.73 18 2.27 4.47 38 H=15.15 0.0005 Education 10.62 5.03 13 7.18 3.84 17 4.78 3.15 36 H=15.08 0.0005 Hours fishing 1.13 2.26 15 30.71 56.14 17 37.90 37.89 37 H=23.48 <<0.0001 Size 3 3.10 15 5.78 2.26 18 6.92 2.56 38 H=20.61 <<0.0001 90

PAGE 105

Table 5-4. Continued Direct Influence x Std Dev n x Std Dev n x Std Dev n Test Statistic p-value Eat 1 0.76 15 0.83 0.79 18 0.84 0.97 38 H=1.34 0.51 Send turtle 0.47 1.25 15 0.06 0.24 18 0.03 0.17 36 H=2.33 0.31 Send eggs c 3% = yes 19% = no 15 1% = yes 25% = no 18 2% = yes 51% = no 36 2 =2.16 df=2 0.34 Hours turtling 0 6.35 16.92 17 9 16.52 35 H=6.07 0.05 Prefer c 7% = 1 3% = 2 12% = 3 15 7% = 1 1% = 2 17% = 3 18 14% = 1 6% = 2 33% =3 37 2 =0.94 df=4 0.92 Sell c 0% = yes 50% = no 6 8% = yes 0% = no 1 33% = yes 8% = no 5 2 =8.71 df=2 0.01 Note: See Table 5-2 for an explanation of the variables. Standard deviation calculated at 95% confidence level. Means only shown for interval variables and no calculations made for groups with n < 1. Pearsons chi square test applied to categorical variables, Kruskal-Wallis test applied to ordinal and interval data. Variables with significant differences between wealth groups at 95% confidence level shown in boldface. 91

PAGE 106

92 only significant difference among wealth groups in Piso Firme was sell while in Remanso sell and hours turtling were significantly different. In Piso Firme the incidence of sale increases as wealth decreases while in Remanso the group with the highest reporting of sale is group one. One observation that bears mention is that for the variables eat, send turtle, and send eggs the highest mean in Remanso is found among the wealthiest group. In Piso Firme it is found in the middle group. However, hours turtling in both communities is highest in the lowest wealth group. These findings suggest that turtling may be 1) a subsistence activity, 2) a specialized activity among poorer households or, 3) an activity that earns cash for those in the lowest wealth group, who sell it to the wealthier households. Differences Between Subsistence and Market Consumption of Turtle Among Wealth Groups The fifth hypothesis (H5: there is a negative relationship between wealth and turtle protein sale/trade such that as personal wealth increases the consumption of turtle protein decreases) was examined using the Kruskal-Wallis or Pearson Chi-square tests (Table 5-4). Of the interval variables examined, hours turtling in Remanso was the only one significantly different (p<0.05) between wealth groups. To determine where the specific differences lie for hours turtling among the three wealth groups in Remanso, two groups at a time were compared using a Mann-Whitney U test (Table 5-5). The results of the Mann-Whitney U test indicate that for hours turtling groups two and three are not different from each other but each is different from group one. People in the highest wealth group do not spend time looking for turtles. People in

PAGE 107

93 groups two and three spend an average of 6.35 and nine hours per week, respectively, turtling. Table 5-5. Differences in hours turtling between wealth groups in Remanso. Mann-Whitney U test shows that the wealthiest group does not dedicate time to catching turtles, while households in groups two and three do. Wealth Group 1 Wealth Group 2 Wealth Group 3 Hours turtling U p x sd n x sd n x sd n Wealth 1 vs. Wealth 2 105 0.09 15 6.35 16.92 17 Wealth 1 vs. Wealth 3 180 0.02 15 9 16.52 35 Wealth 2 vs. Wealth 3 259.5 0.34 6.35 16.92 17 9 16.52 35 Note significant result in boldface. Whether someone in the household sold turtle (sell) was an additional indicator of market use. To test for differences among wealth groups, the Pearson chi-square test was used. There were 19 responses in Piso Firme and 12 in Remanso to the question Does someone in your household sell turtle? The test indicates that the majority of households selling turtle are in wealth group three (11 of 15 in Piso Firme and four of five in Remanso) and none in wealth group one. This form of market involvement is clearly more evident among poorer households (Table 5-4). Considering that households in the lowest wealth group do more selling, hours dedicated to turtling may be for purposes of sale. Hypothesis five states that as wealth increases, the consumption of turtle protein decreases. The results of tests for difference in sell turtle between wealth groups indicate differences between wealth groups in both communities. The differences indicate that poorer residents are more likely to sell turtle. For this measure of turtle consumption the null hypothesis is rejected. As household wealth decreases there is

PAGE 108

94 increased incidence of turtle sale. Poorer households are more dependent upon turtle as a source of cash. Discussion Network and Income Differences between Piso Firme and Remanso Results suggest that market integration does not lead to greater turtle consumption in this area of eastern lowland Bolivia. The a priori classification based on less distance and greater connection to market as well as more cash availability in Remanso leading to its greater market integration that would affect turtle consumption does not hold. These characteristics, which provide a greater influx of people, goods and cash to Remanso, were also thought to negatively affect turtle consumption. Although I anticipated that this would be reflected in the data on income source, the availability of cash-generating job was not significantly different between Remanso and Piso Firme. Between the communities there was only one significant difference between the socioeconomic variables tested, for chaco. This difference may be an indicator of the different degrees of market integration between the two communities that is consistent with Von Thnens intensity theory. This is because Remanso, which is closer to market, has a higher average area of cultivated land. But it does not appear to increase the harvest of turtles. Piso Firme is much farther from its market than Remanso, which may make the cost-benefit ratio of producing agricultural goods not worth the effort and costs involved in transport. But turtle and eggs are more commonly shipped out of Piso Firme, suggesting that the increased market distance or the lack of access to cash generating income negatively affect them. Remanso may be more market integrated but this integration may not lead to more turtle consumption.

PAGE 109

95 All consumption variables tested (eat, send turtle, send eggs, hours turtling and prefer (turtle)) indicate greater turtle consumption in Piso Firme than Remanso. Hence, turtles represent an important resource in Piso Firme, more so than in Remanso. I argue that greater access to/reliance on cash, less reliance on agriculture, more education, more frequent trips out of the community, and smaller families are all indicators that households in Remanso have moved farther away from a subsistence livelihood than in Piso Firme. This appears to have a benefit for the consumption of turtle. Subsistence and Market Use by Community I believe that turtle is a useful indicator representing this difference between locations along the subsistence to market continuum. Although the data indicate that market integration is not a major factor in turtle consumption, other factors add nuance to the interpretation. These additional factors include: reliance on subsistence agriculture, time dedicated to fishing, and household size. The statistical analyses reveal significant differences between all indicators of subsistence and market consumption of turtles between Piso Firme and Remanso. Hours turtling, which is greater on average in Piso Firme than Remanso, was considered a subsistence activity because in general there appears to be less cash exchange in this community and more people prefer turtle in Piso Firme vs. the monetary equivalent in Remanso. However, considering additional variables this original assumption is not as straightforward. Based on the incidents of send turtle and send eggs reported, there is more market consumption in Piso Firme. This may indicate that hours turtling is not as much of a subsistence activity as originally thought, based on the definition of turtle consumed outside of the community as market consumption. Because most of the turtle shipped out of Piso Firme goes to family and not in exchange

PAGE 110

96 for cash, I propose this flow is an indication of cultural practices that are subsistence in origin. Residents in Piso Firme are more likely to have been born and raised in the floodplain and therefore have longer experience/greater preference for eating turtles. Remanso, being more of a colonist community, has less inclination towards turtle consumption as preference and therefore the demand for the turtle resource outside of the community is less. This difference in cultural origins may also explain why there is greater incidence of sale if turtles are less preferred for their meat than their cash equivalent in Remanso. Turtle may be more of a subsistence than market commodity in this area of Bolivia. Some people I spoke with in San Ignacio said they enjoy eating turtle and turtle eggs, and I heard an advertisement on the radio in San Ignacio for river turtles from the Itenz River. The stronger social connection between San Ignacio and Piso Firme suggests that some of the turtles that go out on the bus may be involved in commercial trade as I have defined it, although the residents I spoke with do not consider this commercialization. Regardless of who is defining commercialization, the end result of a captured turtle is less reproductive potential for the population and represents a net loss. This fact must be addressed in communities attitudes towards turtles and in any conservation and management plan. Most people who reported sending turtles or eggs out on the bus to Santa Cruz stated that they were going to family members who were working or studying in the city. Anyone I asked who grew up in Santa Cruz reacted with surprise or horror that people ate turtles, which suggests that there is not a large food market for turtles in Santa Cruz, although a larger sample is needed to make a confident statement to this effect. During

PAGE 111

97 the time I spent in Santa Cruz I never saw turtle meat or eggs for sale in food markets or on restaurant menus. On one occasion I saw turtles from the Beni River being sold as pets in the Siete Calles market in Santa Cruz. So, the turtles that are sent to Santa Cruz fall into a gray area in terms of subsistence consumption or market consumption, depending upon the definition used. This is a critical point to be incorporated into a management plan for the turtles. In terms of developing a conservation plan for the species, the majority of consumption does occur in the communities and is based on social practices and values and must also be considered. Turtle meat is not a highly commercialized product outside of the communities at this time, according to low numbers of turtles and eggs that go out on the bus. Analyses of difference suggest that market integration, which is less in Piso Firme than in Remanso, is not the main causal factor of greater turtle consumption. This is contrary to what I expected and may be positive in terms of potential growth of market integration in Piso Firme, which may ultimately lead to less turtle harvest. The implications for Remanso do not fully remove the threat of market integration to turtle harvest. While there is less turtle eaten and sent out from Remanso, the abundance data suggest that turtles still face hunting pressure there. Of course it is possible that the pressure is mainly from residents of Piso Firme, who appear to hunt less on the river adjacent to their community, which may be a result of the Parks presence or the earlier depletion of the turtle population. This is an area for further inquiry. Many community residents expressed to me on multiple occasions that they valued turtle meat as part of their cultural and social activities. The annual arrival of turtle season is a welcome and talked about event for many residents who describe turtle as a

PAGE 112

98 valuable source of strength for the physical labor that most people perform on a daily basis. A few comments were also made to me about the value of turtle meat as an aphrodisiac, particularly for men. The annual founding day celebration (September 24 in Piso Firme and September 29 in Remanso) is one of the most important celebrations for residents, who begin preparations more than three months in advance and mobilize in large numbers to prepare the communities. These celebrations, which last for several days, are characterized by abundant food and drink. It is a poor reflection on the community and households that are not able to offer food and drink to their guests, some of whom make multiple days journeys to help celebrate. Because of its cultural value, turtle is one of the most important dishes served to guests at the founding day celebrations. In Remanso I was invited to the guests dinner where a large P. expansa was served. On multiple occasions I walked by houses while someone was obviously preparing turtle only to be told later when I came to interview that no one in the household had eaten turtle at all that season. While it is possible that people simply forgot what they ate because eating turtle is an historic and common practice and problems with recall are well-documented (Bernard 2000; Bernard et al. 1984), another possibility was fear about the question. Perhaps the interviewees misunderstood the objective of my question, and by sharing such information with me they would disclose information that would negatively affect their consumption of turtle because that is one of the Park administrations objectives. Finally, there is confusion in the communities about what is and what is not legal turtle consumption, which caused a fair amount of fear and differing interpretations of the question. Because of this, people were not always willing to answer

PAGE 113

99 turtle-related questions. So, I take the results from the random sample as indications of trends, not explicit quantification of turtle meat consumption. Differences in Turtle Consumption Among Wealth Groups Wealth rank is related to two of the variables thought to directly represent turtle consumptionsell turtle in both communities and hours turtling per week in Remanso. Of the general socioeconomic characteristics considered indirect indicators of turtle consumption, wealth offers more of an explanation for differences in Remansobeing a source of significant difference in four of the variables considered. As wealth increases there is less sale of turtle by householdsno household in wealth group one claimed selling turtle while 47% of those households that reported selling turtle were in wealth group three. This may be related to the observation that wealth group three spends more time fishing, which provides more time to opportunistically catch turtle. But this group also spends more time specifically looking for turtles, which may be because selling turtles is one of the few options for acquisition of cash available to them. Also, none of the households in either community in wealth group three reported jobs that generated cash on a regular basis compared to 44% of those in wealth group one who reported having regular cash income opportunities. It may be that turtles are one of the few commodities that provide access to cash for people in wealth group three who do not report as many cash-generating jobs (but higher incidence of leaving on trips, which requires cash). Although the effect of wealth on turtle consumption was analyzed separately by community, there were many commonalities stated by the people who made the ranks, so the discussion covers them together (Appendix D). A common description of the poorest families in the wealth-ranking exercise was that these families do not have regular

PAGE 114

100 sources of income and rely much more on natural resources available in the surrounding rivers and forests. Possibly, group three is more dependent upon turtles as a source of cash income than the other groups, which are more likely to have consistent cash-generating regular employment because their education facilitates such employment opportunities. Household size, which has a negative relationship with wealth rank, is another source of significant difference among wealth groups in Remanso. This may contribute to turtle consumption. Larger families are more often observed in wealth groups two and three and larger household size brings with it increased needs. These needs may include increased need for trips outside the community because of increased probability of medical emergencies or educational purposes (schools in both communities end at the 8 th grade). This may increase the incidence of these households selling turtles for cash. Increasing numbers of goods and services are acquired only with cash and the promise of cash-generating activities provides an incentive to leave the communities. For example, residents frequently noted school supplies for children and fuel for boats as necessities they were pressed to find cash to acquire. The fact that many families have one or more member(s) outside of the community who sends remittances or goods to them is relevant. In the censuses of the communities almost 100% of the households in both communities noted at least one member living outside the community (this could be for school, cash-generating or other motives). In the population samples, 32% of families in Piso Firme and 30% of families in Remanso claim at least one household member working outside or sending remittances or goods to them from outside the community. Increased dependency on goods and services from outside could translate into increased

PAGE 115

101 sales of turtles and eggs for acquisition of goods or it could translate into decreased sales due to diversion of funds to other goods. My research is the starting point for the generation of questions aimed at teasing out the relationship between consumption of turtle and level of market integration, some of which were answered in my study, others serve as reminders that these are complex relationships, not as straightforward as they appear at first glance. Limitations of Data Collected The most significant shortcoming of the data from my perspective is the level of trust that was achieved between my interviewees and me. The presence of the Park and legislation (Bolivian, Brazilian and international) restricting quantity of turtle and egg harvest make turtles an illicit product. My status as an outsider dealing with this sensitive topic in an area where there is confusion and disagreement about it undoubtedly affected peoples perceptions of me. This in turn affected the quality of data I was able to collect. Some amount of skepticism is unavoidable in work that crosses cultural, national and economic boundaries. I made efforts to remain objective and maintain an appropriately naive approach to interviews and interactions in the field (Bernard 1995: 158-64). I have assumed that any biases in the data collected are consistent across samples. An additional potential bias concerns the question used to measure the marketing of turtles. Possibly my question about sending turtles/eggs out on the bus was only one of two appropriate questions. A second appropriate question would have dealt with sending turtles to the St. Simn mine, which lies just south of Remanso, in private vehicles. This question was not asked because this flow of turtles was not discovered until late in the field season. While I never witnessed it, I was told that truckloads of turtles were sent up to the mine, especially during the week preceding Easter, and sold to workers there, many

PAGE 116

102 who are from the communities (five households in Remanso and none in Piso Firme reported mining as their primary source of income). The demand for turtle meat at the mine is likely due in part to the combination of the Catholic Churchs having declared turtle meat a white meat, which is relevant to the majority of residents in this area who are practicing Catholics. Additionally, the mines are an employment source for residents of Remanso who take their cultural values of eating turtle with them when they leave their communities. While I was not able to get to the mine to confirm shipments of turtles, a Canadian manager working there reported seeing two turtles in the kitchen. Although he could not identify whether they were river turtles, it is likely that they were because they were immobilized by being laid on their backs, which is only effective for the Podocnemis turtles due to their wide shallow carapaces that prevent them from rolling over, unlike the land tortoises (the South American yellow-footed tortoise Geochelone denticulata and the South American red-footed tortoise G. carbonaria), which can right themselves and are immobilized by tying them up. My late discovery of this market consumption of turtles is due to one of several factors or some combination of them. One possibility is that my line of inquiry regarding market consumption of turtles was too narrow and did not elicit responses regarding this local marketI only asked about sending turtles out on the bus, which does not go to the mine, so I did not formally collect some potentially insightful information. Another possibility is that providing this information deflected attention away from practices in Remanso, where I was informed of this market, to the next community downriver. A third possibility is that the turtle issue, which was sensitive to begin with, was associated

PAGE 117

103 with the Park and affected my work because of my association with the Park. I suspect a combination of these factors was at play since I noted a change in peoples attitudes towards me after an event in Remanso crystallized two distinct groups within the communityone for the Park and the other against it. The event occurred while I was away from the community in early March 2000, so I missed firsthand many of the details involved. The main outcome of the event was the resignation of the mayor and the chief magistrate of the community, because of their purported involvement in the alleged sale of Remanso to the Park for the sum of US$2,000. The mayor was an immigrant from another part of Bolivia and had been controversial in the communitythis due in part to his outsider status as well as personal style (pers. comm. by resident, name withheld, Remanso October 2000). Because of research procedures and a choice madeI sought permission to conduct my research in the community from the mayor and then stayed with his family in January 2000I was also affiliated with the Park and the mayor in the eyes of the community. The end result for me, as an outsider asking questions about turtles and associated with the mayor, was that by extension I too was involved in the sale of the community, as a spy. Interestingly though, my classification as a spy provided grounds for some of the more insightful conversations with community members in regards to their use and consumption of turtle. Becoming part of local gossip opened up avenues for conversation about turtles, because it seems to have made me more of a local. In these conversations we covered topics ranging from peoples dependence on the turtles and their eggs, history of this use, what the future holds if nothing changes, and what conservation of turtles means for the people and the resource.

PAGE 118

104 One insightful piece of information acquired during this late phase of research had to do with the transport of turtles out of the community on the bus. I was told by a woman who had sent turtles on the bus as unregistered cargo that the manager of the bus in Remanso, in an attempt to make the community appear environmentally sound, refused to accept any turtles for shipment during the course of my study. That way he would not have to lie to me about the numbers of turtles or eggs nor would there be evidence in the bus records about this cargo going out of the community. People would also have fewer incidents to tell me about. This behavior affected responses to my question about the number of times households sent turtles or eggs out on the bus, but exactly how much was reported and whether or not it was consistent across households is conjecture. In both communities I am certain that some people were uncomfortable sharing this information with me. The best way to deal with this specific element was to watch as the bus was loaded up each week and to observe the cargo, which provided evidence that turtles did in fact go out on the bus from Remanso. The greater preference for cash in Remanso is likely to be driven by the greater amount of interaction with outside forces, a result of the influx of people and goods through the greater amount of river transit. But, this interaction also drives Remanseos greater need for cash as well, so they can participate in these external markets. It is not possible to say, based on the data here, whether the need for cash drives Remanseos to sell turtles for cash or if access to the market provides them with cash to buy turtles. Relevance of this Work to Future Research In contrast to the mainly subsistence patterns of turtle consumption observed, several incidences of external market consumption of turtles was seen during the research period. On at least two occasions a restaurant owner from San Ignacio came to Piso

PAGE 119

105 Firme and Remanso and bought turtlesI counted 16 turtles on one of these trips. If this is a regular occurrence or if it becomes more frequent this is ground for concern about the threat it poses to turtle populations in the area. Presently, these export activities are small-scale in the two communities studied in this research. A larger sample is required to determine region-wide effects of subsistence use of Podocnemis turtles on remaining populations. Extrapolating from my study and considering the additional research done to date does suggest that current use negatively affects turtle populations in the Amazon Basin. A basin-wide study of practices will provide insight into how different types of use (subsistence vs. market) impact these turtles. That is beyond the scope of this work, which was to quantify use patterns and turtle abundance near the Park. Conclusion Based on the differences observed between communities for subsistence vs. market use of turtle there are significant differences between the communities. A surprising result is the direction of the differences between the communities. There is more of both types of consumptionsubsistence and marketin Piso Firme than in Remanso. These results indicate greater market consumption of turtle (send turtle and send eggs) in Piso Firme than expected, which was initially thought to be characterized by exclusively subsistence consumption. Distance to market being greater is not enough of a deterrent to stop turtles or turtle eggs from being shipped out. However, as I discussed earlier, it appears that turtles/turtle eggs going out of Piso Firme are more cultural phenomena than economic. Tests of difference among wealth groups indicate that poorer residents dedicate more time to turtling and more selling of turtles than do wealthier households. These

PAGE 120

106 results indicate that there is a negative relationship between wealth and consumption of turtle protein. Turtle meat is an inferior good and its consumption declines as wealth increases. The implications of these findings for conservation are: 1) if road improvements occur, the impact on turtle harvest for sale and shipping should be closely monitored for changes that will need to be addressed; 2) cultural attitudes and related practices must be addressed in management planning; and, 3) monitoring of the wealth-turtle consumption relationship should continue in the event changes occur or additional measures are found to better understand the links.

PAGE 121

CHAPTER 6 COMMUNITY-BASED CONSERVATION AND MANAGEMENT Need for River Turtle Conservation Over-hunting of turtles and eggs has at least three negative outcomes: 1) a decline or loss of the species; 2) an alteration of the ecosystem when a component no longer fills its niche; and, 3) loss of a resource for humans (Moll and Moll 2000). This chapter discusses justification for community-based conservation and management based on the implications of abundance data on P. unifilis and P. expansa turtles encountered at three sites in eastern Bolivia. This comparison between two hunted and one less-hunted site was made to assess differences that might be due to hunting pressure. Distance from settlements and size of turtles are metrics that offer an indication of a negative impact on turtle populations due to human activity. These data may be used as part of the necessary ecological information for development of a community-based conservation and management plan for the turtle resource. I believe that community-based management and conservation is a relevant choice and worth the effort. There are limitations to community-based work. Some of the limitations include secure funding, consistent and effective involvement of people and political infighting. However, given these limitations I still believe that conservation efforts are more apt to benefit from local involvement than top-down approaches designed remotely. There are situations when community-involvement will not work and another approach will be more successful. When this is the case, alternative approaches should be developed. Given the local reliance on the turtle resource and the precedent of 107

PAGE 122

108 local involvement set in my research area I believe there are strong possibilities for community-based conservation and management and they should be attempted before other types of management and conservation. This is the approach I suggest be initiated first. As noted earlier, turtles provide food and cash to ribereos and provide ecosystem services. Turtles, especially hatchlings, are generally viewed as charismatic animals, which helps to rally support for their conservation after scientific studies have noted their declining numbers (Burke et al. 2000; Cantarelli 1997; Castro de Len 1986; Cavalcanti 1999; Ergueta and de Morales 1996; Fachn Tern 1994; Johns 1987; Klemens and Thorbjarnarson 1995; Luxmoore et al. 1988; Ministerio de Agricultura y Cria 1972; Pritchard and Trebbau 1984; Pritchard 1997b; Ramirez 1956; Thorbjarnarson et al. 2000; Tortoise and Freshwater Turtle Specialist Group 1991). Although 34% of identified turtle species are considered threatened or endangered, as a taxonomic group they have received much more attention than other herpetofauna (Mittermeier et al. 1992). This is promising for turtles given the scarcity of funds available for reptiles and amphibians overall as it may increase the number of potential conservation partners. One approach to conservation where concentration of efforts may prove productive is with "flagship species" or "charismatic megavertebrates", which include podocnemine turtles. As knowledge of turtle ecology grows, revealing their multifaceted roles, a wider variety of actors involved in projects aimed at achieving turtle survival becomes essential (Mittermeier et al. 1992). Role of Protected Areas and Wildlife Use in the Neotropics In the evolution of thought about protected areas as conservation units, different planning schemes have been tried since the founding of the idea of human-inhabited

PAGE 123

109 multiple-use protected areas in the early 1990s (Wells and Brandon 1992). One key factor that has been established as critical to the success of protected areas and species conservation is local participation in planning and implementation (Bromley and Cernea 1989; Chase 1989; Child 1995; Hurri Salmn et al. 1998; Little 1994; Shaw 1991; Songorwa 1999; Warner 1997; West and Brechin 1991; Western and Wright 1994). Agrawal and Yadama (1997) found that community institutions are mediating forces in forest conservation and unless communities possess and exercise the right to participate in management, benefits of other forms of participation are lost. Studies have shown that use is not necessarily detrimental to wildlife (Bodmer et al. 1997; Francis 1997; Vickers 1991). There are instances when wildlife use as a livelihood strategy creates incentives for communities to conserve the species they depend upon for protein or economic benefits (Godoy 2001). However, there are many more studies in direct contradiction to sustainable wildlife harvest by rural dwellers (Alvard 1994; Campbell 1998; Licata and Elguezabal 1997; Martin 1978; Ministerio de Agricultura y Cria 1972; Murray 2003; Peres and Lake 2003; Robinson and Redford 1994; Shaw 1991; Terborgh et al. 2002). Findings from ecological studies and monitoring specifically indicate that current uses of freshwater turtles of the genus Podocnemis in lowland Bolivia are threatening populations (Knothe and Mhlteich 1996; Quiroga Vera 2000). My observations at Piso Firme and Remanso suggest that human use does have a negative impact on turtles. Thus, conservation efforts are critical now. Conservation efforts that involve local residents are necessary because people are well-settled in the area.

PAGE 124

110 Community participation in natural resource conservation and management is a critical element for success because it facilitates involvement of local residents who are important allies for achieving program goals (Bodmer 2000; Bodmer et al. 1997; Borrini-Feyerband 1997; Noss 1998; Sols Rivera and Ayales 1997; Rubio et al. 2000; Ulloa et al. 1996; Warner 1997; West and Brechin 1991; Western and Wright 1994). Furthermore, experience has demonstrated that analyses that inform protected area planners how best to promote community involvement must necessarily take political, ecological and cultural factors into consideration (Bebbington 1996 and 1999; Bodmer 2000; Little 1994; Robinson and Redford 1994). This requires at least minimal community participation. Influence of Social Norms In a different context illustrating the power of social norms, Bernard (1987) found that sponge fishers in Greece, aware of risks involved, continued to engage in dangerous diving practices and accept less than "fair" payment for work. The implication of Bernards finding for Podocnemis turtles lies in the influence of social norms on group behavior. Ribereos who rely upon the turtles note difficulty in finding and harvesting turtles, yet they continue harvest patterns that are negatively affecting turtle abundance. Information from various sources (residents, previous studies and the results of my study) suggests that continued turtling at present levels poses a threat to turtle abundance and population maturity. Without changes in harvest practices, which may be brought about through changed social norms, the resilience of turtle populations may eventually be extinguished. One means of achieving change in social norms may begin with community participation in conservation and management planning. In 1996 an amendment

PAGE 125

111 increasing the Parks extent exacerbated conflict between Park officialswhose objective is the protection of wildlife and plant speciesand local residentswho rely on these natural resources around and within the Park and have established traditions of use, including turtles, that they do not want taken away. In the communities of Piso Firme and Remanso, some community leaders and residents have expressed interest in participating in a conservation and management plan for the turtles if their voices are heeded in the planning process. Also, the Park director, Gonzalo Pea Bello, states an interest in working with the communities to conserve and manage these turtles (Pea Bello personal communication 1999 and 2003), which is highlighted in the management plan as a key objective (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). This is a positive starting point for discussion. Harvest Rules In areas where harvest of species is permitted by law, impact monitoringthe observation of target species diversity and abundance in buffer zones and core protected areas for changes in ecological process as a result of human activitiescan be especially helpful (Kremen et al. 1994). In such areas, impact monitoring is necessary to insure the sustainability of integrated conservation and development programs. Development of conservation and management programs targeted at species such as P. unifilis and P. expansa will aid the development of sustainable use patterns. One confounding factor to determining sustainable harvest is that the edge of a species range is often the result of a physical or ecological boundary. Hence, populations in lowland eastern Bolivia may be more stressed than other populations and less resistant to additional pressures, such as hunting (Pritchard 1997a). Lowland eastern Bolivia, where Piso Firme and Remanso lie, marks the southwestern limit of the ranges of

PAGE 126

112 P. unifilis and P. expansa (Iverson 1992) (see Figures 6-1 and 6-2). It follows that individuals here may be at a greater risk of extirpation from continued anthropogenic stress. Historic records cite greater turtle abundance, which concurs with current local accounts of declining numbers. Thus, continued assessment of contemporary use is a priority, especially in areas where these turtles persist today. Assessment of turtle abundance and consumption are potential forms of impact monitoring that community members can undertake. Figure 6-1. Map of P. unifilis range (dots represent sightings) in the Amazon and Orinoco River Basins, (Iverson, J.B. 1992. A Revised Checklist with ond, Distribution Maps of the Turtles of the World. Privately Printed. RichmIndiana: 77).

PAGE 127

113 Orinoco River Basins (Iverson, J.B. 1992. A Revised Checklist with ond, In spite of this being the edge of their range, one reason for the persistence of P. unifil lt Figure 6-2. Map of P. expansa range (dots represent sightings) in the Amazon and Distribution Maps of the Turtles of the World. Privately Printed. RichmIndiana: 74). is and P. expansa may be Bolivias relatively low human population densitythe national average is 7 people/km 2 The even lower density in Santa Cruz Department5people/km 2 (de Mesa et al. 1999)is among the lowest densities in South America. Relative to other locations in their range, the lower human density in Bolivia may resuin less hunting pressure. Of all the countries where Podocnemis spp. are found, Bolivia has the second lowest human population density (Table 6-1). Developing conservation

PAGE 128

114 rules that effectively involve local residents now sets a positive precedent for the future, when human density is potentially greater. Table 6-1. Comparison of geopolitical areas, human populations and population densities in the ranges of P. unifilis and P. expansa. Country Area (km 2 ) Population b Population Density (people/km 2 ) Department of Beni, Bolivia a 213,564 2 Department of Santa Cruz, Bolivia a 370,621 5 Bolivia 1,098,587 8,445,134 7 Brazil 8,511,999 176,029,560 19 Colombia 1,141,752 41,008,227 33 Ecuador 283,563 13,447,494 41 Guyana 214,970 698,209 3 Peru 1,285,222 27,949,639 19 Venezuela 912,055 24,287,670 25 (Source: Allen 1999 except where a, Source: de Mesa et al. 1999 and b, Source: CIA World Factbook 2002) Common-Pool Resources and Community-Based Management Common-pool or common-property resources may be defined as a class of resources for which exclusion is difficult and joint use involves subtractability (Berkes 1991: 7). River turtles in lowland eastern Bolivia are common-pool resources. There are challenges associated with managing common-pool resources. One particularly relevant challenge associated with wildlife is assessing stocks and flows into and out of the inventory (Dolak and Ostrom 2003). Because of their movement and activity patterns, wildlife populations are often hard to quantify. Additionally, there are often geographical and scale differences as well as conflicts surrounding common-pool resource governance (Dolak and Ostrom 2003). The complexity associated with common-pool resources presents a set of challenges that need to be considered in the development of management plans for these resources. Continued turtle abundance counts are one means of dealing with the issue of stocks and flows for this common-pool wildlife resource. Community

PAGE 129

115 participation in governance is one way of potentially reducing conflicts and complexity and achieving conservation success. It may be argued that user groups are the best managers of common-pool resources. One argument for community-based management is based on empirical evidence that suggests traditional-based local level management is effective (e.g. Bray et al. 2003) and literature that suggests non-local state management plans are costly and ineffective (Child 1995; Freeman 1991). For conservation to be supported, incentives must exist for users to want to conserve. As Ostrom (1999) articulates, rules-of-use created by locals are more likely to be appropriate because they are based on local knowledge, incorporate the appropriate parties, have lower enforcement costs and adapt to the situation more quickly. In fact, multiple parties simultaneously working on aspects of management solutions are more likely to come up with rules that work more quickly than isolated efforts (Ostrom 1999), because more heads are better than one. One advantage to local residents conducting baseline studies to assess the effects of hunting is a long-term benefit for conservation by developing a sense of ownership among local users. Local users who live in the forests and depend upon the wildlife in the forests and adjoining rivers often have extensive life history and ecological knowledge of wildlife. They may also have significant incentives to conserve (c.f. Baskin 1994; Brockelman et al. 2002; Child 1995; Fitzgerald 1994; Moran 1991; Neumann 1998; Redford and Padoch 1992; Redford and Robinson 1985; Salafsky and Wollenberg 2000; Schwartzman et al. 2000; Songorwa 1999; Wadley et al. 1997). Local involvement in use studies is potential insurance that long-term work will be carried on.

PAGE 130

116 Working Towards Community-Based Wildlife Management and Conservation There are advocates for the exclusion of humans from areas dedicated to conservation (Redford and Sanderson 2000, Terborgh 2000), and advocates for working reserves as tenable means of achieving conservation (Wells and Brandon 1992; Schwartzman et al. 2000). The most effective means of conserving wildlife is in areas without humans, for "[m]uch research confirms that humans and wild nature are incompatible except where humans practice a low-impact pre-modern lifestyle at densities of no more than a few individuals per square kilometer. People damage the ecological system by clearing land, hunting, fishing, persecuting predators and commercializing natural resources" (Terborgh 2002: 307). However, large areas void of humans are more the exception than the rule in much of the world, and, in areas where humans are presently settled, removal is costly. Because relocation is most often financially and ethically untenable, conservation plans that involve local people are imperative. In the case of Podocnemis turtles, inclusion of Piso Firmeos and Remanseos in management and conservation has implications for the state of the resource because turtles are food resources desired by the people and their consumption is affected by socioeconomic factors (e.g. perception of wealth, access to cash, amount of cultivated land). Accounting for and understanding the role of socioeconomic driving forces improve chances of locally-appropriate management plan design. This in turn improves the probability of implementation success. At my research site, understanding relationships between local residents, the protected area and influences between them has implications for the effectiveness of the Park as a means of wildlife protection (e.g. Abbot and Thomas 2001).

PAGE 131

117 Locally-appropriate design, based on a conscientious understanding of the local socioeconomic context is important (Bebbington 1996). When local residents, who generally have the greatest stake in continued access to local resources, are included in the development process, projects are more likely to be accepted (Conway 1997). Ultimately, program approval by locals is an important element in achieving success (Casey-Lefkowitz 1997). Local involvement may increase popular support and contribute to project success if it includes information sharing, communication between the enforcing entity(ies) and the local public, recognition of environmental rights and means for citizen action (Casey-Lefkowitz 1997; Peres 1994). These elements can make it possible for people to participate in change without feeling as if they are being imposed upon without recourse. As Neumann (1998) noted in his study of forest and wildlife conservation in Tanzania, local residents have no incentive to participate in a conservation program that only takes rights away from them, offering no compensation in return. Over the past few decades, protected area planners have shifted their perceptions about conservation. Ulloa et al. (1996) describe conservation as having begun to be conceived of as a more complex process [than preservation] that implies the search for solutions together with local people and social movements (Ulloa et al. 1996: 50-1 my translation). Thus, participatory conservation efforts should include a diverse array of playersamong them resource users, planners, biologists, and managersat all stages from planning through implementation, as well as the delegation of responsibility for maintaining local biodiversity (Blaikie 2002; Bodmer and Puertas 2001; Thorbjarnarson et al. 2000; Schwartzman et al. 2000; Ulloa et al. 1996).

PAGE 132

118 Empowerment, decentralization of power and increased likelihood of project success are justifications for local participation in conservation. They should be considered important because people potentially involved in community-based conservation and management projects are often subsistence agriculturalists or extractivists, which generally means they have fewer options available to change their livelihood strategies. People with subsistence livelihoods typically do not have large savings from which to draw, nor the inclination to plan risky projectsalthough they may have an inclination to protect local biodiversity when their livelihoods depend on it (Agrawal and Yadama 1997; Twyman 2001). Here it is important to distinguish clearly between use of wildlife resources and strict dependence upon them, because those who are dependent face serious problems if access to a resource is cut (Bennett 2002). Knowledge of the local socioeconomic context (to the extent possible by an outsider) can elucidate peoples level of dependence on a resource, which is needed to appropriately develop conservation and management plans. Such understanding and inclusion of the local context are also important if new conservation and development interventions are to break from the status quo of past colonial practices that imposed Western ideas and images on conservation programs (Neumann 1997; Schwartzman et al. 2000; and see Pritchard 1997b). Inclusion of local people in wildlife conservation is an important step in shifting to sustainable use, albeit a difficult one. In spite of the challenges associated with converting unsustainable hunting practices, Bodmer (2000) gives the following reasons for involving local people in wildlife conservation: 1) they are the ones who hunt for subsistence food and monetary income; 2) hunting will be compatible with conservation

PAGE 133

119 only if harvest amounts are known; and 3) communities must be the ones to determine areas to be set aside as non-hunted in order that boundaries are respected and wildlife populations may recover from hunting pressure. Bodmer (2000) does not suggest an unconditional approval of community participation, but rather cautions that the influence of wildlife use must be investigated and the effects determined before going ahead with participation. I believe this approach is well-founded. Structure for Development of Community-Based Conservation If the premise of community-based management of common-pool resources is accepted as a valuable starting point for conserving them, a framework for the process is a next step towards implementation. Common sense suggests that resource users are more likely to accept rules that they have participated in drafting and modifying. An ample array of literature suggests this may be an appropriate strategy for conserving natural resources (Berkes 1991; Bodmer et al. 1997; Bodmer and Puertas 2001; Bray et al. 2003; Horowitz 1998; Larson 2002; Little 1994; Loomis 2000; Moffat et al. 1998; Mosse 1997; Noss 1998; Ostrom 1999; Peluso 1996; Pritchard 1997b; Redford and Padoch 1992; Songorwa 1999; Stevens 1997; Thibault and Blaney 2001; Warner 1997). In reference to forests in Java, Peluso articulates the tension that arises between resource users and controllers when not the same entity"[b]oth forest-based peasants and the state have constructed ideologies intended to justify their own rights to control forest access. Each side resists structural changes or concepts of management that would confer legitimacy on the other" (1994: 6). Centrally originating rules and laws may not be efficient in terms of conserving natural resources, often because they are not crafted with an understanding of the context in which they are to be implemented. However, the step from central to local government management is a large leap in many cases, which

PAGE 134

120 requires certain conditions for success. There are some minimal conditions that must be present for locally-based conservation and management plans to have even a hope of realizing success. These characteristics may be stated as (the list is adapted from Agrawal and Yadama 1997; Bray et al. 2003; Larson 2002): 1. presence of functioning internal organization 2. capacity 3. incentives for conservation or loss of use 4. transparency and accountability in the process The need for some form of internal organization in the community to support conservation is key (Bray et al. 2003) and can significantly affect the success of plans (Agrawal 1995; Peluso 1996). Ejidos and communal lands in Mexico are an example of devolution of local power resulting in conservation and benefits to local communities, due in large part to realization of benefits as a result of internal organization successfully responding to outside organizing efforts and returning these benefits to the communities (Bray et al. 2003). In the ejido example, the social capital is present and rules are set at the local level, facilitating adaptive management. The outcomes of this community-based organization and restrictions are conservation success in terms of increased forest canopy area, willingness to reduce unsustainable extraction practices, and in social terms in the form of monetary benefits flowing back into communities and being used for infrastructure improvements (Bray et al. 2003). One potential difference between forest management and wildlife management is the lower potential for monetary benefits associated with wildlife compared with timber. Lower monetary benefits may pose a challenge for turtle conservation but should not justify dispensing with the idea because while lower they may still be sufficient to generate local interest in conservation.

PAGE 135

121 The presence of the four conditions mentioned above is much more likely to result in successful conservation and management plans when these plans are drawn from an understanding of the local social, economic and political contexts of resource use (Berkes 1991; Peet and Watts 1996). Thus, management plans are unique in some way to each organizational context, in part because they may address common-pool resources that cross legal, political or cultural boundaries. For example, a management plan for turtles that is appropriate for Piso Firme will differ from a management plan for Remanso because each will address conservation and management of the same resource in a geographically similar area where cultural and economic contexts are distinct. Participatory Possibility in Bolivia A better understanding of market effects as measured through social and economic systems is one way of developing a community-based conservation and management plan. In the next section I will discuss possibilities for community based conservation and management based on my findings. Several precedents are encouraging for successful community participation. Currently residents of Piso Firme are in the process of acquiring title to their land, with the assistance of Fundacin Amigos de la Naturaleza, the NGO arm of the Park. This effort, begun over five years ago, has involved capacity building and participation of residents. Also, the present extent of the Park was achieved through Park officials working with community members in the five communities in and immediately surrounding the Park (Bella Vista, Florida, Piso Firme, Porvenir and Cachuela) to obtain their consent. Finally, a fisheries management plan was developed with active participation of Piso Firmeos in 2001. These events demonstrate cooperative precedent between the Park and communities in the area.

PAGE 136

122 In 2001, a team of community members from Piso Firme, Porvenir, Bella Vista and Florida (the four communities within the Park), Park biologists and employees completed a management plan for edible fish. The broad participation in this management plan set a positive precedent for additional participatory resource management in the area. Because the Podocnemis turtles found in this area are endangered and play an important role in human livelihoods, they are target animals for the next management plan (Ministerio de Desarollo Sostenible y Medio Ambiente 1996; Pea pers. comm. 2003). In terms of turtle population recovery, relocation of humans farther from the Park would be ideal, but realistically will not happen. However, the precedents stated above offer a positive outlook for the state of the turtle resource. I believe that because fishers in Piso Firme report catching most of their turtles outside of the Paragu River, most likely because large turtles have been hunted out of it, some change has occurred. Part of a management plan should consider substitutes for turtle, so that hunting pressure is not simply relocated to another area, thereby negatively affecting other turtle populations. Adaptive Management of Common-Pool Resources In field and laboratory studies Ostrom (1999) has documented cooperative group behavior, adjustment of resource use to maximize group benefits, and sanctioning when agreements are brokenwhen communication among actors is allowed. These observations suggest that policies set by multiple local parties are more effective than when decided upon by a central agency (Casey-Lefkovitch 1997; Ostrom 1999). This is an abbreviated explanation of adaptive community-based management. The possibility of policy constructed at the local level being adaptable is greater because it is more accessible and more appropriate to begin with, thereby facilitating changes when appropriate. The interaction of ecological and social systems provides a very complex

PAGE 137

123 arena within which to write policy. Elimination of as many biases as possible is in the best interest of developing appropriate policy, specific to each situation. One method worth investigating, which may eliminate bias, is to bring the policy making process to a local level, where policymakers are the resource users and policy enforcers. Many locally-organized user groups are complex adaptive systems, making changes as they go to systems comprised of myriad interconnections that do not readily reveal direct connections between interacting elements (Ostrom 1999). If policy is viewed as an experiment based on informed expectations about potential outcomes and the distribution of these outcomes for participants across time and space (Ostrom 1999) important ends towards conservation may be accomplished. Framework for Community-Based Turtle Conservation and Management I include here an outline of a structure for developing a community-based conservation and management plan for river turtles in Bolivia. I acknowledge the imperative involvement of local residents in development of natural resource conservation and management plans and recognize the fact that I, an outsider, collected the data reported. Because much discussion must still take place at the local level my recommendations are preliminarythey are intended as starting points for discussion. Perhaps the most important contribution of this data will be as baseline information for the two ribereo communities examined, whose residents depend upon the turtle resource. Parts of the research was collected in collaboration with the help of residents and there is no reason in terms of social capital that local residents could not continue data collection where I left off. I suggest the following four points, adapted from Agrawal and Yadama 1997; Bray et al. 2003; and, Larson 2002, as necessary foundations

PAGE 138

124 for successful community-based conservation and management and describe them in reference to turtle resources near Piso Firme and Remanso. Functioning Internal Organization Both Piso Firme and Remanso have functioning internal governance systems in the form of a mayor and a magistrate. Piso Firme also has an indigenous cabildo, the indigenous equivalent to the municipal government that works parallel to the municipal government. In both communities the municipal governments receive their mandates and some funds from the capital of their respective departmentTrinidad in the case of Remanso, San Ignacio for Piso Firme. The cabildo in Piso Firme handles some of the tasks the municipal government is responsible for in Remanso, mainly issues of dispute between parties. In both communities, the governing structure in place provides a stable framework upon which the community is based. Communities that are cohesive, self-reliant and possess a well-organized institutional structure appear to respond most positively to development interventions (this may be linked to size, with smaller communities exhibiting greater propensity to possess these qualities) (Abbot and Thomas 2001). The governing structures in Piso Firme and Remanso generally fit this description, which also provides them with a structure to handle changes. The governments are logical points from which to base a conservation plan. In each community there is at least partial interest among the governing individuals in a turtle conservation plan. Thus, there is a possibility for discussion at the very least about turtle conservation and management.

PAGE 139

125 Capacity Both communities have elementary schools that contribute to development of human capital through education. Additionally there is great local knowledge about wildlife and ecosystem function. Training in data collection on turtle abundance and consumption are necessary for scientific investigation purposes. These data are necessary for analyses of turtle abundance and consumption and methods can be taught to willing residents. Analyses of findings such as number of turtles per kilometer over time can be taught to participants so they may make informed management decisions. Incentives for Participation and Conservation Community members are valuable partners in the process of defining incentives for turtle conservation. By participating, communities can help determine incentives that can be extended to reduce hunting pressure. External organizations such as the Park and nongovernmental conservation organizations will also need to be included in some capacity for provision of funding, capacity-building and program development. Goals should include finding alternatives to turtle harvest for protein and cash and development of non-consumptive uses of turtles and other resources (ecotourism). Transparency and Accountability Processes that involve local residents from the beginning of the planning and development stages are more likely to be successful. Committees and books should be open to the public, which will require that the communities define members who are willing and able to do the extra work this entails. This will also involve outside organizations, at least initially, to provide funding.

PAGE 140

126 Points for Discussion Towards a Conservation and Management Plan Based on the abundance and use findings, several alternatives are possible for turtle conservation. If the residents in Piso Firme and Remanso are interested in being involved, the following points for discussion towards a conservation and management plan based on my research are relevant: turtles and eggs are sent out of the communities o decisions must be made regarding this practice I suggest complete a ban beginning immediately I suggest some form of enforcement determined and administered by community members turtles close to communities are less abundant o decisions must be made on catch limits and time of year for harvest based on turtle life history parameters I suggest communities determine catch limits based on criteria of their choice (i.e. size of household, size of chaco, access to cash) I suggest reduced or complete temporary ban on harvest for community-defined time frame I suggest catch be limited to smaller turtles (20 to 35 cm) as an attempt to limit reduction of reproductive females small turtles are more abundant than large turtles o decisions must be made on size limits of catch I suggest large males as prime targets (turtles 20 to 35 cm)

PAGE 141

127 turtles are less abundant closer to humans o decisions must be made as to catch limits I suggest limits based on household size and/or income

PAGE 142

CHAPTER 7 CONTRIBUTIONS AND FURTHER RESEARCH Introduction Throughout the world the interplay of economic, social, political and physical systems influences natural resource exploitation (Cutter and Renwick 1999). Market integration is one instance of this interplay that has implications for wildlife resources. Human-environment interaction is a core area of investigation in geography. Research dealing with human environment relationships calls for bridging between disciplines (Turner 1997). In this dissertation I have looked at a human-environment interaction and conducted my investigation using an interdisciplinary framework. The objective of my study was to quantify turtle consumption and its effects on turtle populations and assess the social and economic driving forces behind this consumption. An additional objective was quantifying turtle abundance as an added measure of the status of these wildlife resources. The data on human and ecological systems generated here offers baseline information on socioeconomic and ecological systems involving ribereos in two communities and river turtles at three sites in eastern lowland Bolivia. Social science methods and wildlife ecology provided frameworks for assessing natural resource use and conservation. Drawing upon these frameworks highlighted current and potential market influences on turtle use, based on a case study of human use of Podocnemis river turtles (summaries of the findings presented in Chapters Four and Five). I have also reviewed elements of market economics, wildlife use and common 128

PAGE 143

129 pool resource theory, adding turtle natural history and community-based conservation as each applies to my case study. Theoretical and Practical Contributions My main contribution to geographic theory lies in the realm of human-environment relationships. Specifically, I tested Von Thnens intensity theory, which proposes a negative spatial relationship between distance to market and intensity of harvest, adapting and applying it to abundance and harvest of Podocnemis river turtles. For turtles, distance to market is not a critical determinant of market consumption. This is likely due to the fact that turtles may survive long trips to market. Secondly, I have also contributed to the body of interdisciplinary work on relationships between market integration and wildlife-use patterns. Greater market integration, as indicated by proximity, access to cash and traffic flow, does not necessarily lead to more consumptive use of turtles. Third, my research generates additional information on the potential effects of human presence on turtle abundance and population maturity indicated by turtle size. My analyses indicate negative relationships between human presence and turtle abundance and population maturity, as indicated by turtle size. Finally, in terms of a potential contribution to applied conservation, my research may benefit the people in Piso Firme and Remanso, who depend directly upon the turtles, in development of a conservation and management plan. My baseline data will be returned to the communities in its entirety along with the recommendations I have suggested as starting points for discussion towards plan development. Residents may continue the data collection and analyses and apply them towards a community-sanctioned management and conservation plan if the interest and appropriate elements

PAGE 144

130 permit. I am hopeful my research will benefit the Park in its mission to work towards community-based conservation and management. Effects of Hunting Pressure on Turtle Abundance and Size Composition I adapted the spatial relationship Von Thnen (1966) observed for harvest of agricultural goods to turtle abundance and combined it with wildlife-human interaction theory at three sites. This was done because the hunting patterns that negatively affect turtle abundance also affect the value of turtles in a market economy context and these effects logically affect turtle harvest rates and therefore abundance and maturity. Expending great amounts of time and energy to harvest turtles make them more expensive, which may drive demand down. This is one possible explanation for what is occurring at Remanso. Additionally, in Remanso it appears that the availability of cash-generating work makes turtle harvest less attractive as a job. Piso Firme, on the other hand, is a more subsistence-based community and my data suggest that residents still rely heavily on turtle meat as a cash crop, most likely because there are few available substitutes. These explanations do not directly coincide with my abundance data. At Piso Firme, turtles are less abundant than at Remanso but do not exhibit as dramatic an absence around the community as do the turtles at Remanso. This is likely due to a combination of earlier reduction of the population by over hunting and the lower order of the Paragu River. This combination of factors likely causes fishers to leave the Paragu in search of turtles. It is possible they harvest from the area immediately surrounding Remanso, because the Paragu River empties into the Itenz River just upstream from Remanso. Potentially there are also fewer large turtles at Piso Firme, potentially a result of hunting pressure or ecological limiting factors. The net result for fishers is a greater time investment in turtling, causing them to forego other activities as

PAGE 145

131 they travel farther to capture larger turtles to make their efforts worthwhile. Again, this may lead to greater time investments among Piso Firmeos. The spatial relationship between humans and turtles was negative at Remanso and Piso Firme. Conversely, at Mangabalito, where there is presently no permanent human settlement, the abundance of turtles did not exhibit a negative spatial relationship radiating from the center of the transect. Because the main difference between Remanso and Mangabalito is the presence of humans, the conclusion that human presence negatively affects turtle abundance is appropriate. At Piso Firme the spatial relationship indicates more turtles downstream than upstream, in relation to humans. But, additional factors may be at play here. The Paragu River alongside Piso Firme is a lower order river than the Itenz, which is alongside Remanso and Mangabalito. At points upstream the Paragu narrows to a width of 1.5 m, which the Itenz does not do at any point along the Remanso and Mangabalito transects. I believe this is a contributing factor in the lower abundance counts at Piso Firme. There is also the fact that traveling upstream from Piso Firme one enters the Park. While patrolling by Park guards upriver at Piso Firme is no more frequent than at Mangabalito or Remanso, there is the possibility that presence of a Park station in Piso Firme has an impact, most likely through educational programs that the guards perform. Being stricter with ecological characteristics, a comparison between Remanso and Mangabalito is more appropriate. At Remanso the difference in turtle abundance exhibits a very clear pattern of less abundance radiating from the center of the transect. Here there is clearly a negative spatial relationship between humans and turtles, most likely due to hunting pressure. Hunting for turtles at Remanso has had a depressing effect on

PAGE 146

132 the turtle abundance, as has been observed in other parts of the Amazon Basin. If left unchecked, this pressure may result in local extirpation of P. unifilis and P. expansa. Demography of wildlife populations indicates maturity. In the case of Podocnemis turtles, which exhibit sexual dimorphism, size offers insight into the age and sex of individuals. My data show that there are more large turtles at Mangabalito than either Piso Firme or Remanso. The Gamma test of association indicates a negative association between turtle size and proximity to humans. The proportions of turtles in the different size classes indicate that the population at less-hunted Mangabalito is the most mature. The implication of maturity is greater long-term survival potential. If larger/more mature individuals are more frequent, the possibilities of reproduction and thus recruitment are improved. At Piso Firme, where abundance is lowest among the three sites, small turtles are more abundant than at Remanso. This suggests that the Piso Firme turtle population may be experiencing increased levels of recruitment after severe hunting pressure on the large turtles, which are the most attractive from catch per unit effort and economic perspectives. Fishers may be going elsewhere to capture turtles because the size of turtles available on the Paragu River is not worth the effort involved. This suggests the possibility that the Park may be exerting a positive effect on turtle recruitment through education, presence and/or patrolling efforts. Alternatively, residents of Piso Firme may be harvesting turtles from the Itenz River because the catch is so poor on the Paragu River. This may simply be explained as being a result of greater gains per unit effort or an informal common pool resource management plan. Further investigation and analysis of ecological differences are necessary to rule out confounding effects on turtle abundance such as river channel order and nesting site

PAGE 147

133 availability or food availability. Based on the consumption data and time dedicated to turtling observed in my data, human predation is a reasonable explanation for the patterns of reduced turtle abundance and the trend towards smaller turtles observed at Remanso and to a lesser extent at Piso Firme compared to Mangabalito. Intensity of Turtle Harvest as Related to Distance to Market The spatial relationship I observed between distance to market and intensity of turtle harvest did not follow the pattern observed by Von Thnen. Market consumption of turtles, as measured by sending turtles and turtle eggs out of the community, was a more frequent event at Piso Firme, which is farther from the market, than at Remanso. Turtle is not a commodity whose harvest is explained based solely on distance to market. A third measure of market consumption, preference for cash equivalent of turtles, was greater at Remanso than Piso Firme. This suggests that turtle is not as valuable as its cash equivalent when the market is closer. Considering additional indicators of intensity of turtle harvestfrequency of household consumption, sale of turtle, hours dedicated to turtlingintensity of harvest is greater at Piso Firme or farther from market rather than closer to it. Thus, Von Thnens intensity theory is not relevant to turtles as market commodities when distance is the main distinction used for the comparison. The lack of a significant relationship between distance to market and intensity of turtle harvest may be explained by other factors, including cultural context, access to cash/jobs that earn cash and wealth. Piso Firme, which has more of an indigenous background linking it to the geographic area than Remanso, also exhibits greater consumptive use of river turtles. This may be an artifact of the stronger riverine dweller history in this community. Remanso, which is located on the border with Brazil and has a population that is more heterogeneous in terms of geographic origin, exhibits less consumptive use of turtles.

PAGE 148

134 Access to cash through employment and other opportunities, is generally greater in Remanso. This too may influence turtle harvest. In Remanso, the stated preference for cash equivalent to turtle may be due to greater involvement in a cash economy, which is potentially linked to distance to market. This again suggest that turtle is not an important market commodity. The observation that most turtles and eggs sent out of Piso Firme went to family as gifts augment this potential relationship. These differences are suggestive of differing locations along the subsistence-market continuum. Because turtle is a cash-generating resource, these locations on the continuum likely affect the resource. Wealth and its Effect on Turtle Consumption In societies where natural resources have cash value and markets exist where these resources can be sold, different patterns have been observed. Godoy (2001), drawing on microeconomic theory, defined three relationships between neotropical wildlife and incomeinferior, necessity and superior goods. My analyses of the relationship between wealth (as a proxy for income) and river turtle consumption indicate that turtle is an inferior goodits consumption is negatively related to wealth. The magnitude of sale by the poorest households is much greater in Piso Firme than in Remanso and sale of turtle takes place on a larger scale in Piso Firme, where fewer households reported cash-generating work as the primary income source. Limited involvement in cash-generating work and greater involvement in turtling among poorer households suggest that turtle is an important source of food as well as cash, particularly for the poorest households. People of all wealth groups report eating turtle, but it is the poorest households that are more heavily involved in harvest and sale. These trends point to turtle as a likely source of cash income where other opportunities for cash-generating work are lacking.

PAGE 149

135 Additionally, cash is more highly valued by all wealth groups in Remanso, which is a more cash-based economy in comparison to Piso Firme. In Piso Firme, the preference for cash vs. turtle shows a trend towards turtle meat as wealth decreases. The direction of this relationship is likely the result of differences in cultural context as well as wealth distribution differences between the two communities. These relationships illustrate that turtle provides cash to households in a cash-scarce environment where there is a transition in process towards greater market integration. Practical Application of Consumption and Market Integration Findings It is important to consider the context of each of the three sites examined in my research. The community where the turtles are least abundant, Piso Firme, is closest to the Park. This fact emphasizes the vulnerability of wildlife to humans in spite of protection measures like the Park. But, the greater incidence of small turtles at Piso Firme, more time involved in turtling and information from fishers indicating that they rarely harvest turtles close to the community all suggest that this population may be in initial stages of recovery as well. This is potentially an instance of user group self-regulation, which has been observed by scholars of common property resources. This scenario also calls for greater involvement of residents in conservation and management efforts to insure consistent management practices and population recovery. Management and conservation of turtles is especially critical at Remanso, where there is the clearest link between human presence and reduced turtle abundance. If turtles at Mangabalito are recovering under conditions of fewer humans it is probable that the populations at Piso Firme and Remanso could rebound under similar conditions.

PAGE 150

136 Further Research My research presented here is a beginning effort to understand the context of river turtle use in eastern lowland Bolivia. Conducting my research has provided additional questions and forms of analyses that will provide further insight into the phenomena surrounding the human component of turtle consumption. Two areas of further research strike me as relevant to further development of this investigation. These topics will be the beginning of further research on this topic. First, further habitat analyses more closely linking habitat type and turtle abundance would be helpful. This would involve fine-tuning a habitat classification system and recording this for each turtle observed. This information will help eliminate doubt as to whether or not there are ecological differences among the three transects leading to differences in abundance. A second research extension would involve further investigation into socioeconomic driving forces in addition to distance to market, cash-generating employment, and volume of river traffic. Wealth is an area of particular interest that appears to have bearing on turtle consumption. To clarify the wealth-preference relationship, further analyses should consider size of turtles consumed aggregated by wealth groups.

PAGE 151

APPENDIX A BRIEF HISTORY OF PARQUE NACIONAL NOEL KEMPFF MERCADO Parque Nacional Huanchaca, created on June 28, 1979 by Supreme Decree number 16,446, was a precursor to Parque Nacional Noel Kempff Mercado. Parque Nacional Huanchaca had an area of 541,200 hectares. In March 1988 Parque Nacional Huanchaca was renamed Parque Nacional Noel Kempff Mercado, in honor of Professor Noel Kempff Mercado, a scientist who had contributed significantly to the knowledge of Bolivian flora and fauna and was killed on the Huanchaca Plateau during a scientific expedition. Parque Nacional Noel Kempff Mercado (hereafter the Park) was expanded three more times: first to 706,000 hectares by Supreme Decree number 21,997 on 31 August 1988, again by 21,900 hectares by Supreme Decree number 22,020 on 19 September 1988, and ultimately to its current extent of 1,523,446 hectares by Supreme Decree number 24,457 on 23 December 1996 (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). Administration of the Park has a dynamic history, involving government, industry and local residents. The Park, first specified by institutional foundation act number 720, was signed in April 1987. The Act established a multi-institutional board made up of the regional corporation for sustainable development of Santa Cruz (CORDECRUZ), the decentralized technical unit of the sustainable forestry center, the Ren Moreno autonomous university and the honorable municipal government of Santa Cruz de la Sierra. The board was granted responsibility of administrative aspects of Park implementation. On June 23, 1989, Supreme Decree number 22,231 was signed, creating 137

PAGE 152

138 the regional Center of Nature Conservation (CERCONA) as a decentralized unit of CORDECRUZ. The directorate of CERCONA consisted of the organizations named in the Institutional Foundation Act number 720 of 1987 plus the civic committee of San Ignacio de Velasco. A convention of cooperation signed on June 29, 1990 added Fundacin Amigos de la Naturaleza (FAN) as a member of the directorate of the Park for three years. This convention provided for the provision of financial and technical assistance intended to promote and reinforce conservation efforts of the Park, by FAN. When the structure of the national government changed in 1994, power over national parks was transferred to the National Secretary of Natural Resources and Environment under the Ministry of Sustainable Development and Environment (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). In 1994 FAN together with FONAMA (Fondo Nacional para el Medio Ambiente) and the National Secretary of Natural Resources and Environment administered funds from the Biological Diversity Conservation Project and Ecosystems of Protected Areas in Bolivia (PCBB) to pay personnel salaries, vehicular maintenance, storage of fuel and food for guards patrolling the Park. At the same time, funds from debt for nature swaps given by JP Morgan began the protection program for the Park (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). In April 1995 a contract between FAN and The Nature Conservancy was signed for elaboration of a management plan for the Park. Early in the process of elaborating a management plan, FAN called for a rapid ecological assessment of the area included in the Park to determine whether adequate habitats and areas were included within the limits for conservation purposes (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). Results of the assessment indicated

PAGE 153

139 that with an area less than one million hectares the Park was insufficient for conservation of minimal viable populations of birds of prey and carnivores of medium size (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). This information was presented to logging companies and human communities in the area with the hope of reaching an agreement to put more land under the Parks administration. Negotiations with the five logging companies in the area (San Martin, Moira, El Chore, Paragu, and El Paso) and logging concession owners resulted in reverting of the concessions to the Park. Meetings with community members who would be affected by an expansion (Piso Firme, Porvenir, Florida and Bella Vista) included discussions of the advantages and disadvantages of expanding Park boundaries as well as anticipated benefits and technical assistance that might be available if the Park was amplified. Based on these discussions, agreement was reached by community members for Park amplification. In December 1996 the Park was amplified to its present area of just over 1.5 million hectares (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The inclusion of tropical wet forests and seasonal wetlands within the final boundaries of the Park are significant because the forest area is believed to contain the greatest species diversity per unit area in the region and the sheer size of the Park offers a strong possibility for conservation. Simulations of minimum viable populations indicate that even given its large extension, the Park just meets the size requirement for conservation of some species found in the area (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). What is more, it is estimated that the present area of the Park contributes 131 million tons of organic material, the major part of which is transformed

PAGE 154

140 into atmospheric carbon dioxide, a significant contribution to global carbon sequestration (Ministerio de Desarollo Sostenible y Medio Ambiente 1996).

PAGE 155

APPENDIX B DESCRIPTION OF VEGETATION CLASSIFICATION SYSTEM Habitat Type Description Steep Barren Cliff 90 cut walls, usually clay or soft rock, leading directly into river Steep Cliff Aquatic Vegetation 90 cut walls with aquatic vegetation at base Steep Cliff with Sticks 90 cut walls with sticks/logs at base Tarope Thick, leafy aquatic vegetation eaten by Podocnemis turtles Canuela Grass-like aquatic vegetation eaten by Podocnemis turtles Mixed Aquatic Vegetation Mixture of tarope and canuela and possibly other aquatic vegetation High Sandy Beach Sandy beach more than 3 m high from river edge, appropriate for P. expansa nesting High Vegetative Beach Soil or sandy beach that slopes more than 3 m and is covered at least partially with vegetation Low Vegetative Beach Soil or sandy slope of less than 3 m with vegetation Muddy Flat Low muddy area with slope less than 45 Bays Opening into a bay Rocks Granite boulders Low Sandy Beach Gradual slope reaching less than 3 m beyond the river. Potentially appropriate nesting site. 141

PAGE 156

APPENDIX C INTERVIEW QUESTIONS Demographic Data Number and ages of people in the house Nombre y edades de los personas en la casa Years of education of each resident Hasta donde ha estudiado/a que curso lleg? Income How do you support your family? What is your work? Com mantiene su familia? Qu trabajo realiza? Where do you get your food? Com consigue su comida? Do you or other household members travel outside of Piso Firme/Remanso? If so, how often? Viaja ud. o alguien de la casa a fuera de Piso Firme/Remanso? Cada cuando? When you travel outside, what is your motive? A qu proposito viaja ud(s). a fuera de la comunidad? Do you have relative/friends who send you items from other places? Hay alguien afuera de Piso Firme/Remanso quien le manda cosas (dinero, comida, vibres)? Tell me about your chaco. Hablame de su chaco. How big is your chaco? Qu tamao tiene su chaco? How far is your chaco from your home? A que distancia est su chaco de aqui? Do you have pasture? Tiene ud. portrero? 142

PAGE 157

143 Turtles How much time do you or someone in the household dedicate each week to fishing? Cunto tiempo dedica ud(s). a la pesca cada semana? How much time do you or someone in the household dedicate each week to turtle fishing? Cunto tiempo dedica ud(s). a la busqueda de petas cada semana? Do you or does someone in the household sell turtles? Vende(n) petas? Do you ever send turtles or turtle eggs out of Piso Firme/Remanso? Envia ud(s). petas o sus huevos afuera de Piso Firme/Remanso? If you send turtles/eggs out, to whom do you send them? A quien quien manda la peta o los huevos? How often do you send turtles or eggs outside of Piso Firme/Remanso? Cada cuando manda petas o huevos a fuera de Piso Firme/Remanso? Did you eat more turtle/eggs in the past? Antes, comia ud. ms peta/huevos de peta que hoy dia? Which would you prefer a female Podocnemis unifilis, 20bs (the price of one) or they are the same? Cul preferiria tener ud. una tracaya, los 20 bolivianos o son igual? How many times did you eat turtle last week? Cuantos veces comian peta la semana pasado?

PAGE 158

APPENDIX D CHARACTERISTICS OF THREE WEALTH GROUPS Wealth Rank Characteristics Described in Piso Firme Characteristics Described in Remanso Participant One (self-classified in group 2, ranked 2) (self-classified in group 2, ranked 3) One have more of everything than all other groups have constant income from hotel or restaurant also sell meat to earn money have constant income source Two have regular work with the municipality or state or own their own business have enough from chaco to sell, but profits from sales not as important as for group three have large families eat bushmeat/fish/turtle as most of protein tied to natural resources have constant source of income Three have chaco that provides some food for year most have a large family have regular work or a chaco often have enough to sell to others (yucca, corn, rice, bananas, etc.) many have children studying in local school mainly farmers all own their house do not have enough to get by each day do not have disposable income many are fishermen who earn small amount, not enough to cover their costs have the largest families Four have no chaco have no regular work do not own a home live day to day in terms of livelihood most have small families 144

PAGE 159

145 Wealth Rank Characteristics Described in Piso Firme Characteristics Described in Remanso Participants Two and Three (couple) (self-classified in group 2, ranked 2) (self-classified in group 2, ranked 1) One have cattle they use for income or fixed salary that covers all costs store & ranch owners Two have a salary or cattle that covers most costs have some money because they have jobs if fishermen, have their own boat & motor have access to cash-generating work Three fishers, have domestic animals no set salary use their heads to find work, have personality live off their chacos sometimes find day labor practice trueque (barter) Four do not have everything they need work to acquire their basic necessities if confronted with an emergency need assistance from the community or institution Participant Four (self-classified in group 5, ranked 3) (self-classified in group 1, ranked 1) One have enough to eat & live very secure have access to cash-generating income (hotel/restaurant, store owners, fuel vendors, mine, transport, ranch, logging) Two have enough to eat, secure have money for necessities dont need to work daily for their food or living expenses have cattle (200+) access to cash-generating income, not as much as group 1 (state & day laborers) own their house have few head of cattle have chacos Three have money mostly store owners do not fish or hunt buy their meat live day to day dont all own their houses, not best houses fishermen lack training/avenues to improve situation have problems (i.e. drunkenness) that hinder Four have money but need to supplement with hunting & fishing sell fish or domestic animals for cash Five do not have cattle have domestic animals mostly fishermen all have chaco Six do not have a chaco do not fish or hunt no house, dont live in community permanently

PAGE 160

146 Wealth Rank Characteristics Described in Piso Firme Characteristics Described in Remanso Participant Five (self-classified in group 4, ranked 3) (self-classified in group 2, ranked 1) One offer work to others with less have enough to eat & live with have good salaries mostly storeowners & transportation vendors Two salaried workers give more credit to people in fourth group most have regular salary state or mine employees or hotel owners Three give credit to those in fourth group have cattle do not have enough economic resources to meet basic needs do not have regular salaries small, poor quality houses depend heavily on bushmeat/fish/turtle Four farmers work for others in the community day laborers live day by day covering their expenses/needs

PAGE 161

LIST OF REFERENCES Abbot, J. I. O. Thomas D. H. L. 2001. Understanding the Links Between Conservation and Development in the Bamenda Highlands, Cameroon. World Development 29 (7): 1115-36. Adams, A. M., T. G. Evans, R. Mohammed, and J. Farnsworth. 1997. Socioeconomic Stratification by Wealth Ranking: Is It Valid? World Development 25 (7): 1165-72. Agrawal, A. 1995. Population Pressure = Forest Degradation: An Oversimplistic Equation? Unasylva 181 (46): 50-58. Agrawal, A. 2001. Common Property Institutions and Sustainable Governance of Resources. World Development 29 (10): 1649-72. Agrawal, A., and G. N. Yadama. 1997. How Do Local Institutions Mediate Market and Population Pressures on Resources? Forest Panchayats in Kumaon, India. Development and Change 28: 435-65. Alho, C. J. R. 1985. Conservation and Management Strategies for Commonly Exploited Amazonian Turtles. Biological Conservation 32: 291-98. Allen, J. L. (Ed.). 1999. Student Atlas of World Geography. United States: Dushkin/McGraw-Hill. Allen, J. C., and D. F. Barnes. 1985. The Causes of Deforestation in Developing Countries. Annals of the Association of American Geographers 75 (2): 163-84. Alvard, M. S. 1994. Conservation by Native Peoples: Prey Choice in a Depleted Habitat. Human Nature 5 (2): 127-54. Anderson, A. B. 1990. Extraction and Forest Management by Rural Inhabitants in the Amazon Estuary. In Alternatives to Deforestation: Steps Toward Sustainable Use of the Amazon Rain Forest. (Ed.) Anthony B. Anderson, 65-85. New York: Columbia University Press. Angelsen, A. 1999. Agricultural Expansion and Deforestation: Modelling the Impact of Population, Market Forces and Property Rights. Journal of Development Economics 58: 185-218. 147

PAGE 162

148 Aramayo Cuenca, T. X. 1989. "Estuido Preliminar del Uso y Estado de Las Tortugas Acuaticas del Ro Maniqui (Departmento del Beni)." Universidad Mayor de San Andres, Cochabamba, Bolivia. Ayres, J. M., D. de Magalhes Lima, D. de Souza Martins, and J. L. K. Barreiros. 1991. On the Track of the Road: Changes in Subsistence Hunting in a Brazilian Amazonian Village. In Neotropical Wildlife Use and Conservation. (Eds.) J. G. Robinson, and K. H. Redford, 82-92. Chicago: University of Chicago Press. Baskin, Y. 1994. There's a New Wildlife Policy in Kenya: Use it or Lose It. Science 265: 733-34. Bates, H. W. 1962. The Naturalist on the River Amazons. Second ed. Los Angeles, California: University of California Press. [1863]. Bebbington, A. 1996. Movements, Modernizations, and Markets: Indigenous Organizations and Agrarian Stratgegies in Ecuador. In Liberation Ecologies: Environment, Development, Social Movements. (Eds.) R. Peet, and M. Watts, 86-109. London: Routledge. . 1999. Capitals and Capabilities: A Framework for Analyzing Peasant Viability, Rural Livelihoods and Poverty. World Development 27 (12): 2021-44. Bennett, E. L. 2002. Is There a Link Between Wildlife Meat and Food Security? Conservation Biology 16 (3): 590-592. Berkes, F., (Ed.). 1991. Common Property Resources: Ecology and Community-Based Sustainable Development. Dehra Dun, India: International Book Distributors. Bernard, H. R. 1987. Sponge Fishing and Technological Change in Greece. In Technology and Social Change. Second ed., (Eds.) H. Russell Bernard, and Pertti Pelto, 167-206. Prospect Heights, Illinois: Waveland Press, Inc. . 1995. Research Methods in Anthropology: Qualitative and Quantitative Approaches. Walnut Creek, California: Altamira Press. . 2000. Social Research Methods: Qualitative and Quantitative Approaches. Thousand Oaks, California: Sage Publications. Bernard, H. R., P. Kilworth, D. Kroenfeld, and L. Sailer. 1984. The Problem of Informant Inaccuracy: The Validity of Retrospective Data. Annual Review of Anthropology 13: 495-517. Blaikie, P. 2002. Participation, Sustainability and Devolution or, Kleptocracy, Privatisation and Degradation: Resource Management in Four African Countries. Presentation at Association of American Geographers Annual Meeting. Los Angeles, California. March 23.

PAGE 163

149 Bodmer, R. E. 2000. Integrating Hunting and Protected Areas in the Amazon. Priorities for the conservation of mammalian diversity: has the panda had its day? (Eds.) A. Entwistle, and N. Dunstone, 277-90. Cambridge: Cambridge University Press. Bodmer, R. E. 1995. Managing Amazonian Wildlife: Biological Correlates of Game Choice by Detribalized Hunters. Ecological Applications 5 (4): 872-77. Bodmer, R. E., J. Penn, P. Puertas, L. Moya, and T. G. Fang. 1997. Linking Conservation and Local People Through Sustainable Use of Natural Resources: Community-based Management in the Peruvian Amazon. In Harvesting Wild Species: Implications for Biodiversity Conservation. (Ed.) C.H. Freese, 315-58. Baltimore: The Johns Hopkins University Press. Bodmer, R. E. and E. Pezo. 2001. Rural Development and Sustainable Wildlife Use in Peru. Conservation Biology 15 (4): 1163-70. Bodmer, R. E. and P. Puertas. 2001. Community-based Co-management of Wildlife in the Peruvian Amazon. In Hunting for Sustainability in Tropical Forests. J.G. Robinson and E.L. Bennett (Eds.). New York: Columbia University Press. Borrini-Feyerabend, G. 1997. Collaborative Management of Protected Areas: Tailoring the Approach to the Context. Gland, Switzerland: Social Policy Group, International Union for the Conservation of Nature. Bozo, J. M. 1991. Historia de las Montaas, con sus Animales, Plantas, Rios y Gusanos. (Ed.) Germn Coimbra Sanz. Santa Cruz, Bolivia: Alcada Municipal. [1816]. Bray, D. B., L. Merino-Prez, P. Negreros-Castillo, G. Segura-Warnholtz, J. M. Torres-Rojo, and H. F. M. Vester. 2003. Mexico's Community-Managed Forests as a Global Model for Sustainable Landscapes. Conservation Biology 17 (3): 672-77. Brockelman, W. Y., M. Griffiths, M. Rao, R. Ruf, and N. Salafsky. 2002. Enforcement Mechanisms. Making Parks Work: Strategies for Preserving Tropical Nature. (Eds.) J. Terborgh, C. VanSchaik, L. Davenport, and M. Rao, 265-78. Washington, D.C.: Island Press. Bromley, D., and M. Cernea. 1989. The Management of Common Property Natural Resources: Some Conceptual and Operation Fallacies. Brondizio, E. S., E. F. Moran, P. Mausel, and Y. Wu. 1994. Land Use Change in the Amazon Estuary: Patterns of Caboclo Settlement and Landscape Management. Human Ecology 22 (3): 249-78. Bruno Coca, F. 1999. "Diversidad Alimenticia en Dos Especies de Tortugas Acuaticas: Podocnemis expansa y Podocnemis unifilis, en la Cuenca del Ro Itenez; Parque "Noel Kempff Mercado". Universidad Autonoma "Gabriel Rene Moreno". Santa Cruz de la Sierra, Bolivia.

PAGE 164

150 Burke, V. J., J. E. Lovich, and J. W. Gibbons. 2000. Conservation of Freshwater Turtles. In Turtle Conservation. Ed. M. W. Klemens, 156-79. Washington: Smithsonian Institution Press. Burt, J. E., and G. M. Barber. 1996. Elementary Statistics for Geographers. New York: The Guilford Press. Caballero Guerrero, J. 1994. "Manejo y Exito Reproductivo de la Tataruga Podocnemis expansa y la Tracaya Podocnemis unifilis en las Playas de Rio Itnez." Informe Trimestral Nro. 1, Santa Cruz, Bolivia. Campbell, L. M. 1998. Use Them or Lose Them? Conservation and the Consumptive Use of Marine Turtle Eggs at Ostional, Costa Rica. Environmental Conservation 25 (3): 305-19. Cantarelli, V. H. 1997. The Amazon Turtles Conservation and Management in Brazil. Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles An International Conference, New York Turtle and Tortoise Society, 407-10. New York. Casey-Lefkowitz, S. 1997. A Comparative Look at the Role of Citizens in Environmental Enforcement. National Environmental Enforcement Journal June: 29-44. Castells, M., and R. Laserna. 1994. The New Dependency: Technological Change and Socioeconomic Restructuring in Latin America. In Comparative National Development: Society and Economy in the New Global Order. (Eds.) A. Douglass Kincaid, and Alejandro Portes, 57-83. Chapel Hill: University of North Carolina Press. Castro de Len, Z. 1986. "Geografia Historica de la Tortuga del Orinoco." Universidad Central de Venezuela, Facultad de Humanidades y Educacin, Escuela de Historia. Caughley, G. 1977. Analysis of Vertebrate Populations. London: John Wiley and Sons. Cavalcanti, K. 1999. Matana amaznica. Veja 25 August: 100-01. Central Intelligence Agency. 2002. "CIA World Factbook 2002. "web page, [accessed 28 March 2002]. Available at http://www.cia.gov/cia/publications/factbook/index.html. Chase, A. 1989. The Janzen Heresy: Parks that Divorce Man from Nature are Not Natural. Cond Nast Traveler November: 122-27. Chernela, J. M. 1994. Tukanoan Fishing. National Geographic Research and Exploration 10 (4): 441-57.

PAGE 165

151 Child, G. 1995. Wildlife and People: the Zimbabwean Success. Harare, Zimbabwe: Wisdom Institute. Chinese University of Hong Kong. 2000. Web page, [accessed 29 August 2002]. Available at http://intl.econ.cuhk.edu.hk/exchange_rate_regime/index.php?cid=18 Cole, L. C. 1954. The Population Consequences of Life History Phenomena. The Quarterly Review of Biology 29: 103-37. Congdon, J. D., A. E. Dunham, and R. C. Van Loben Sels. 1993. Delayed Sexual Maturity and Demographics of Blanding's Turtle (Emydoidea blandingii): Implications for Conservation and Management of Long-Lived Organisms. Conservation Biology 7 (4): 826-33. Conway, K. 1997. Una Comparacin Entre las Actitudes de la Gente Aledaa hacia la Conservacin en Areas Protegidas en Costa Rica, Panam y Honduras. In Agroecologa: Memorias del Curso de la Organizacin para Estudios Tropicales de 1997. (Eds.) M. E. Swisher, and J. M. Mora, 183-92. Durham, North Carolina: Organization for Tropical Studies and Duke University Press. Coomes, O. T. 1996. State Credit Programs and the Peasantry under Populist Regimes: Lessons from the APRA Experience in the Peruvian Amazon. World Development 24 (8): 1333-46. Cutter, S. L., and W. H. Renwick. 1999. Exploitation, Conservation, Preservation: A Geographic Perspective on Natural Resource Use. New York: John Wiley and Sons, Inc. Darwin, R., M. Tsigas, J. Lewandrowski, and A. Raneses. 1996. Land Use and Cover in Ecological Economics. Ecological Economics 17: 157-81. de Mesa, J., T Gisbert, and C. D. Mesa Gisbert. 1999. Historia de Bolivia. La Paz, Bolivia: Editorial Gisbert. Denevan, W. M. 1996. A Bluff Model of Riverine Settelment in Prehistoric Amazonia. Annals of the Association of American Geographers 86 (4): 654-81. Dolak, N., and E. Ostrom, ((Eds.) 2003. The Commons in the New Millennium: Challenges and Adaptations. Cambridge: The MIT Press. Edwards, P. J., and C. Abivardi. 1998. The Value of Biodiversity: Where Ecology and Economy Blend. Biological Conservation 83 (3): 239-46. El Deber. 30 November 2000. La Humanidad tiene un Bien, el Parque Noel Kempff. El Deber, sec. Especial, pp. A16-17. Ergueta, P., and C. de Morales, (Eds.). 1996. Libro Rojo de los Vertebrados de Bolivia. Gland, Switzerland: International Union for the Conservation of Nature.

PAGE 166

152 Escalona, T., and J. E. Fa. 1998. Survival of Nests of the Terecay Turtle (Podocnemis unifilis) in the Nichare-Tawadu Rivers, Venezuela. Journal Zoology 244: 303-12. Fa, J. E., J. Juste, J. Perez del Val, and J. Castroviejo. 1995. Imapct of Market Hunting on Mammal Species in Equitorial Guinea. Conservation Biology 9 (5): 1107-15. Fachn Tern, A. 1992. "Alimentao de Cinco Espcies de Quelnios em Costa Marques, Rondnia Brasil." Instituto Nacional de Pesquisas da Amaznia/Universidade do Amazonas. . 1994. Depredacin de la Taricaya Podocnemis unifilis en la Reserva Nacional Pacaya-Samiria, Loreto. Boletin De Lima XVI (91-96): 417-23. Fachn Tern, A., A. Acosta Daz, and I. Vlchez Ramrez. 1992. Tortugas Podocnemis Mantenidas en Cautiverio en los Alrededores de Iquitos, Loreto-Peru. Boletin De Lima (84): 79-88. Fachn Tern, A., R. C. Vogt, and M. de Fatima Soares Gomez. 1995. Food Habits of an Assemblage of Five Species of Turtles in the Rio Guapor, Rondnia, Brazil. Journal of Herpetology 29 (4): 536-47. Fitzgerald, L. 1994. Tupinambis Lizards and People: A Sustainable Use Approach to Conservation and Development. Conservation Biology 8 (1): 12-15. Francis, R. C. 1997. Sustainable Use of Salmon: Its Effects on Biodiversity and Ecosystem Function. In Harvesting Wild Species: Implications for Biodiversity Conservation. (Ed.) C.H. Freese, 626-70. Baltimore: The Johns Hopkins University Press. Freeman, M. M. R. 1991. Graphs and Gaffs: A Cautionary Tale in the Common-Property Resources Debate. In Common Property Resources: Ecology and Community-Based Sustainable Development. Ed. F. Berkes, 92-109. Dehra Dun, India: International Book Distributors. Freese, C. H., (Ed.). 1997a. Harvesting Wild Species: Implications for Biodiversity Conservation. Baltimore, MD: The Johns Hopkins University Press. . 1997b. The "Use it or Lose it" Debate: Issues of a Conservation Paradox. In Harvesting Wild Species: Implications for Biodiversity Conservation. (Ed.) C.H. Freese. Baltimore, MD: The Johns Hopkins University Press. . 1998. Wild Species as Commodities. Washington, D.C: Island Press. Funes, M. C., and A. J. Novaro. 1999. Rol de la Fauna Silvestre en la Economa del Poblador Rural, Provincia del Neuquen, Argentina. Revista Argentina De Produccin Animal 19 (1): 265-71.

PAGE 167

153 Gibbons, J. W. 1968. Population Structure and Survivorship in the Painted Turtle, Chrysemys picta. Copeia 2: 260-68. Godoy, R. 2001. Indians, Markets, and Rainforests: Theory, Methods, Analysis. New York: Columbia University Press. Godoy, R., N. Brokaw, and D. Wilkie. 1995. The Effect of Income on the Extraction of Non-Timber Tropical Forest Products: Model, Hypotheses, and Preliminary Findings from the Sumu Indians of Nicaragua. Human Ecology 23 (1): 29-52. Godoy, R., N. Brokaw, D. Wilkie, D. Coln, A. Palermo, S. Lye, and S. Wei. 1998. Of Trade and Cognition: Markets and the Loss of Folk Knowledge Among the Tawahka Indians of the Honduran Rain Forest. Journal of Anthropological Research 54 (26): 219-33. Goulding, M. 1980. The Fishes and the Forest: Explorations in Amazonian Natural History. Berkeley: University of California Press. Gunderson, L. H., and C. S. Holling, (Eds.). 2002. Panarchy: Understanding Transformations in Human and Natural Systems. Washington: Island Press. Hames, R. 1987. Game Conservation or Efficient Hunting? In The Question of the Commons: The Culture and Ecology of Communal Resources. (Eds.) Bonnie J. McCay, and James M. Acheson, 92-107. Tucson: The University of Arizona Press. Herndon, W. L. 1952. Exploration of the Valley of the Amazon. (Ed.) Hamilton Basso. New York: McGraw-Hill Book Company, Inc. [1854]. Hiraoka, M. 1985. Mestizo Subsistence in Riparian Amazonia. National Geographic Research 1 (2): 236-46. Horowitz, L. S. 1998. Integrating Indigenous Resource Management with Wildlife Conservation: A Case Study of Batang Ai National Park, Sarawak, Malaysia. Human Ecology 26 (3): 371-403. Humboldt, A. 1852. Personal Narrative of Travels to the Equinoctial Regions of America 1799-1804. London: H.G. Bohn. Hurri Salmn, J., J. C. Schulze, and R. Haudry de Soucy. 1998. We Just Want What Belongs to Us. U.S. Regional Programme in Support of Indigneous Peoples of the Amazon Basin with International Fund for Agricultural Development and Andean Development Foundation. Iverson, J. B. 1992. A Revised Checklist with Distribution Maps of the Turtles of the World. Iverson. Johns, A. D. 1987. Continuing Problems for Amazon River Turtles. Oryx 21 (1): 25-28.

PAGE 168

154 Junk, W. J., and K. Furch. 1993. A General Review of Tropical South American Floodplains. Wetlands Ecology and Management 2 (4): 231-38. Keller, F. 1874. The Amazon and Madeira Rivers: Sketches and Descriptions from the Notebook of an Explorer. London: Chapman and Hall. Klemens, M. W., and J. B. Thorbjarnarson. 1995. Reptiles as a Food Source. Biodiversity and Conservation 4 (1): 281-98. Knothe, H., and A. Mhlteich. 1996. Manejo de la Tortuga Podocnemis unifilis en la Estacin Biolgica del Beni: Proyecto Quelonios 1991-1995. Informe Presentado 3/1996. Santa Cruz, Bolivia. Koper, N., and R. J. Brooks. 1998. Population-size Estimators and Unequal Catchability in Painted Turtles. Canadian Journal of Zoology 76: 458-65. Kremen, C., A. M. Merenlender, and D. D. Murphy. 1994. Ecological Monitoring: A Vital Need for Integrated Conservation and Development Programs in the Tropics. Conservation Biology 8 (2): 388-97. Kubitzki, K. Ziburski A. 1994. Seed Dispersal in Flood Plain Forests of Amazonia. Biotropica 26 (1): 30-43. Larson, A. M. 2002. Natural Resources and Decentralization in Nicaragua: Are Local Governments Up to the Job? World Development 30 (1): 17-31. Le Cointe, P. 1945. O Estado do Par: A Terra, A Agua e o Ar. Brasiliana Biblioteca Pedaggica Brasileira. So Paulo: Companhia Editora Nacional. Licata, L. 1992. La Tortuga Arrau y Su Conservacin, Ministerio del Ambiente y de los Recursos Naturales Renovables, Apure, Venezuela. Licata, L., and X. Elguezabal. 1997. Management Plan for the Giant Amazonian Turtle, Podocnemis expansa, in De La Tortuga Arrau Wildlife Refuge, Orinoco River, Venezuela. In Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles An International Conference, 171-73. New York Turtle and Tortoise Society. Little, P. D. 1994. The Link Between Local Participation and Improved Conservation: A Review of Issues and Experiences. Natural Connections: Perspectives in Community-Based Conservation. (Eds.) D. and R. M. Wright Western, 347-72. Washington, D.C: Island Press. Loomis, T. M. 2000. Indigenous Populations and Sustainable Development: Building on Indigenous Approaches to Holistic, Self-Determined Development. World Development 28 (5): 893-910.

PAGE 169

155 Luxmoore, R., B. Groombridge, and S. Broad, (Eds.) 1988. Significant Trade in Wildlife: A Review of Selected Species in CITES Appendix II., Vol. 2: Reptiles and Amphibians. South American River Turtle Podocnemis Expansa. Cambridge, UK: IUCN Conservation Monitoring Centre. Marcoy, P. 1875. Travels in South America: From the Pacific Ocean to the Atlantic Ocean, Vol. II Tumbuya-Sarayacu-Tierra Blanca-Nauta-Tabatinga-Santa Maria de Beni. New York: Scribner, Armstrong and Co. Martin, C. 1978. Keepers of the Game: Indian-Animal Relationships and the Fur Trade. Berkeley: University of California Press. Maw, H. L. 1989. Narativo da Passagem do Pacfico al Atlntico Atravs dos Andes nas Provncias do Norte do Peru e Descendo pelo Rio Amazonas At ao Par. Manaus, Brasil: Associao Comercial do Amazonas/Fundo Editorial. [1831]. Ministerio de Agricultura y Cria. 1972. La Problematica de la Tortuga Arrau, Oficina de Divulgacin Agricola, Caracas. Ministerio de Desarollo Sostenible y Medio Ambiente. 1996. Plan de Manejo Parque Nacional Noel Kempff Mercado. Santa Cruz, Bolivia: Fundacin Amigos de la Naturaleza and The Nature Conservancy. Mittermeier, R. A. 1975. A Turtle in Every Pot: A Valuable South American Resource Going to Waste. Animal Kingdom April/May: 9-14. Mittermeier, R. A. 1978. South America's River Turtles: Saving Them by Use. Oryx 14 (3): 222-30. Mittermeier, R. A., J. L. Carr, I. R. Swingland, T. B. Werner, and R. B. Mast. 1992. Conservation of Amphibians and Reptiles. In Herpetology: Current Research on the Biology of Amphibians and Reptiles: Proceedings of the First World Congress of Herpetology. (Ed.) Kraig Adler, 59-80. Oxford, Ohio: Society for the Study of Amphibians and Reptiles. Moffat, P., V. Fuentes, S. Langel, and K. Conway. 1998. A Model for Incorporating Community Participation in the Establishment and Management of a Protected Area. Journal of the Public Interest Environmental Conference 2 (1): 119-25. Moll, D., and K. P. Jansen. 1995. Evidence for a Role in Seed Dispersal by Two Tropical Herbivorous Turtles. Biotropica 27 (1): 121-27. Moll, E. O., and D. Moll. 2000. Conservation of River Turtles. In Turtle Conservation. Ed. M. W. Klemens, 126-55. Washington: Smithsonian Institution Press. Moran, E. F. 1991. Human Adaptive Strategies in Amazonian Blackwater Ecosystems. American Anthropologist 93 (2): 361-82.

PAGE 170

156 Morley, R. J. 2000. Origin and Evolution of Tropical Rain Forests. Chichester, England: Wiley and Sons. Mosse, D. 1997. The Symbolic Making of a Common Property Resource: History, Ecology and Locality in a Tank-irrigated Landscape in South India. Development and Change 28: 467-504. Murray, M. 2003. Overkill and Sustainable Use. Science 299 (21 March): 1851-53. Navarro, G., and M. Maldonado. 2002. Geografa Ecolgica de Bolivia: Vegetacin y Ambientes Acuticos. Cochabamba, Bolivia: Centro de Ecologa Simn I. Patio. Neumann, R. P. 1997. Primitive Ideas: Protected Area Buffer Zones and the Policies of Land in Africa. Development and Change 28: 559-82. . 1998. Imposing Wilderness: Struggles over Livelihood and Nature Preservation in Africa. Berkeley: University of California Press. Noss, A. 1998. El Monitoreo Comunitario de Caceria en el Izozog: Datos Preliminares. Ecologia En Bolivia 31 (April): 53-66. Odum, E. P. 1993. Ecology and Our Endangered Life-Support Systems. Sunderland, Massachusetts: Sinauer Associates, Inc. Oficina Regional de la FAO para Amrica Latina y el Caribe. 1988. Las Tortugas de Ro. Flora Fauna y Aras Silvestres 3 (7): 23-29. Ojasti, J. 1971. Un Recurso Natural Impropiamente Utilizado: La Tortuga Arrau del Orinoco. Defensa De La Naturaleza 1 (2): 3-9. . 1996. Wildlife Utilization in Latin America: Current Situation and Prospects for Sustainable Management. FAO Conservation Guide ed. FAO. Ostrom, E. 1999. Coping with Tragedies of the Commons. Annual Review of Political Science 2: 493-535. Paez, V. P., and B. C. Bock. 1998. Temperature Effect on Incubation Period in the Yellow-Spotted River Turtle, Podocnemis unifilis, in the Colombian Amazon. Chelonian Conservation and Biology 3 (1): 31-36. Peet, R., and M. Watts, (Eds.) 1996. Liberation Ecologies: Environment, Development, Social Movements. London: Routledge. Peluso, N. L. 1994. Rich Forests, Poor People: Resource Control and Resistance in Java. Berkeley, California: University of California Press.

PAGE 171

157 . 1996. Fruit Trees and Family Trees in an Anthropogenic Forest: Ethics of Access, Property Zones, and Environmental Change in Indonesia. Society for Comparative Study of Society and History: 510-48. Peres, C. A. 1994. Indigenous Reserves and Nature Conservation in Amazonian Forests. Conservation Biology 8 (2): 586-88. Peres, C. A., and I. R. Lake. 2003. Extent of Nontimber Resource Extraction in Tropical Forests: Accessibility to Game Vertebrates by Hunters in the Amazon Basin. Conservation Biology 17 (2): 521-35. Polisar, J. 1992. "Reproductive Biology and Exploitation of the Central American River Turtle Dermatemys mawii in Belize. University of Florida. Pritchard, P. C. H. 1997a. Conservation Strategies An Overview: Implications for Management. Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles An International Conference, New York Turtle and Tortoise Society, 467-71New York. . 1997b. Turtles as a Resource: Avoiding the "Tragedy of the Commons". Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles An International Conference, New York Turtle and Tortoise Society, 444-45. New York. Pritchard, P. C. H., and P. Trebbau. 1984. The Turtles of Venezuela. Oxford, Ohio: Society for the Study of Amphibians and Reptiles. Quiroga Vera, P. G. 2000. "Abundancia y Estructura Poblacional en Funcion de la Intervencion Humana y las Preferencias de Nidificacion de Podocnemis unifilis en el rio Maniqui." Universidad Mayor de San Simon: Cochabamba, Bolivia. Ramirez, M. V. 1956. La Tortuga: Estudio Biolgico de la Tortuga "Arrau" del Orinoco, Venezuela. Agricultor Venezolano 190: 44-63. Redford, K. H., and C. Padoch, (Eds.) 1992. Conservation of Neotropical Forests: Working from Traditional Resource Use. New York: Columbia University Press. Redford, K. H., and J. G. Robinson. 1985. Hunting by Indigenous Peoples and Conservation of Game Species. Cultural Survival Quaterly 9 (1): 41-44. . 1991. Subsistence and Commercial Uses of Wildlife in Latin America. In Neotropical Wildlife Use and Conservation. (Eds.) J. G. Robinson, and K. H. Redford, 6-23. Redford, K. H., and S. E. Sanderson. 2000. Extracting Humans from Nature. Conservation Biology 14 (5): 1362-64.

PAGE 172

158 Riester, J. 1977. Los Guarasug'w : Crnica de Sus ltimos Das. La Paz: Editorial Los Amigos del Libro. . 1985. CIDOB's Role in the Self-Determination of the Eastern Bolivian Indians. In Native Peoples and Economic Development: Six Case Studies from Latin America. Ed. T. Macdonald Jr., 55-74. Cambridge, Massachusetts: Cultural Survival, Inc. Robinson, J. G., and R. E. Bodmer. 1999. Towards Wildlife Management in Tropical Forests. Journal of Wildlife Management 63 (1): 1-13. Robinson, J. G., and K. H. Redford, (Eds.). 1991. Neotropical Wildlife Use and Conservation. Chicago: University of Chicago Press. Robinson, J. G., and K. H. Redford. 1994. Measuring the Sustainability of Hunting in Tropical Forests. Oryx 28 (4): 249-56. Rubio, H., A. Ulloa, C. Campos, and L. Piedrahita. 2000. Manejo de la Fauna de Caza, Una Construccin a Partir de lo Local: Mtodos y Herramientas. Bogot, Colombia: OREWA, Fundacin de Natura, Ministerio del Medio Ambiente, Organizacin de Estadso Iberoamericanos, Instituto Colombiano de Anthropologa e Historia, Fondo Mundial para la Naturaleza. Salafsky, Nick, and Eva Wollenberg. 2000. Linking Livelihoods and Conservation: A Conceptual Framework and Scale for Assessing the Integration of Human Needs and Biodiversity. World Development 28 (8): 1421-38. Salati, E., M. J. Dourojeanni, F. C. Novaes, A. Engrcia de Oliveira, R. W. Perritt, and H. O. R. Schubart. 1990. Amazonia. In The Earth as Transformed by Human Action: Global and Regional Changes in the Biosphere over the Past 300 Years. (Eds.) B. L. Turner II, W. C. Clark, R. W. Kates, J. F. Richards, J. T. Matthews, and W. B. Meyer II, 480-493. Cambridge: Cambridge University Press with Clark University. Santos Brito, W. L. dos, and M. Ferreira. 1978. Fauna Amaznica como Preferida como Alimento uma Anlise Regional. Brasil Florestal 9 (35): 11-17. Schwartzman, S., D. Nepstad, and A. Moreira. 2000. Arguing Tropical Forest Conservation: People versus Parks. Conservation Biology 14 (5): 1370-1374. Schweik, C. M., H. Nagendra, and D. R. Sinha. 2003. Using Satellite Imagery to Locate Innovative Forest Management Practices in Nepal. Ambio. Shaw, J. 1991. The Outlook for Sustainable Harvests of Wildlife in Latin America. In Neotropical Wildlife Use and Conservation. (Eds.) J. G. Robinson, and K. H. Redford, 24-34. Chicago: University of Chicago Press.

PAGE 173

159 Sierra, R., F. Rodriguez, and E. Losos. 1999. Forest Resource Use Change During Early Market Integration in Tropical Rain Forests: The Huaorani of Upper Amazonia. Ecological Economics 30: 107-19. Slocum, R., L. Wichhart, D. Rocheleau, and B. Thomas-Slater, (Eds.). 1995. Power, Process and Participation: Tools for Change. London: Intermediate Technology Publications. Smith, N. J. H. 1974. Destructive Exploitation of the South American River Turtle. Yearbook of the Association of Pacific Coast Geographers 36: 85-102. . 1999. The Amazon River Forest: A Natural History of Plants, Animals, and People. New York: Oxford University Press. Smyth, W., and F. Lowe. 1973. Narrative of a Journey from Lima to Par, Across the Andes and Down the Amazon: Undertaken with a Veiw of Ascertaining the Practicability of a Navigable Communication with the Atlantic. Boston: Milford House. [1836]. Soini, P. 1995a. "Bioecologa de la Taricaya (Podocnemis unifilis): Datos Nuevos y Actualizados." Reporte Pacaya-Samiria: Investigaciones en la Estacin Biolgica Cahuana 1979-1984, Centro de Datos para la Conservacin. Proyecto "Employment and Natural Resources Sustainability on Pacaya-Samiria National Reserve". Peru. . 1995b. "Ecologa y Situacin de la Charapa (Podocnemis expansa): Informe Preliminar." Reporte Pacaya-Samiria: Investigaciones en la Estacin Biolgica Cahuana 1979-1984, Centro de Datos para la Conservacin. Proyecto "Employment and Natural Resources Sustainability on Pacaya-Samiria National Reserve". Peru. Soini, P, and M. Cppula. 1995. "Estudio, Reproducin y Manejo de los Quelonios del Gnero Podocnemis (Charapa, Cupiso y Taricaya) en la Cuenca del Ro Pacaya, Loreto-Per." Reporte Pacaya-Samiria: Investigaciones en la Estacin Biolgica Cahuana 1979-1984, Centro de Datos para la Conservacin. Proyecto "Employment and Natural Resources Sustainability on Pacaya-Samiria National Reserve". Peru. Sols Rivera, V., and I. Ayales Cruz. 1997. El Manejo Comunitario de Vida Silvestre en Centroamrica: Algunos Indicadores y Mecanismos para su Sostenibilidad. Proceedings Tercer Congreso Internacional Sobre Manejo de Vida Silvestre en la AmazonaSanta Cruz, Bolivia. Songorwa, A. N. 1999. Community-Based Wildlife Management (CWM) in Tanzania: Are the Communities Interested? World Development 27 (12): 2061-79. Spix, and von Martius. 1968. Viagem pelo Brasil 1817-1820: Excertos y Ilustraes. So Paulo: Edies Melhoramentos.

PAGE 174

160 Spruce, R. 1908. Notes of a Botanist on the Amazon and Andes. (Ed.) Alfred Russell Wallace, Vol. II. London: Macmillan. Stevens, S., Ed. 1997. Conservation Through Cultural Survival: Indigenous Peoples and Protected Areas. Washington, D.C.: Island Press. Tang, C., and S. Tang. 2001. Negotiated Autonomy: Transforming Self-Governing Institutions for Local Common-Pool Resources in Two Tribal Villages in Taiwan. Human Ecology 29 (1): 49-67. Takasaki, Y., B. L. Barham, and O. T. Coomes. 2000. Rapid Rural Appraisal in Humid Tropical Forests: An Asset Possession-Based Approach and Validation Methods for Wealth Assessment Among Forest Peasant Households. World Development 28 (11): 1961-77. Terborgh, J. 1992. Diversity and the Tropical Rain Forest. New York: Scientific American Library. . 2000. The Fate of Tropical Forests: A Matter of Stewardship. Conservation Biology 14 (5): 1358-61. Terborgh, J., and C. VanSchaik. 2002. Why the World Needs Parks. In Making Parks Work: Strategies for Preserving Tropical Nature. (Eds.) J. Terborgh, C. VanSchaik, L. Davenport, and M. Rao, 3-14. Washington, D.C.: Island Press. Terborgh, J., C. VanSchaik, L. Davenport, and M. Rao, (Eds.). 2002. Making Parks Work: Strategies for Preserving Tropical Nature. Washington, D.C.: Island Press. Thibault, M., and S. Blaney. 2001. Sustainable Human Resources in a Protected Area in Southwestern Gabon. Conservation Biology 15 (3): 591-95. Thorbjarnarson, J., C. J. Lagueux, D. Bolze, M. W. Klemens, and A. B. Meylan. 2000. Human Use of Turtles. In Turtle Conservation. (Ed.) M. W. Klemens, 33-84. Washington: Smithsonian Institution Press. Tortoise and Freshwater Turtle Specialist Group. 1991. Tortoises and Freshwater Turtles: An Action Plan for their Conservation. IUCN/SSC/TFTG. Broadview, Illinois: International Union for the Conservation of Nature, Gland. Turner II, B. L. 1997. Spirals, Bridges and Tunnels: Engaging Human-Environment Perspectives in Geography. Ecumene 4 (2): 196-217. Twyman, C. 2001. Natural Resource Use and Livelihoods in Botswana's Wildlife Management Areas. Applied Geography 21: 45-68. Ulloa, A., H. Rubio, and C. Campos. 1996. Trua Wuandra: Estrategias para el Manejo de Fauna con Comunidades Embera en el Parque Nacional Natural Utra, Choc, Colombia. Colombia: Editorial buena Semilla.

PAGE 175

161 Urea Aranda, C. 2000. "Algunos Aspectos Sobre la Reproductin de Podocnemis unifilis en el Ro Maniqui, Estacin Biolgica del Beni." Universidad Mayor de San Simn. Vanzolini, P. E. 1977. A Brief Biometrical Note on the Reproductive Biology of some South American Podocnemis (Testudines, Peloedusidae). Papis Avulsos De Zoologia 31 (5): 79-102. Vickers, W. T. 1991. Ten Years in an Amazon Indian Territory. In Neotropical Wildlife Use and Conservation. (Eds.) J. G. Robinson, and K. H. Redford, 53-81. Chicago: University of Chicago Press. Vogt, R. (Dickturtle@aol.com). 13 October 2000. "Podocnemis in Brasil." E-mail to K. Conway (kmconway@hotmail.com). Von Thnen, J. 1966. Isolated State, An English Version of Der Isolierte Staat. New York: Pergamon Press. Wadley, R. L., C. J. Pierce Colfor, and I. G. Hood. 1997. Hunting Primates and Managing Forests: The Case of Iban Forest Farmers in Indonesia. Human Ecology 25 (2): 243-71. Warner, G. 1997. Participatory Management, Popular Knowledge, and Community Empowerment: The Case of Sea Urchin Harvesting in the Vieux-Fort Area of St. Lucia. Human Ecology 25 (1): 29-46. Wells, K., and K. Brandon. 1992. People and Parks: Linking Protected Area Management with Local Communities. Washington, D.C.: The World Bank. West, P. C., and S. R. Brechin, (Eds.). 1991. Resident Peoples and Protected Areas: Social Dilemmas and Strategies in International Conservation. Tuscon, Arizona: University of Arizona Press. Western, D., and R. M. Wright, (Eds.). 1994. Natural Connections: Perspectives in Community-Based Conservation. Washington, D.C: Island Press. Wilcox, B. A. 1995. Tropical Forest Resources and Biodiversity: The Risks of Forest Loss and Degradation. Unasylva 181 (46): 43-49. Wilkie, D. S., and R. A. Godoy. 2001. Income and Price Elasticities of Bushmeat Demand in Lowland Amerindian Societies. Conservation Biology 15 (3): 761-69. Wilson, E. O. 1988. The Diversity of Life. Cambridge: Harvard University Press. Yapa, L. 1998. The Poverty Discourse and the Poor in Sri Lanka. Transactions of the Institute of British Geographers 23: 95-115.

PAGE 176

BIOGRAPHICAL SKETCH Kristen Marie Conway was born on May 4, 1969, in Fairview Park, Ohio, to Joyce Marie Michals Conway and Timothy John Conway. She spent her childhood in northeastern Ohio. Two teachers along the way her seventh grade social studies teacher and her eleventh grade biology teacher inspired her to pursue what ultimately evolved into her interdisciplinary path of study. On graduating from high school, with these teachers combined examples and the encouragement and support of her family, she set off first to Massachusetts and then to Minnesota for college. Kristen graduated from Macalester College, St. Paul, Minnesota, in June 1992. After graduation Kristen moved to Washington, D.C., where she worked for two years on wildlife policy. For an additional year she worked in Carara Biological Reserve, Costa Rica, conducting fieldwork and combining efforts with park guards to develop an environmental education project for the Scarlet macaw (Ara macao). Ultimately, the sum of these experiences propelled her to graduate school, to further develop tools necessary to participate in community-based conservation and development. 162


Permanent Link: http://ufdc.ufl.edu/UFE0001481/00001

Material Information

Title: Human Use of Two Species of River Turtles (Podocnemis spp.) in Lowland Eastern Bolivia
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: UFE0001481:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001481/00001

Material Information

Title: Human Use of Two Species of River Turtles (Podocnemis spp.) in Lowland Eastern Bolivia
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: UFE0001481:00001


This item has the following downloads:


Full Text












HUMAN USE OF TWO SPECIES OF RIVER TURTLES (Podocnemis spp) IN
LOWLAND EASTERN BOLIVIA
















By

KRISTEN MARIE CONWAY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Kristen Marie Conway




























In my life there have been some women who have modeled success for me in special
ways. I wish to dedicate this work to them, beginning with my grandmothers-Ruth
Colletta Michals and Irene Francis Dool-who demonstrated strength and the value of
perseverance. Several new friends in Piso Firme (particularly Dofia Lucida Paz G6mez,
Dofia Nilda G6mez Pefia, Pura G6mez Pefia, Pura Montera Pefia, and Mariluz Cesari
G6mez) opened their lives and homes to me unselfishly as I faced new challenges, each
sharing her individual strengths and optimism, and demonstrating true virtue.















ACKNOWLEDGMENTS

For financial support of my research I am thankful to the National Science

Foundation (Grant # BCS-0082250) and The Chelonian Research Foundation. Funding

for preliminary research was provided by the InterAmerican Foundation (Grant # F-019)

and the Tropical Conservation and Development Program of the University of Florida.

For his patience in teaching me and sharing with me his enthusiasm for reptiles as

well as untiring editing assistance, I am greatly indebted to Dr. J. Parren Ross. For his

guidance, strong mentoring and exigencia I am deeply indebted to Dr. Charles Wood.

For her encouragement to reach new levels of inquiry and understanding of scientific

investigation I am grateful to Dr. Marilyn Swisher who has taught me valuable lessons

through two graduate degrees. Every young scholar who has the benefit of mentors such

as I found in Dr. Ross, Dr. Wood and Dr. Swisher is truly gifted. Dr. Michael Binford

and Dr. Nigel Smith offered thoughtful insights along the way. Two past committee

members-Dr. Richard Bodmer and Dr. Edward Malecki-were extremely helpful in the

intellectual development of my course of study and the design stage of my research. I

benefited greatly from their involvement. Dr. Ken Portier's guidance on categorical data

analysis was absolutely critical. I am deeply grateful to the community members of Piso

Firme and Remanso, Bolivia, who allowed me to enter into their lives and ask some

strange questions, all for the sake of my research. Their time, help, and input are

appreciated. I look forward to continuing relationships with them in the years to come.

For his help with transect setup and data logging I am indebted to wildlife biologist Stan









Howarter. I am grateful to my family, Joyce Marie Michals Conway, my mother-from

the beginning the ceaseless supporter of my education who taught me that happiness

comes in many forms; my father, Timothy John Conway-who has always encouraged

and supported my independence; and my sister, Julie Anne Beckles-who regularly

teaches me the importance of taking time to enjoy life, because the little details do take

care of themselves. Many dear friends have helped me with diverse tasks related to this

project as well as life in general. Without friends as good as these I might have lost my

way. For their generous help and support I thank Alisa Coffin, Amanda Stronza-Rojas,

Andreana Ososki, Ann Angelheart, Beatriz Alvarado Heath, Carol Lee Hilewick, Claudia

Stickler, Elli Sugita, Jocelyn Peskin Col6n, Naomi Trenier Sosa, Noemi and Roberto

Porro, Pete Gozar, Samantha Stone, Tracy VanHolt and Zulma Villegas. Ultimately, I

am truly grateful for the support-spoken and unspoken-love, encouragement

(especially in those tough moments), humor and wisdom of my husband, Luis G6mez

Paz.
















TABLE OF CONTENTS

Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ................................................................... ....... ....... ix

LIST OF FIGURES ......... ........................................... ............ xi

A B S T R A C T .......................................... .................................................. x iii

CHAPTER

1 INTRODUCTION ............... .................................. ................... 1

2 TU R TLE S IN A M AZON IA ............................................................................. 5

H history of Turtle U se .................. ...................................... .. ........ .. 5
Turtles in the R iver E cosystem ............................................................... ............... ....8
Implications of Large Turtle Life History for Conservation and Management ..........10
Legislative Stopgaps ................................................. .. ...... .................. 12
V alue of W wildlife, Including Turtles ............... ..................................................... 12
Value of Podocnemis Turtles in Eastern Lowland Bolivia .....................................13
Price of River Turtles in Lowland Bolivia ................................................... 16
M markets as they Apply to Turtles............................ ..... ......................... 17
Influence of the Market on Neotropical Wildlife ...................................................17
N negative Im pacts of M arkets................................................................... ...... 18
Positive Effects of Markets .................................................20
Market Integration and Hunting Intensity-Effects on People...........................20
Influence of Incom e on H unting..................................................................... ...... 21
Theoretical Derivation of Hypotheses ............................ ........ ............. .................. 26
Im plications of Findings ..............................................................................29
Note on my Research Style and Findings .............. ............................................30

3 RESEARCH AREA: EASTERN BOLIVIA ...................................... ...............32

River and Floodplain Habitats of Parque Nacional Noel Kempff Mercado .......32
Parque Nacional Noel Kempff Mercado and Ecological Benefits it Provides....34
The C om m unity of Piso Firm e .................................................................. .. ... 36
Indigenous Links to Piso Firm e................................... .......................... 38
The C om m unity of R em anso.................................... .......................... .. ......... 42









D iverse Cultural Links to R em anso .................... ... .. ...................... .... 43
Note on the History of Natural Resource Use in the Area from the 1940's to
1 9 8 0 's .........................................................................4 4

4 EFFECTS OF HUMAN SETTLEMENT ON THE NUMBER AND SIZE
DISTRIBUTION OF TURTLES ........ .............................................47

Intro du action .................................................................................................... 4 7
S tu d y A rea ................................................................4 7
Survey P rocedure........... .................................................. .......... ...... ..... .. 49
H habitat D description ......... ...................................................... ...... .... .... 53
Results ............... .............. ....................54
Mean Numbers of Turtles at Different Distances from Transect Center ............54
River Edge H habitat Characteristics ........................................ ............... 61
Size D differences am ong Sites........................................................ ................... 63
Comparison of Size of Turtles at Different Distances from Transect Center .....66
D iscu ssion ....................... ..... .. ... ............. .... .... ................ ........... ...... 69
Differences in Turtle Abundance in Relationship to Humans...........................69
Differences in Sizes in Relationship to Humans .............................................71
Conclusion ................................. .. .................................. ......... 72

5 SOCIAL STRATIFICATION, MARKET INTEGRATION, AND HOUSEHOLD
C O N SU M PTIO N ................. .................................. ................ .. ............. 74

O objectives and H ypotheses........... ............... ............................. ............... 74
M e th o d s ...........................................................................7 6
Sample Size Calculation.................................................. 76
Interview Administration and Questions.................. ........................77
W health R ranking ................... .... .. .................... .............. .... ...... .. .......... 78
Definition of Market Goods and Market Integration ................................ ....79
Socioeconomic Characteristics and Their Implications for Livelihoods ............80
Turtle Meat Consumption in Piso Firme and Remanso ...................................84
Differences Between Subsistence and Market Consumption of Turtle in Piso
Firm e and Remanso ................................. ........ ........ .. ...... ... 86
General Socioeconomic Characteristics by Wealth Groups ................................88
Differences Between Subsistence and Market Consumption of Turtle Among
W health G groups .......................................... ................... .. ...... 92
Discussion ...................................... .. ....... ........ ..... ..... .................... 94
Network and Income Differences between Piso Firme and Remanso ................94
Subsistence and Market Use by Community....................................................95
Differences in Turtle Consumption Among Wealth Groups ..................................99
Limitations of Data Collected ............................ .................... 101
Relevance of this W ork to Future Research.................... .................. ................104
C conclusion ........... ................................................... ................... 105









6 COMMUNITY-BASED CONSERVATION AND MANAGEMENT ................... 107

N eed for River Turtle Conservation ......... ........... ........................................ 107
Role of Protected Areas and Wildlife Use in the Neotropics.............................. 108
Influence of Social N orm s...... .............................................. .. ............... 110
H arv est R u les ........................... ......... .... ... ......................... ....... 111
Common-Pool Resources and Community-Based Management ...........................114
Working Towards Community-Based Wildlife Management and Conservation .....116
Structure for Development of Community-Based Conservation.............................119
Participatory Possibility in B olivia............................................... ........ ....... 121
Adaptive Management of Common-Pool Resources ............................................122
Framework for Community-Based Turtle Conservation and Management .............123
Functioning Internal Organization ....................................... ............... 124
Capacity ...................................................... ............ .......... 125
Incentives for Participation and Conservation ..............................................125
Transparency and A ccountability................................................................. 125
Points for Discussion Towards a Conservation and Management Plan ................. 126

7 CONTRIBUTIONS AND FURTHER RESEARCH ............................................128

Introdu action ......................................................... ............ ................. 12 8
Theoretical and Practical Contributions ..................................................................129
Effects of Hunting Pressure on Turtle Abundance and Size Composition .......130
Intensity of Turtle Harvest as Related to Distance to Market .........................133
Wealth and its Effect on Turtle Consumption..................................134
Practical Application of Consumption and Market Integration Findings .........135
Further Research ........... .... .......... .. .................. ......... 136

APPENDIX

A BRIEF HISTORY OF PARQUE NATIONAL NOEL KEMPFF MERCADO ......137

B DESCRIPTION OF VEGETATION CLASSIFICATION SYSTEM ...................141

C IN TER V IEW Q U E STIO N S ......................................................... .....................142

D CHARACTERISTICS OF THREE WEALTH GROUPS ......................................144

L IST O F R E F E R E N C E S ...................................................................... ..................... 147

BIOGRAPHICAL SKETCH ............................................................. ............... 162
















LIST OF TABLES


Table p

2-1 Average values of carapace length and body weight of female Podocnemis
expansa turtles in the Orinoco and Amazon River Basins, an indication of their
value as protein resources .............................................. .............................. 14

2-2 Research questions addressed in this research and the hypotheses used to test the
questions .................... ...... ................................ .... ... ...... .. 29

3-1 Numbers of flora and fauna species known to exist in Parque Nacional Noel
Kempff M ercado, indicating biodiversity present...............................................3..35

4-1 Descriptive statistics for the turtles counted along the three 40-km transects (Piso
Firm e, Rem anso and M angabalito).........................................................................54

4-2 Kruskal-Wallis test statistics comparing median number of turtles in segments
within (Close) and beyond (Far) 10 km of the center point between the three sites.
These results indicate that the difference is not random. .......................................55

4-3 Mann-Whitney test statistics for the comparison of close vs.far segments within
sites. Results indicate a significantly higher median number of turtles farther
from the center of the Remanso transect only ......................................................55

4-4 Frequencies of counts greater than or less than/equal to median values within 10
km (Close) or beyond (Far) transect center points of the three sites.....................56

4-5 Test statistics for test of difference between numbers of turtles counted at close
vs. far segments from center point of three transects...................... ..................58

4-6 Comparison of differences between sites in numbers of turtles counted at close
vs. far segm ents from center points....................................................................... 58

4-7 Proportions of turtle size classes seen within sites at far and close categories with
averages over all three sites and at Mangabalito, the less-hunted site ...................66

4-8 Kruskal-Wallis test statistics comparing all four size classes at close vs. far
segments from center point at all three sites. Results indicate significant
differences between size classes in close vs. far segments overall ........................68









4-9 Frequencies of size classes counted in relationship to medians for far vs. close
transect segments when all three sites are combined. ............................................68

4-10 Gamma test for known size classes at different distances compared between
hunted and less hunted sites indicating a negative association between turtle size
at close vs. far distances in hunted and less-hunted areas .....................................69

5-1 Motivation for trips taken out of communities to San Ignacio, Bolivia and
Pimenteiras, Brazil by residents of Piso Firme and Remanso..............................82

5-2 Descriptive and test statistics of socioeconomic characteristics that directly and
indirectly influence turtle consumption detailed by community.............................83

5-3 Distribution of households in three wealth groups, with relative percent and
percent of number of households in each wealth group ..........................................88

5-4 Descriptive statistics of socioeconomic characteristics that indirectly and directly
influence turtle consumption detailed by wealth group within Piso Firme and
R em anso. .............................................................................90

5-5 Differences in "hours turtling" between wealth groups in Remanso.
Mann-Whitney U test shows that the wealthiest group does not dedicate time to
catching turtles, while households in groups two and three do.............................93

6-1 Comparison of geopolitical areas, human populations and population densities in
the ranges of P. unifilis and P. expansa ......... ............................. ......... .......... 114
















LIST OF FIGURES


Figure p

2-1 Conibo Indians along Ucayali in 1866 Turtle eggs being crushed to extract oil .......6

2-2 Turtle being cooked on A m azon R iver ........................................... .....................7

3-1 Map indicating the communities where research was conducted, nearby cities that
community members have links with, the Park and the Paragua and
Itenez Rivers............................................... ......... ...................... 33

3-2 Aerial view of seasonally flooded savanna habitat, a scarce habitat type in the
tropical Americas, pictured here near the Flor de Oro guard station in Parque
Nacional Noel Kempff M ercado..................................... ........ ....... 35

3-3 Arco Iris Waterfall, a popular ecotourism destination on the Paucerna River in
Parque Nacional N oel Kempff M ercado............................................................... 36

3-4 Piso Firme from the Paragua River in the dry season..............................................37

3-5 Distribution of jobs that provide the major source of income for households
interviewed in the community of Piso Firme, located along the Paragua River......37

3-6 Example of a ceramic pot presumed of Guarasug'we origin, unearthed in Piso
F irm e in N ovem ber 200 1 ................................... ............. ................................... 39

3-7 Panero basket made from motacu palm fronds (Attaleaphalerata) and used to
carry harvested products in Piso Firm e ....................................... ....................... 41

3-8 Distribution of jobs that provide the major source of income for households
interviewed in the community of Remanso, located along the Itenez River............43

4-1 Podocnemis unifilis turtles basking along the Itenez River ...................................51

4-2 Schematic diagram of the 40-km transects illustrating location of Piso Firme and
Remanso at the center of the 40-km transects and categorization of km into
distance categories.......... .............................................. ...... .....52

4-3 Number of turtles counted at Piso Firme, along the Paragua River, shown in
10-km segments. Each line represents a complete survey along the 40-km
tran sect. .............................................................................57









4-4 Number of turtles counted at Remanso, along the Itenez River, shown in 10-km
segments. Each line represents a complete survey along the 40-km transect.........57

4-5 Number of turtles counted at Mangabalito, along the Itenez River, shown in
10-km segments. Each line represents a complete survey along the 40-km
tran sect. .............................................................................58

4-6 Average number of turtles seen in each km over eight samples at populated sites
(Piso Firme and Remanso) with one standard deviation indicated by bars on each
kilometer and the communities at KM 20 ..................................... ............... 59

4-7 Average number of turtles seen in each km over eight samples at the unpopulated
site (Mangabalito) with one standard deviation indicated by bars on each
kilom eter ............................................................................ ........ ....... 60

4-8 Meters of each habitat type at the three sites, shown individually by site. ..............62

4-9 Number of turtles in each size class observed with one standard deviation of the
number of turtles in each size class at each site and relative percent of the turtles
at the site in each of the four size classes noted .............. ....................................64

4-10 Proportions of turtles of different size classes by distance categories at the three
site s .......................................................................................... . 6 7

5-1 Main source of household income, in increasing order of cash-generating
potential in Piso Firm e and Rem anso................................. ........................ 81

5-2 Proportions of meat consumed in a purposive sample of households from July to
A ugust 2000. .........................................................................86

6-1 Map ofP. unifilis range (dots represent sightings) in the Amazon and Orinoco
River B asins, .................................................................... ........ 112

6-2 Map of P. expansa range (dots represent sightings) in the Amazon and Orinoco
R iver B asins ...................................................................... .........113















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HUMAN USE OF TWO SPECIES OF RIVER TURTLES (Podocnemis spp) IN
LOWLAND EASTERN BOLIVIA


By

Kristen Marie Conway

August 2004
Chair: Michael W. Binford
Major Department: Geography


Centuries of harvest (of adults and eggs) have left remaining populations of two

species of Amazonian freshwater turtles, Podocnemis unifilis and P. expansa, seriously

reduced throughout their ranges. Today, these turtles remain important wildlife resources

for riberefo communities in and near National Park Noel Kempff Mercado, in the

Bolivian Amazon for subsistence and market consumption. I quantified turtle abundance

and investigated the impacts of socioeconomic factors driving hunting pressure to

determine whether turtle abundance differs depending on whether sites are adjacent to

human communities or not and if different socioeconomic characteristics affect turtle

consumption. The communities where my study was conducted may ultimately use this

information to inform a management plan for their turtle resources.

Basking counts at three sites-two adjacent to human communities and one

adjacent to an abandoned community site-were conducted to obtain information on the









impact of hunting on turtle abundance. Tests of difference applied to these data show

differences in turtle abundance between sites with and without human communities,

which may be a result of hunting pressure. Size classes of turtles among the three sites

were negatively associated with proximity to humans.

Analyses of socioeconomic data show differences between subsistence and market

consumption patterns of turtle between the two communities. My findings suggest

different levels of market integration, not influenced by distance to market as anticipated,

do affect turtle consumption. Local phenomena including household wealth, subsistence-

based livelihoods and indigenous origins are suggested as explanations for turtle

consumption.

The combined abundance, size class and socioeconomic data gathered in my study

have implications for the effects of increased access to cash income on wildlife resources

and community-based conservation and management of species important in subsistence

and market economies. Poorer households are more involved in turtle harvest and sale.

The more subsistence-based community, which has indigenous origins, reports more

turtle consumption. Increased access to cash, through road improvement or greater

market integration may actually reduce turtle consumption, positively affecting turtle

abundance.

Since the communities studied lie in the buffer zone of a protected area, the results

have implications for protected area management. The impacts of hunting and

information on some of the driving forces behind it, provide baseline information for a

management plan in one eastern lowland Bolivia setting, which may provide insight for

community-based wildlife conservation elsewhere in Amazonia.














CHAPTER 1
INTRODUCTION

The concern for the preservation of biodiversity in fragile environments has been

voiced by researchers and many others the world over, and has been reflected in the

environmental policies adopted by numerous countries. The importance given to

biodiversity has defined a research agenda that seeks to identify endangered species, and

strives to understand the dynamics of human behavior that threaten the sustainable

reproduction of particular animal populations. Critical dimensions of this larger research

agenda are studies that address the ways in which particular rural communities in lowland

tropical regions exploit the environment for their livelihood, and harvest animals for sale

in local and regional markets.

My dissertation focuses on the harvest for consumption of two species of turtles,

Podocnemis unifilis and P. expansa in rural lowland tropical communities. The site for

my research is eastern Bolivia, on the Paragua and Itenez/Guapore Rivers. The research

design targets three locations: Piso Firme-a smaller and comparatively subsistence-

oriented settlement, Remanso-a larger, wealthier, and more market-oriented settlement,

and a third site, Mangabalito-the site of a human settlement that was abandoned in

1996.

Data collection and analyses are organized in terms of two different yet related

topics. The first is based on field observations of the abundance and size distribution of

turtles at varying distances from the human settlements. The objective is to generate

quantitative estimates of the effects of turtle harvesting on the abundance/size structure of









the turtle populations. The second analytical priority is based on socioeconomic data

obtained through random surveys of households in Piso Firme and Remanso, the rural

lowland tropical communities. In this case, the objective was to identify community and

household-level variables associated with the harvest of river turtles.

Taken together, the findings provide insights into human behaviors that profoundly

affect wildlife species that have long played a major role in the riverine environment of

Eastern lowland Bolivia and across the Amazon Basin. The results provide the baseline

data that can be used to promote community-based conservation and management

programs-a strategy that has the most promising potential to alter human behavior in a

manner that promotes the sustainability of threatened species.

To set the stage for the empirical analysis, my initial objective is to review the

historical uses of P. unifilis and P. expansa, describe the ecology of turtles, and

summarize research that documents the ecological effects of the harvesting and

commercialization of animals in the lowland tropics. Chapter Two is therefore divided

into four sections: historic use; turtle ecology; value of turtles; and, turtles and markets.

The characteristics of the research area are described in Chapter Three.

Chapter Four presents data on changes in the number and sizes of turtles as one

travels along the river from the human settlement sites to more distant regions, upstream

and downstream. The organizing theme of Chapter Four is derived from geographic

theory that examines the spatial relationship between distance to market and value of

agricultural goods by adapting this relationship to a wildlife resource. The specific

hypotheses tested are: HI: turtles will be less abundant closer to human communities

(where there is hunting); H2: there will be more turtles at Mangabalito (less hunting









pressure) than at Remanso (more hunting pressure); and, H3: there will be fewer large

turtles closer to human communities than farther away (because hunters prefer larger

turtles).

Chapter Five turns attention to characteristics of the communities of Piso Firme and

Remanso. The analyses are based on effects of market integration and social

stratification on household consumption of turtles. The following hypotheses are tested:

H4: there is greater market consumption of turtle protein in Remanso than in Piso Firme

and greater subsistence consumption of turtle in Piso Firme than in Remanso (because

there is closer access to market, more cash and more traffic in and out of the latter); and,

H5: there is a negative relationship between wealth and turtle protein sale/trade such that

as personal wealth increases, the consumption of turtle protein decreases (based on initial

field observations). The findings point to a number of important conclusions: the

greatest threat to turtles is local subsistence consumption; and geographic location

influences cash-generating possibilities, which in turn affect social stratification and

market integration.

The results of Chapters Four and Five can provide the initial basis for a

community-based conservation and management initiative. The justification for

endorsing this environmental strategy is summarized in Chapter Six, which is organized

around five major themes: 1.) the need for turtle conservation; 2.) the role of protected

areas and wildlife use; 3.) common property resource use and community-based

management; 4.) participatory possibilities; and, 5.) a roadmap for community-based

conservation and management in Piso Firme and Remanso.






4


Chapter Seven synthesizes results of the different research topics and offers

conclusions based on my findings. Future research is suggested.














CHAPTER 2
TURTLES IN AMAZONIA

History of Turtle Use

For centuries, humans in the Amazon and Orinoco River Basins have exploited

freshwater turtles of the genus Podocnemis. Records from both river basins relayed by

early explorers (including Bates [1863] 1962; Bozo [1816] 1991; Herndon [1854] 1952;

Humboldt 1852; Keller 1874; Le Cointe 1945; Maw [1831] 1989; Smyth and Lowe

[1836] 1973; Spix and Martius 1968; and Spruce 1908) cite the abundance ofP. unifilis

and P. expansa and their extensive harvest by local inhabitants particularly during nesting

season. Anthropological evidence for turtle use by humans dates to pre-Hispanic times

(Castro de Le6n 1986; Licata 1992) with the earliest written records of human use of

Podocnemis unifilis (Paez and Bock 1998) and P. expansa (Castro de Le6n 1986) dated

prior to the 16th century. Illustrations contributed by early explorers such as Marcoy

(1875) support historic records that cite the utility of Podocnemis turtle eggs to

indigenous, colonist and mestizo people throughout the Amazon Basin (Castro de Le6n

1986; Robinson and Redford 1991) (Figures 2-1 and 2-2). Prior to the introduction of

kerosene and vegetable oil substitutes, P. expansa eggs provided the raw material for

cooking and lighting oil production. An estimated 48 million eggs, likely to have

represented the nesting efforts of thousands of turtles annually, were used to make oil in

the 19th century (Bates 1962; Smith 1974). Some of these early explorers quantified the

harvests of turtles and their eggs, noting an apparent decline in their abundance over the

course of time. One of these explorers, Bates, noted that









At least 6,000 jars, holding each three gallons of the oil, are exported annually from
the Upper Amazons and the Madeira to Para, where it is used for lighting, frying
fish, and other purposes. It may be fairly estimated that 2000 more jars-full are
consumed by the inhabitants of the villages on the river. Now, it takes at least
twelve basketfull of eggs, or about 6000, by the wasteful process followed, to make
one jar of oil. The total number of eggs annually destroyed amounts, therefore, to
48,000,000. As each turtle lays about 120, it follows that the yearly offspring of
400,000 turtles is thus annihilated. A vast number, nevertheless, remain
undetected; and these would probably be sufficient to keep the turtle population of
these rivers up to the mark, if the people did not follow the wasteful practice of
lying in wait for the newly-hatched young, and collecting them by thousands for
eating; their tender flesh and the remains of yolk in their entrails being considered a
great delicacy.... The universal opinion of the settlers on the Upper Amazons is,
that the turtle has very greatly decreased in numbers, and is still annually
decreasing. (1863: 364-365).


Figure 2-1. Conibo Indians along Ucayali in 1866 Turtle eggs being crushed to extract
oil (Marcoy. 1875. Travels in .n,,ml America: From the Pacific Ocean to the
Atlantic Ocean. Vol II. New York. Scribner, Armstrong and Co.: 35).

Writing even earlier, Herndon noted

prolificc as they are, I think the turtle is even now diminishing in number on the
Amazon. Vast numbers of the young are eaten by the Indians, who take them by









the time they are able to crawl and when they do not measure more than an inch in
diameter, boil them, and eat them as a delicacy (1854: 157-8).

These passages, among the earliest written records of Podocnemis turtles, mention

their use and possible vulnerability in a setting shared with humans dating back at least

two centuries. This pressure on turtles has likely increased as human populations in the

Amazon Basin have grown (Allen and Barnes 1985; Darwin et al. 1996; Godoy 2001).

As European influence spread in the Amazon and Orinoco River basins, so too did

large-scale harvest of Podocnemis turtle eggs for oil and turtles for protein by I il,' eeli,,

as depicted by Keller (1874) in Figure 2-2. Since the 16th century, these turtles have

exhibited declining population trends similar to those of other threatened and endangered

wildlife, mainly believed to be the result of overhunting (Alho 1985; Ergueta & de

Morales 1996; Escalona & Fa 1998; Johns 1987; Luxmoore et al. 1988; Salati et al.

-- -


Figure 2-2. Turtle being cooked on Amazon River (Keller. 1874. The Amazon and
Madeira Rivers: Sketches and Descriptions from the Notebook of an Explorer.
London. Chapman and Hall: 74.).









1990). Local residents and scientists have noted significant declines in populations of

these turtles throughout their ranges (Alho 1985; Ergueta and de Morales 1996; Escalona

and Fa 1998; Johns1987; Luxmoore et al. 1988).

Turtles in the River Ecosystem

River turtles, including P. unifilis and P. expansa, inhabit lowland higher-order

sections of rivers characterized by broad, deep channels and rapid water velocity (Moll

and Moll 2000). In the Bolivian Amazon, as in other areas of Amazonia, turtles share

these habitats with humans. People are attracted to rivers for their abundant water

supply, food resources, agricultural foundation and transportation (Brondizio et al. 1994;

Denevan 1996; Hiraoka 1985; Moll and Moll 2000; Moran 1991).

River turtles are thought to contribute a significant proportion of biomass to river

ecosystems and through their scavenging provide a valuable contribution to the cycling of

nutrients (Moll and Moll 2000). Due to their sensitivity to environmental change, P.

unifilis and P. expansa are also indicators of potentially negative ecosystem changes (e.g.

pollutants that harm ecosystem function) (Oficina Regional de la FAO para America

Latina y el Caribe 1988).

In the rainy season, P. expansa, a frugivore, enters the flooded forests where it

finds abundant food in the form of fruits and seeds from trees (Moll and Jansen 1995;

Ojasti 1971; Smith 1999). This behavior is significant because it contributes to "the

relationship of the tropical forest fauna on tree diversity" (Wilcox 1995: 45). Multiple

studies show Podocnemis species disperse seeds in the flooded forests of Amazonia

(Bruno Coca 1999; Fachin Teran 1992; Kubitzki and Ziburski 1994; Ojasti 1971).

Moll and Jansen (1995) specifically found that turtles play a role in the dispersal of

plants, by acting as seed dispersers, secondary to birds and monkeys. In their study of









aquatic black wood turtle (Rhinoclemmysfunerea) and terrestrial brown wood turtle (R.

annulata) turtles at Tortuguero, Costa Rica, Moll and Jansen (1995) found the majority of

seeds in turtle stomachs to be from riparian, wetland and second-growth plant species

commonly found in and adjacent to canals and streams. Germination of eaten versus

uneaten seeds was not significantly different-evidence for the role of turtles in the

dispersal of plants (Moll and Jansen 1995). "[T]he potential for this [P. expansa] and

other frugivorous turtle species (e.g., P. unifilis) to act as seed dispersers (with fish) in

these cyclically flooded, forest ecosystems seems especially promising and warrants

further investigation" (Moll and Jensen 1995: 126). A benefit to flooded forests of

frugivorous turtles is that the low concentration of oxygen in the stagnant water does not

negatively affect the turtles because they get their oxygen from the air. In fact, this

characteristic confers an advantage to turtles, by reducing the concentrations of

frugivorous fish that compete with them for fruit (Ojasti 1971). This biological attribute

enhances the mutually beneficial relationship between turtles and tropical forest diversity.

Frugivory in P. expansa has been reported throughout most of its range (e.g. Bruno

Coca 1999 in Bolivia; Fachin Teran 1992 and Goulding 1980 in Brazil; Ojasti 1971 in

Venezuela; Soini 1995b in Peru). Ojasti (1971) examined the stomach contents often

adult P. expansa turtles from the Orinoco River and found that 86% of the ingested

material was fruit from trees, 4% was leaves and stalks of diverse plants and the

remainder was aquatic and terrestrial invertebrates, bones from cattle, fish and turtles in

insignificant amounts. In Brazil, Fachin Teran et al. (1995) found that P. unifilis ate

more seeds and fruits from the flooded forest than from rivers and lakes. In Bolivia,

Bruno Coca (1999) showed that the diet of P. expansa is largely composed of fruits and









seeds from the forest while P. unifilis ingests fruits and seeds more as a complement to its

diet, which is mainly composed of leaves and stalks of terrestrial and aquatic plants.

Bruno Coca's study (1999) suggests that P. unifilis and P. expansa in and around Parque

Nacional Noel Kempff Mercado (the Park) act consistently with their conspecifics in

other parts of the Amazon Basin as seed dispersers. These studies in Costa Rica, Brazil,

Bolivia, Peru and Venezuela suggest that river turtles, including Podocnemis sp., depend

upon access to the flooded forests and are important links in the life cycles of trees and

other vegetation in the forest floodplains.

Implications of Large Turtle Life History for Conservation and Management

Some general life history traits of large turtles are thought to reduce their

population resilience in the presence of hunting pressure-the main threat that

accompanies sharing habitat with humans. Freshwater turtles like P. unifilis and P.

expansa are long-lived species characterized by low adult mortality and high losses at

both embryonic and hatchling stages (Gibbons 1968). Turtle populations are further

threatened because nesting beaches overlap with human settlements, which has resulted

in P. unifilis and P. expansa being particularly vulnerable to hunting and their

populations reduced throughout most of their ranges (Cantarelli 1997; Ergueta and de

Morales 1996; Johns 1987; Luxmoore et al. 1988; Mittermeier 1978; Pritchard and

Trebbau 1984; Shaw 1991; Smith 1974). Because of the annual nesting season of

Podocnemis turtles, populations are under heavy pressure for approximately 120 days

each year, depending on location (June to September in my research area) (Pritchard and

Trebbau 1984).

The main threats to eggs are flooding and human predation (Escalona and Fa 1998;

Luxmoore et al. 1988). If eggs survive their incubation period, upon the hatchlings'









emergence from the nest a host of avian, amphibian and reptilian predators await (Alho

1985). The most serious predators however, are humans. In a study ofP. unifilis nest

survival on the Nichare and Tawadu Rivers of Venezuela, Escalona and Fa (1998)

reported that human collection of eggs affected 84.9% of nests (N = 351), with removal

of all eggs from 70.6% of the human-disturbed nests (animals and flooding also

destroyed nests, but to a much lesser extent). As a colonial nester (large numbers of

turtles gather concurrently on particular beaches over several days to lay their eggs) P.

expansa is even more susceptible to large-scale nest excavation, which can lead to higher

percentages of nests being destroyed.

In population ecology, five factors influence the rate of reproduction. These factors

are clutch size; number of clutches per year; minimum and maximum breeding ages; sex

ratios; and density effects. The harvest of nesting adult females is especially harmful

because it poses a significant threat to the turtle populations' survival by removing key

reproducers (Congdon et al. 1993). P. expansa's nesting-style vulnerability combined

with the sizeable amount of meat they offer makes them particularly attractive prey for

humans. Although P. unifilis exhibits different nesting site preferences and is not a

colonial nester, humans have learned and pass on the necessary insight to detect their

nests efficiently and these turtles are also threatened. The combination of a slow

reproductive rate, ease of capture-particularly of reproductive females during nesting

season-and historic harvest pressure for subsistence and market consumption has

decimated turtle populations throughout the Amazon Basin (Johns 1987; Mittermeier

1978; Pritchard and Trebbau 1984). These factors negatively influence the reproductive

efforts of freshwater turtles of the genus Podocnemis found in eastern lowland Bolivia in









the rivers adjacent to the communities of Piso Firme and Remanso. Greater market

integration and increased demand for turtle meat may make populations even more

susceptible (Campbell 1998) and are considered in my study.

Legislative Stopgaps

Multiple forms of legislation have been drafted and passed in attempts to reduce

hunting pressure on these turtles. Both P. unifilis and P. expansa are listed on the

Convention on International Traded in Endangered Species (CITES) Appendix II, which

legislates regulation and control of international trade for their protection. Additionally,

both species are rated "one" on the Tortoise and Freshwater Turtle Specialist Group

action plan rating, a priority conservation-rating plan indicating that they are "in trouble

through over-exploitation for eggs and meat, and some are particularly vulnerable at

colonial nesting sites" (Tortoise and Freshwater Turtle Specialist Group 1991: 10).

Brazilian law prohibits hunting of any wildlife found in its territory for sale nationally or

internationally (Article 1 Law 5,197 January 3, 1967 and Article 29 of Law 9,605

February 12, 1998). While these laws and treaties exist, use that threatens turtle

existence continues.

Value of Wildlife, Including Turtles

The forests and rivers of Amazonia contain tremendous faunal and floral diversity,

including many species that are important resources for humans (Redford and Padoch

1992; Smith 1999; Terborgh 1992; Wilson 1988). Among the important resources are

river turtles (Aramayo Cuenca 1989; Moll and Moll 2000).

Surveys assessing the consumption of wildlife have cited the importance of wildlife

as food resources to humans (Bennett 2002; Edwards and Abivardi 1998; Fa et al. 1995;

Funes and Novaro 1999; Robinson and Redford 1991; Santos Brito and Ferreira 1978).









Santos Brito and Ferreira (1978) quantified Amazonian wildlife consumption in Manaus

in 1976 and seven additional cities in 1977. Their study identified species preferred in

restaurants, suggesting the pressure facing animals in the wild. Both P. unifilis and P.

expansa were mentioned by participants in these studies as being desirable (Santos Brito

and Ferreira 1978). Sale of bushmeat, which was outlawed in Brazil in 1967 and is

restricted within and near the Park, continues today (e.g. Johns 1987 and Conway 1999-

2001 personal observation). This translates into hunting pressure that is greater than just

subsistence hunting pressure alone.

In his work on Amazon River turtles in Tefe Johns (1987) noted that during the

1985 season, the average market price per turtle for a P. unifilis was US$11-14 but

dropped to US$7 if several boats arrived in port simultaneously. Podocnemis expansa

was reputed to have a less desired taste among / il ,i ei),s but was more highly prized in

the city where individual turtles sold for US$60 (range $40-$130) in the market while a

portion in a restaurant was US$13-$20 (Johns 1987: 26).

The giant South American river turtle (P. expansa) is unusual among bushmeat

because it tends to be more expensive than domestic meat (Ojasti 1996). In 1978, the

market price for P. unifilis and P. expansa meat in seven Brazilian cities was US$0.91

per kilogram while servings in a restaurant were US$7.83 and US$9.56, respectively in

1978 US$ (Chinese University of Hong Kong 2000 for the currency conversion; Santos

Brito and Ferreira 1978 data).

Value of Podocnemis Turtles in Eastern Lowland Bolivia

Podocnemis turtles are large, widely distributed and provide a sizeable amount of

high quality protein (Pritchard and Trebbau 1984). The largest recorded length and

weight of a P. expansa specimen is 89 cm and 45 kg (Luxmoore et al. 1988; Pritchard









and Trebbau 1984). Turtles of this size can provide 10 to 12 kg of high-protein meat

(Table 2-1). In terms of protein content, Cantarelli (1997) states that turtle meat contains

85 to 88% protein vs. 43 to 70% for domestic meats (pig, chicken, and beef). Rilk' eii),

(floodplain residents) in eastern lowland Bolivia state a preference for turtle meat over all

domestic meats and claim it offers more nutritional value (various residents personal

comm. 1999, 2000-2001)

Preliminary fieldwork in three Bolivian communities (Piso Firme on the Paragua

River and Remanso and Bella Vista on the Itenez River) in August 1999 confirmed that

/ il,'ie i,),s there consume P. unifilis and P. expansa meat and eggs. Observation of

middens and interviews with 36 individuals in the two larger communities (Piso Firme

and Remanso) indicated that turtle meat and eggs are sources of both meat and cash/trade

value. The meat and cash/trade values ofP. unifilis and P. expansa appeared to influence

livelihood strategies in this area, which may have long-term effects on the humans, the

turtles and the natural systems.

Table 2-1. Average values of carapace length and body weight of female Podocnemis
expansa turtles in the Orinoco and Amazon River Basins, an indication of
their value as protein resources
Source Year Location Average Average N
Carapace Body
Length (cm) Weight
(kg)
Ojasti (1971) Early Orinoco River, 23.3 100
1960's Venezuela
Ojasti (Pritchard and 1962 Playa del Medio, 63.1 Unknown
Trebbau 1984) Estado Apure,
Venezuela
Ojasti (Pritchard and 1964 & Playa del Medio, 64.1 & 64.0 Unknown
Trebbau 1984) 1966 Estado Apure,
Venezuela
Vanzolini (1967) 1967 Trombetas River, 70.1 38
Brazil
Pritchard (Pritchard 1977 Playa del Medio, 65.1 20
and Trebbau 1984) Estado Apure,
Venezuela









Pritchard (Pritchard 1977 Playa Pararuma, 66.2 20
and Trebbau 1984) Estado Apure,
Venezuela
Paollilo (Pritchard and 1981 Playa del Medio, 66.8 26.0 302
Trebbau 1984) Estado Apure,
Venezuela
Conway Present Itenez/Guapore and 59 22 8
work Paragua Rivers, (sd=10.67) (sd=8.77)
Bolivia

Included in Table 2-1 are the results of a sample of eight (for length) and six (for

weight) P. expansa turtles captured on the Itenez and Paragua rivers in eastern Bolivia. I

was able to weigh the components of one -28 kg specimen of P. expansa as it was

slaughtered for consumption. A 28-kg female P. expansa caught and weighed as it was

slaughtered was constituted as follows: carapace (dorsal shell) 5 kg; plastron (ventral

shell) 11 kg; organs 1.5 kg (including eggs yolks in the uterus); and, meat 10 kg.

Although the turtles I measured were smaller on average they do, nonetheless, provide a

meat resource for local people, who highly value it.

Podocnemis unifilis, the yellow-headed sideneck turtle, shares much of the same

range with P. expansa, extending somewhat further south (Iverson 1992; Pritchard and

Trebbau 1984). It is smaller with a mean carapace length that ranges from 34.7 to 46.3

cm (Pritchard and Trebbau 1984). Reported weights of two female P. unifilis specimens

weighed by Pritchard were 8.1 and 9.0 kg (Pritchard and Trebbau 1984). A sample often

P. unifilis measured during my fieldwork in 2000-2001 yielded an average carapace

length (measured longitudinally/in a straight line from anterior to posterior end) of 38 cm

(sd=7.90) and an average weight of 7 kg (sd=1.53). Because P. unifilis is much more

locally abundant than P. expansa (Pritchard and Trebbau 1984) the yellow-headed

sideneck turtle is a more frequent source of protein for ribereho communities in lowland

eastern Bolivia today.









Price of River Turtles in Lowland Bolivia

In 2000-01 whole P. unifilis in rural eastern lowland Bolivia averaged US$3.25. In

2000-01 the price for a female P. expansa in the same area ranged from US$8.13 to

US$16.26 (50-80 bolivianos) for a 20 to 30 kg turtle, which is similar to Ojasti (1996)

and Johns (1987). Since ilei 'ei,,s in general also have limited access to cash-generating

income, their preference for turtles may be a result of economic scarcity as much as

historic and cultural preferences.

Based on the measurements of a female P. expansa in Bolivia, a yield of 41%

comestible product is possible. Extrapolating from a sample often female P. unifilis

captured in my study area for local consumption that weighed an average of 7 kg

(sd=1.45), a yield of 3 kg of meat each may be expected. If a 7-kg P. unifilis turtle

yielded 3 kg of edible meat, each kilogram would cost roughly the equivalent of US$1.

In 2000-01 the prices of meat in my study area were: beef US$1/kg; chicken US$1/kg;

and fish US$ 0.75/kg. Daily minimum wage during this time was US$5.00. Thus, P.

unifilis turtle meat is valued roughly equivalent to the prices of domestic meat (beef and

chicken). The average daily meat consumption based on a purposive sample taken from

November 2000 to May 2001 was 1.64 kg (sd=1.206 n=12) in Piso Firme and 1.23 kg

(sd=5.157 n=17) in Remanso. These figures indicate a daily range ofUS$1.23 to

US$1.64 in Piso Firme and US$0.92 to US$1.23 in Remanso potentially spent on meat.

If households buy the meat consumed they will spend anywhere from 18 to 33% of daily

wages on this commodity. Relatively speaking this is a high percentage, which suggests

the economic and cultural worth of meat in these communities.









Markets as they Apply to Turtles

Von Thunen ([1826] 1966) described the correlation between distance to market

and intensification of agricultural ecosystems. All other things being equal, the intensity

of production depends upon the price a farmer gets for a product, which is directly

dependent on transport costs-a function of market distance (Von Thunen [1826] 1966).

Von Thunen's intensity theory posits that sites nearer the market, which have higher

prices, should be cultivated/harvested more intensely (Von Thunen [1826] 1966). There

is a distance at which diminishing returns dictate that intensive production is no longer

cost effective.

In 1999 in the northeastern Brazilian Amazon, residents in a community on the Rio

Iriri, a tributary of the Xingu River in the Brazilian state of Para, earned US$5 per P.

expansa turtle from middlemen, who then sold the turtles in Manaus for a profit (R. Porro

personal communication 1999). Reports of market prices for P. expansa in urban

Amazon markets range from $100 per turtle (Smith 1999) to $150 per turtle in Manaus

(Luxmoore et al. 1988). These figures suggest that the $5/turtle earned by I il 'i ,ei,\ may

have been worth more to fishers than the meat provided, which suggests turtles are a

cash-crop in emerging rural markets. It may also indicate market pressure on the turtles

if each ribereho was capturing and selling multiple turtles. Most importantly, it points

out the ongoing cash and market value of Podocnemis turtles.

Influence of the Market on Neotropical Wildlife

In addition to their monetary value, Podocnemis turtles are attractive from a market

perspective because they stay alive for days without food or water. This is important in

rural communities without refrigeration, for fishers on multi-day fishing trips and for

shipping turtles to cities on the bus. Also, turtles are a luxury item for emigrants to urban









areas who live more than a day's journey away from their rural homes but maintain

cultural and dietary preferences from home.

Both positive and negative benefits for wildlife conservation and income-earning

power result from increased value of species. Wildlife with high market value may face

greater hunting pressure, which would in turn cause greater threats to their existence.

Conversely, increased awareness, when accompanied by management or captive breeding

of prized species, are positive outcomes for conservation that may result from market

incentives (Bodmer and Puertas 2001; Bodmer and Pezo 2001). When done in a

sustainable fashion (i.e. not heavily targeting both reproductive females and new

recruits), harvest of wild species can offer benefits for human and wildlife species (Freese

1997b).

Freese notes freshwatertr species that have been depleted during the twentieth

century include several species of crocodilians (Messel et al. 1992), valued for their

hides, and Amazonian river turtles (Podocnemis spp.) (Alho 1985), valued for their meat"

(1998: 26). In another study, Ojasti (1996) cites turtles (specifically Podocnemis spp.) as

being the reptiles most heavily affected by subsistence hunting in the neotropics.

Assessing the market value of and impact on turtles as natural resources is important in

analyses preceding conservation and management planning because of the history of use

and the evidence that market access can have both negative and positive ramifications on

natural resources (Freese 1997a, Godoy 2001, Johns 1987, Klemens and Thorbjarnarson

1995).

Negative Impacts of Markets

Market access has often been associated with increased harvest of goods ranging

from agricultural to nontimber forest products (Godoy 2001; Godoy et al. 1998a; Redford









and Robinson 1985; Sierra etal. 1999; Von Thinen [1826] 1966). Multiple studies

around the world have pointed out negative effects of hunting practices on wildlife

(Alvard 1994; Campbell 1998; Martin 1978; Ministerio de Agricultura y Cria 1972;

Murray 2003; Redford and Robinson 1991; Terborgh and Van Schaik 2002). Redford

and Robinson (1985) suggest that market integration reduces the sustainability of hunting

by indigenous groups because of the introduction of goods and technologies that

accompany integration. These changes can interrupt ways of life through demands for

"the trappings of modem civilization [that] can cause or increase the over hunting of

certain species" (Redford and Robinson 1985: 43, see also Fa etal. 1995).

Hunting wildlife to extirpation negatively effects human populations that depend

on it and the resource must be replaced if people are to maintain familiar lifestyles. Of

course, removal of a faunal or floral component from an ecosystem can disrupt the

functioning of natural systems as well (Gunderson and Holling 2002; Odum 1993).

This is not to say that markets, technology or development should not be

introduced, but it is a cautionary note on the potential effect of such factors on cultural

practices that have ramifications for wildlife. For, as Bebbington (1996) points out in

reference to the introduction of new tools and technologies for indigenous development,

important components for improving lives without compromising social and political

systems include restructuring social relationships and putting power into the hands of the

people so they have the ability to increase their incomes. So, it is possible that shifts in

income sources and empowerment can reduce pressure on wildlife and therefore be

desirable ends of markets, development and technology.









Positive Effects of Markets

Markets have traditionally been observed as negative forces on natural resources,

but research now indicates that there are certain conditions under which markets do not

pose greater negative consequences and are even beneficial to natural resources (Agrawal

1995; Agrawal 2001; Angelsen 1999; DolSak and Ostrom 2003). Agrawal and Yadama

(1997), working in forests in India, surmise that with the presence of paved roads, the

market is brought closer because transportation costs are reduced, which then increases

villagers' awareness of the forest value and in turn their desire to conserve. Increased

awareness and active conservation of resources can occur with greater market access

(Agrawal and Yadama 1997; Anderson 1990; Godoy 2001; Schweik et al. 2003; Tang

and Tang 2001).

Market Integration and Hunting Intensity-Effects on People

Bodmer and Pezo (2001) note a disjunction between rural development schemes

and sustainable wildlife use. In their view, rural people (as opposed to urban dwellers or

international market players) suffer the greatest economic costs by converting

unsustainable wildlife hunting practices into more sustainable practices. This is in direct

conflict with a main objective of most rural development schemes, which strive to

increase economic development (Coomes 1996; Yapa 1998). For rural people to hunt

more sustainably, they must be able and willing to absorb the costs of change. This

means that rural development projects must increase short-term economic income

potential for the gain of long-term sustainable wildlife harvest. In cases where systems of

use are demonstrated to be unsustainable, conservation plans must incorporate

concessions or alternatives for the people in the short-term, and wildlife survival in the

long-term.









Bodmer and Pezo (2001) suggest that economic benefits may be gained in the form

of compensation payments in lieu of hunting. To realistically establish payment plans

that will offer some form of long-term relief from wildlife exploitation pressure, many

factors-such as origin and handling of funds and potential social effects of introduction

of these funds-must be taken into consideration. Additional payment options may

include greater service employment opportunities (e.g. ecotourism) or alternative

subsistence and market crops. A discussion of the amounts and effects of payments is

beyond the scope of this work, it is mentioned here as one alternative for conservation.

Influence of Income on Hunting

Scholars of non-timber forest product extraction in Central America conclude that

as market integration continues, bringing with it higher income-earning capacity,

foraging for animals in these rainforests increases (Godoy et al. 1995). Hunting can and

does cause wildlife populations to decline or disappear from areas (Martin 1978; Murray

2003; Peres and Lake 2003; Robinson and Redford 1994; Shaw 1991). Podocnemis

turtles used by / ilk'i eii, in the Bolivian Amazon are no exception to this phenomenon

(Aramayo Cuenca 1989; Quiroga Vera 2000; Urefia Aranda 2000). Godoy et al. (1998)

found that in some cases involving sale of non-timber forest products, rural dwellers'

knowledge of animals' natural history increases as commercial values rise, an offshoot

that may benefit wildlife.

In his work in Bolivian and Honduran tropical rainforests, Godoy (2001) expands

on the consequences of increased market integration on wildlife use noting that as income

increases, economic development will increase consumption of most game and generally

will lead to increases in human populations. These two factors-economic development

and human population growth-can work synergistically to greatly increase pressure on









wildlife through market availability of game meat. Redford and Robinson (1985) and

Robinson and Redford (1994) found that market integration acted as a deteriorating factor

in the sustainability of hunting by indigenous groups with the giant river turtle

specifically noted. However, Godoy (2001) also notes that in some cases economic

development will lower the price of game meat substitutes through commercial

production, thereby reducing demand for game and thus hunting pressure.

In his examination of market prices of wild-caught meat Godoy applies three

categories. These categories are defined as: inferior goods-their consumption falls as

income rises, especially in households whose income is below the median; necessities-

their consumption goes up less than 1% with each 1% rise in income in bottom-income

categories and becomes and inferior good in the top income category; and, superior

goods-their consumption increases more than 1% with each percent rise in income

(Godoy 2001). According to Godoy's definitions and observations, fish is an inferior

good and game meat is a necessity (Godoy 2001). "The evidence suggests that growth in

income probably discourages the consumption of fish and increases the consumption of

game meat, but at low rates"(Godoy 2001: 94).

Whether Podocnemis turtle meat is analogous to fish or game meat is unknown-

they are aquatic and terrestrial prey that provide greater catch per unit effort than most of

the common fish in the area, but their capture requires some specific knowledge (which

areas of the river/flooded forest to look in and best times for capture) and techniques

(how and where to place baited hooks, how to use harpoons, where to wait on nesting

beaches and when to approach). The specific knowledge and techniques that facilitate

harvest add to the cost, which could serve to raise the price of this commodity. Higher









catch per unit effort combined with the demand for the turtle meat and evidence for its

consumption in San Ignacio (the nearest Bolivian city), at prices greater than in the

communities, suggest turtle is a superior good. If turtle meat is a superior good, current

consumption threats may be exacerbated with augmented market integration. If turtle

meat is an inferior good, which implies it is less expensive, the stress on turtle

populations from hunting will be reduced if incomes increase and other meats are more

heavily consumed. This could protect turtles, barring other stresses. But, if turtle meat is

a necessity and consumption increases at rates just below income increases and turtle

population growth, it will continue to be threatened because consumption will continue to

rise as incomes rise with growing availability of cash through more market integration.

My assumption is that either through improved road access to San Ignacio or more

involvement in cash-generating employment, these communities, especially Piso Firme,

are bound to be more market integrated in the future.

My observations during preliminary fieldwork suggested that turtle meat was an

inferior good for households with greater access to economic resources. However,

increased market integration could make it a necessity or a superior good. Hypotheses

four and five were set up based on the relationships between wealth, market access and

turtle consumption.

Piso Firme and Remanso have ties to the markets of San Ignacio and Santa Cruz de

la Sierra, Bolivia and Pimenteiras, Brazil. In general, residents of Piso Firme rely on

goods coming from San Ignacio while Remansefios rely more heavily on goods

purchased in Pimenteiras, Brazil. Since at least 1999 there has been talk of improving the

road that connects the communities to San Ignacio, which would reduce transportation









costs (currently it is a 12 to 24 hour bus trip, depending on the season). Reduced

transport costs may affect a range of phenomena-from market access to migration

patterns to access to cash. Such changes may influence turtle consumption patterns,

which is why I emphasize the impact of the market here, using distance as a proxy

indicating market integration.

There are presently several cash-generating activities in my research area-fishing

with Brazilian commercial fleets, mining for gold, and cocaine running and other illicit

goods-that boost the facility of purchasing goods. These activities in all likelihood will

continue to exist until the resources are depleted, as in the case of the fisheries and

mining, or the activity is deemed unworthy of the risk, as in the case of cocaine and other

illicit goods.

In the case of cocaine, Castells and Laserna point out that "short of massive

chemical defoliation, which could trigger an ecological catastrophe and a cultural war,

coca production cannot be eradicated" (1994: 76-77), which has direct effects on the

continuance of cocaine running that is a source of income for several households in each

community. In the case of other illicit goods, rural frontier settlements like Remanso are

prime locations for the transport and resulting financial rewards of such practices. The

impact of disposable cash on turtles could be increased or sustained levels of

consumption if turtle is a superior good or a necessity. Conversely, increased cash

availability could have a positive effect on turtle populations if it is an inferior good and

consumption declined with growing incomes.

Because road improvement is a community and Park priority, the effect of the

market on P. unifilis and P. expansa in the Bolivian Amazon must be considered because









a better road will change the lifestyles of these / ili ei,",s through an improved linkage

with the city and its markets. Ayres et al. (1991) had an opportunity to observe wildlife

harvest rates in a Brazilian Amazon village pre- (1978) and post-construction (1980) of a

road. In the case of Dardanelos, hunting yields and number of hunters decreased after the

construction of the road, suggesting reduced hunting pressure (Ayres et al. 1991).

Furthermore, the road into Dardanelos was associated with reduced consumption of game

meat and increased consumption of domestic meat (Ayres et al. 1991), which suggests

that game meat was an inferior good whose consumption declined with greater incomes.

These are positive impacts of a road for wildlife. But a road also represents greater

market access and other changes. So, while there can be positive effects for wildlife, at

least as suggested in the two years immediately following construction, the longer term

effects associated with road construction and their impacts on wildlife must be

considered. As Ayres et al. (1991) pointed out, the road brought with it considerable

sociocultural change because of integration into the national economy. "Despite the fact

that hunting became less important for human subsistence, the integration of Dardanelos

into the national market economy by the road brought other threats to the regional

wildlife, including logging, cattle ranching, mining, and large-scale agriculture" (Ayres et

al. 1991: 92). Particularly because some of these activities already exist on a small-scale

in Piso Firme and Remanso this note is cause for concern.

Contemporary examples ofP. unifilis and P. expansa use indicate that their meat

and eggs hold market value for the people in the Amazon floodplain, both at the local

scale observed in this research as well as at the larger Amazon Basin scale (as noted in

Cavalcanti 1999). Figures from the greater Amazon Basin and my data from Bolivia, are









evidence that turtles remain a sought-after resource at the local scale. Demand for turtles

in the Amazon Basin (Cavalcanti 1999; Johns 1987; Luxmoore et al. 1988; Mittermeier

1978) coupled with growing human populations foretell continued and increased stress

on the species. If international trade demands rise or the level of market integration in

rural communities such as Piso Firme and Remanso grows, the negative effects on the

turtles are sure to grow with them.

Theoretical Derivation of Hypotheses

My research questions in eastern lowland Bolivia were derived from two areas of

inquiry-ecological and social (Table 2-2). First, I tested spatial relationships between

humans and wildlife abundance and population structure based on wildlife ecology.

Previous studies on wildlife use have shown contrasting impacts of hunting on wildlife.

In many cases, hunting is viewed as detrimental to wildlife populations (Hames 1987;

Martin 1978; Murray 2003; Peres and Lake 2003; Redford and Robinson 1985 and 1991;

Robinson and Redford 1994). Yet, other studies have found that hunting does not always

lead to negative impacts on wildlife species (Bodmer 1995; Bodmer et al. 1997; Chernela

1994; Fitzgerald 1994; Francis 1997; Vickers 1991). Furthermore, some studies indicate

that differences in wildlife population structures may indicate hunting pressure and

unsustainable use (Congdon et al. 1993; Robinson and Bodmer 1999). I wanted to see if

similar patterns could be detected in river turtle populations in eastern lowland Bolivia

and examine socioeconomic factors that might impact the effects of market integration on

turtle use.

Multiple studies cite evidence that suggests market integration contributes to

decreased sustainability of hunting (Alvard 1994; Campbell 1998; Martin 1978;

Ministerio de Agricultura y Cria 1972; Murray 2003; Redford and Robinson 1994;









Terborgh and Van Schaik 2002; Robinson and Redford 1985). Measuring abundance is

one relatively simple and quantifiable means of assessing sustainability when a

comparison is made between hunted and less-hunted sites. The comparative assessment

of suspected hunting impacts on turtles can provide data that corroborate market impacts.

If population abundance is observed to be low around human settlements, this is

preliminary evidence of potentially negative impacts of hunting.

Like assessment of abundance, observation of turtle size in hunted vs. less hunted

sites can suggest impacts of hunting pressure on the wildlife resource. Quantification of

reptile population structure offers information about the maturity of the population, which

gives an indication of recruitment and longevity (Caughley 1977). Noting the value of

age structure for populations, Cole (1954) stated whenhn the mortality factors affecting a

population are altered either through natural environmental changes or through human

exploitation or attempts at control there will in general result a change in the age structure

of their population, and this may be observable even before changes in population size or

in birth rates provide evidence of the consequences of the changed mortality factors" (28-

9).

The second component of my research examined relationships between

socioeconomic factors affecting market integration and its impact on wildlife use. I argue

that the relationship between distance to market and intensity of turtle harvest is similar

to the one Von Thtinen described for agricultural goods-increasing intensity of turtle

harvest will coincide with closer market access (as suggested by Godoy et al. 1995).

This relationship may be caused directly by increased value of meat or indirectly through









increased demand for cash acquired through the sale of meat. Multiple socioeconomic

factors contribute to the actual impact of the market on wildlife.

Market integration can lead to increased harvest of natural resources, including

wildlife (Godoy 2001; Godoy et al. 1995 and 1998a; Santos Brito and Ferreira 1978;

Sierra et al. 1999; Von Thunen [1826] 1966). Specifically I wanted to examine the

relationship between market integration and river turtle use in an eastern Bolivian context

where there are different levels of market integration. Quantifying turtle consumption at

sites with different ties to markets is one means of assessing the market impact on a

wildlife commodity. The comparison of turtle consumption between Piso Firme and

Remanso offers a chance to determine what the market impact on turtles is in these

settings. Comparing several additional socioeconomic characteristics provides an

opportunity to add further insight into turtle consumption. This additional information

may offer insight into the driving forces of turtle consumption if market influence does

not prove a main force affecting consumption.

Income has varying effects on wildlife meat consumption (Godoy 2001).

Perception of wealth, as a proxy for income, is one measure of potential impact on turtle

harvest. Quantifying this relationship offers an opportunity to better understand factors

driving turtle consumption and ultimately informs conservation planning.

The socioeconomic characteristics examined were chosen based on the preceding

review of literature on markets and natural resources, prior knowledge of the

communities from a preliminary site visit and observations during my field research that

suggested characteristics that directly or indirectly influence turtle consumption. A









summary of the questions that drove my research and the specific hypotheses to test these

questions are stated below.

Table 2-2. Research questions addressed in this research and the hypotheses used to test
the questions.
Question Addressed Hypothesis Tested
What is the spatial relationship between HI: turtles will be less abundant closer to
human communities and Podocnemis human communities
unifilis and P. expansa turtle H2: there will be more turtles at
abundance? Mangabalito (less hunting pressure) than
at Remanso (more hunting pressure)
What is the spatial relationship between H3: there will be fewer large turtles closer
location of human communities and size to human communities than farther
classes noted ofPodocnemis unifilis and away
P. expansa turtles?
How does distance to market affect H4: there is greater market consumption of
subsistence vs. market consumption of turtle protein in Remanso than Piso
turtle? Firme and greater subsistence
consumption of turtle protein in Piso
Firme than in Remanso
How does household wealth affect turtle H5: there is a negative correlation between
consumption? wealth and turtle protein sale/trade such
that as personal wealth increases, the
consumption of turtle protein decreases


Implications of Findings

Fewer turtles near human communities is one indication of hunting pressure as a

depressing element on turtle populations. The implication here is that this effect can

continue spreading out away from communities and threatening even greater portions of

turtle populations, as has been observed in other parts of the Amazon Basin. Extinction

is the ultimate threat. Rejection of null hypotheses one and two are indicators of negative

impacts of hunting pressure.

Fewer large turtles closer to humans imply negative effects of hunting on

population maturity. Populations that are not able to maintain sufficient recruitment as

measured by mature females, will eventually disappear. Rejection of null hypothesis









three is another potential indicator of negative hunting pressure on turtles in my research

area.

If there is more market consumption (sale of turtles, sending turtles and eggs out of

the community, preference for cash value of turtles) in Remanso, which is closer to its

market links, then the market may be a negative influence on turtles. If there is more

subsistence consumption (hours turtling, eating turtle and preference for turtle over its

cash equivalent) in Piso Firme, which is farther from its market link, this will indicate

that turtles face threats other than market integration. Rejection of null hypothesis four

will suggest the role of the market and distance to market on turtle populations. Analyses

of further socioeconomic characteristics (hectares of cultivated land, trips out of the

community, years of education, hours fishing), indicate supplementary factors that impact

turtle populations.

As wealth increases, consumption of inferior meat decreases. If turtle is an inferior

meat, greater wealth will be associated with decreased consumption. Increased wealth

could have a positive effect on turtle populations. Because of potential changes coming

to these communities through improved road access-increased access to markets and

cash-quantification of turtle consumption under different wealth conditions has

important implications and can inform potential conservation strategies. Rejection of null

hypothesis five will indicate that turtle meat is an inferior good.

Note on my Research Style and Findings

It is worth noting that this research is largely top-down in spite of the participatory

approach I describe and support. Local participation was included as much as possible in

this research, mainly through training and participation of field assistants. Data have

been returned to the communities to help illustrate portions of the outcome. And, most









importantly, the research was designed to be accessible and it may be easily transferred to

future community-based research efforts where community members collect similar data.

Because of the influence of a national park on my research sites my efforts offer an

opportunity to contribute to applied conservation through return of the data to the

communities. This may be a starting point for discussion of conservation and

management of the turtle resource. Insight gained into how household wealth and

distance to market affect turtle consumption will further understanding of the interplay

between socioeconomic systems and natural resources. This information is valuable in

the design of conservation and management plans for natural resources by user groups.














CHAPTER 3
RESEARCH AREA: EASTERN BOLIVIA

River and Floodplain Habitats of Parque Nacional Noel Kempff Mercado

The river turtles P. unifilis and P. expansa are found in eastern lowland Bolivia,

inhabiting rivers that lie within the boundaries of the Park (Figure 3-1) (Iverson 1992;

Ministerio de Desarollo Sostenible y Medio Ambiente 1996). There are five rivers within

the Park, two of them included in my study-the Paragua and the Itenez. The Itenez

River is the largest river associated with the Park, forming the northern and eastern limits

as well as a portion of the frontier between Bolivia and Brazil. The Itenez River is a

tributary of the Mamore River, which empties into the Madera River, a tributary of the

Amazon River. Tributaries of the Itenez River include the Paragui, Verde and Paucerna

Rivers. The fifth river in the Park, the Tarvo, is a tributary of the Paragua (Ministerio de

Desarollo Sostenible y Medio Ambiente 1996). The Paragua and Tarvo Rivers form the

western limit of the Park.

The watershed of the Paragua River covers 5,500 km2 and dominates the hydrology

of the Park west of the Huanchaca Plateau-the source area for the Paragua and Itenez

Rivers and a landscape feature that defines the eastern boundary of the Park (Ministerio

de Desarollo Sostenible y Medio Ambiente 1996). The Huanchaca Plateau has eroded

over the past 20 million years, exposing Precambrian material of the Brazilian shield and

leaving a gently sloping landscape composed mainly of quartz with some granite

(Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The seasonal flooding of










the Paragua River influences the area that lies to the west of the Huanchaca Plateau and

east of Piso Firme.






5Bella Veta

R| MMagabalito
Piso Fiffm p^PNNKMl
Huanchaca
Sm Igiado_ Plateau
6 I deMasco


*SwiaQuCzdeLa9Sam



250 0 250 Klomelers



Figure 3-1. Map indicating the communities where research was conducted, nearby cities
that community members have links with, the Park and the Paragua and Itenez
Rivers.

From their dark color, the Paragua and Itenez Rivers resemble blackwater rivers.

However their properties are not strictly consistent with blackwater rivers, which are

characterized as oligotrophic with strongly acidic water and rising in white podsolic soils

in the Amazon Basin (Morley 2000). In contrast, the waters of the Paragua and Itenez

Rivers have a nearly neutral pH and do not contain large amounts of suspended organic

podsolic material-characteristics of clearwater rivers (Ministerio de Desarollo

Sostenible y Medio Ambiente 1996; Navarro and Maldonado 2002). Clearwater rivers

originate in Precambrian shield areas and are almost neutral but lack the dark coloration

of blackwater rivers (Junk and Furch 1993; Morley 2000; Salati et al. 1987).









A reflection of the absence of organic material in the Paragua and Itenez Rivers is

the poor quality of nearby soils for agricultural use. Generally, soils in the Park have low

nutrient, cation exchange, and fertility levels and a highly acidic pH. They have been

classified as appropriate for forests, not agricultural activity (Ministerio de Desarollo

Sostenible y Medio Ambiente 1996). As is typical of this type of soils in the neotropics,

the thin layer of leaves and organic material that lies on top provides most of the

nutrients, requiring the presence of forest for maintenance of much biomass (Terborgh

1992). To date there is not extensive information on the Paragua and Itenez Rivers other

than notation of their mixed limnological characteristics (Navarro and Maldonado 2002).

Parque Nacional Noel Kempff Mercado and Ecological Benefits it Provides

The Park includes 1.5 million hectares of undeveloped habitat in an ecological

transition zone in the northeastern tip of the department of Santa Cruz, contiguous with

Brazil (See Appendix A for a detailed history of the Park.). Within Bolivia, the Park is

one of the most ecologically heterogeneous protected areas (El Deber November 30,

2000; Ministerio de Desarrollo Sostenible 1996), home to a great amount of biodiversity

(see Table 3-1). The high level of biological diversity in the Park is related to the

diversity of habitats within it-tropical rainforest, gallery forest, subtropical scrub forest,

seasonally flooded savanna (Figure 3-2), and seasonal dry tropical forest and savanna-

the only place in South America where these five ecoregions converge and are protected.

Human occupation and use of these habitat types in other locations is substantial,

increasing the value of such large, scarcely or wholly unpopulated contiguous areas

(Ministerio de Desarrollo Sostenible 1996).

In addition to its biodiversity assets potential, the Park offers ecotourism

opportunities for national and international travelers. As a combined result of its area,









biological characteristics, scenic and ecotourism opportunities, the Park was named a

Patrimony of Humanity by the United Nations Education, Science and Environment

Organization (UNESCO) in November 2000 (El Deber 2000). The Arco Iris waterfall,

on the Paucema River (Figure 3-3), is one of the impressive sites in the Park that tourists

come specifically to see. Because images of Arco Iris are used in much of advertising for

the Park it is often associated with the Park.

Table 3-1. Numbers of flora and fauna species known to exist in Parque Nacional Noel
Kempff Mercado, indicating biodiversity present.
Flora or Fauna Estimated Number of Species
Vascular Plants 4,000
Mammals 139
Birds 700
Reptiles 74
Amphibians 62
Fish 250
Invertebrates 347
Source: Ministerio de Desarollo Sostenible y Medio Ambiente 1996 and El Deber
November 30, 2000.


Figure 3-2. Aerial view of seasonally flooded savanna habitat, a scarce habitat type in
the tropical Americas, pictured here near the Flor de Oro guard station in
Parque Nacional Noel Kempff Mercado (K. Conway).






























Figure 3-3. Arco Iris Waterfall, a popular ecotourism destination on the Paucerna River
in Parque Nacional Noel Kempff Mercado (K. Conway).

The Community of Piso Firme

Piso Firme, one of the two communities where this research was conducted, is

located alongside the Paragui River, at 13.620 S latitude and 61.73 W longitude, which

lies in the buffer zone of the Park. Piso Firme is 365 km N by road from San Ignacio, the

provincial capital, and 770 km NNE of Santa Cruz de la Sierra, the departmental capital.

Located on the Paragui River Piso Firme was an access point to former rubber

plantations. This combination of access to work and a river for transport and resources

made it an attractive site for settlement (de Mesa et al. 1999). Many aspects of life in the

community center around the river. With the rise and fall of the Paragui River each year

come respective scarcity and abundance of aquatic prey, including Podocnemis turtles.

The photograph in Figure 3-4 shows Piso Firme from the Paragui River in the dry

season, a time of abundant aquatic resources for predators.

















,L .~


Figure 3-4. Piso Firme from the Paragua River in the dry season (K. Conway).

According to the census I conducted in Piso Firme from June to July 2000 there are

473 permanent residents (present in the community for at least six months of the year) in

70 households. The main source of household income for Piso Firmefios was agriculture

followed by salaried and day labor (Figure 3-5). Many families have some livestock-

Sawmill
7% Park
Salaried 7%



i0% Agriculture
--46%


Fishing
9%
Cattle
Day Labor Store 0%
11% 9%


Figure 3-5. Distribution of jobs that provide the major source of income for households
interviewed in the community of Piso Firme, located along the Paragua River
(N=55 households).









commonly chickens, ducks and a few pigs. Some families have several head of cattle,

but only six families have more than 15 head of cattle, with one household claiming 160

head. This is largely a subsistence agriculture community.

Indigenous Links to Piso Firme

Culture being a factor that influences daily life I describe here cultural antecedents

of Piso Firme. I believe this background, which differs from that of Remanso, is worth

pointing out because it influences use of natural resources. As such, this information is

meant as relevant background that further informs the livelihood systems observed today,

although an in-depth comparative analysis of Piso Firme and Remanso on these terms is

beyond the scope of this work.

The residents of Piso Firme are classified by the government and identify

themselves as members of the Chiquitano indigenous group. In the 18th century,

floodplain residents in the area between the Paragua, Itenez and Paucerna Rivers were

presumed to be from the indigenous group known as Guarasug'we and were referred to

as Paucerna by mestizos in the area (Riester 1977). The Guarasug'we rejected Christian

evangelization efforts as well as attempts to involve them in rubber collection. By the

mid 20th century, the Guarasug'we population was greatly reduced-an estimated 50

individuals comprised the population in 1965. Remaining members had moved away

from the banks of the Itenez in 1954, their original place of residence, and with the death

of the last Captain, Tarekuva Miguel Frey, the group was considered culturally extinct by

anthropologist Riester (1977).

Archeological evidence linked to the Guarasug'we culture includes ceramic pottery

found in the communities of Piso Firme, Bella Vista, Remanso, Cafetal, Porvenir, Puerto

Rico and Florida, which lie in or near the Park. Pottery uncovered to date has contained









human bones, tools and ceramic pieces (Ministerio de Desarollo Sostenible y Medio

Ambiente 1996). Figure 3-6 is a photograph of one of these pots unearthed in 2001 while

residents were digging canals for water pipes in Piso Firme.

Today a woman living in the community of Bella Vista, which lies 45 km NE of

Piso Firme via the Paragua and Itenez Rivers inside Park boundaries, claims to be a

Guarasug'we (interview with resident of Bella Vista, name withheld, August 1999).

There are residents in Piso Firme who claim Guarasug'we descendence (interview with

community leader, name withheld, April 2000). So, although at least one anthropologist

considers the Guarasug'we culturally extinct, direct or indirect descendents of this

indigenous group persist in small numbers in this area today, and maintain their cultural

identification alongside the more common Chiquitano group.




















Figure 3-6. Example of a ceramic pot presumed of Guarasug'we origin, unearthed in
Piso Firme in November 2001 (P. Torico).

Overlapping the area previously occupied by the Guarasug'we and that of the

newer cultural group known as Chiquitano, which is partially descended from the









Guarasug'we. The area occupied by Chiquitanos is delimited by the Itenez River to the

north, the Grande River to the west, the frontier with Brazil to the east and the Santa Cruz

- Corumba rail line to the south (Riester 1977). The Chiquitano cultural group is the

result of the work of the Society of Jesus, through whose efforts in the 17th century

approximately 30 different ethnic groups were brought together in reducciones (villages

created by Jesuit missionaries and populated by indigenous people of multiple groups)

and acculturated over the following century (Riester 1985). Among the 30 indigenous

groups were those whose language groups included Chiquita, Otuke and Arauak (Riester

1977). With the expulsion of the Spanish and Portuguese Jesuits in the late 18th century

this acculturated group established itself in the Amazon and Chaco regions of

northeastern Bolivia, its settlements characterized by poverty and marginal conditions

(Riester 1990 in Ministerio de Desarollo Sostenible y Medio Ambiente 1996).

Several artifacts specific to the Chiquitano culture still used in Piso Firme today

include the tacu (a wooden bowl carved from cuchi, tajibo or chonta (Astrocaryum spp)

and used for crushing, mashing or grinding foods), baskets made from motacu (Attalea

phalerata) and cuse (Casearia gossypiosperma) and planting sticks made from chonta

(Riester 1977). See Figure 3-7 for an example of apanero basket made of motacu palm

fronds and used in Piso Firme today. Residents of Piso Firme commented to me about

the loss of additional utilitarian elements of the Chiquitano culture and other items,

including music and the language.

A small number of Piso Firmefios I spoke with about indigenous identity identify

themselves as Chiquitano. But based on five conversations specifically about identity, it

was apparent that there are conflicting sentiments about what it means to be indigenous.









One prominent elder member of the community indicated to me that to be indigenous is

to express a status that is shameful or "low", while another prominent elder indicated that

being indigenous served an end as far as the national government was concerned-

referring to efforts to get land title as an indigenous community. A middle-aged man of

modest means was quite proud of his Chiquitano heritage and lamented the loss of the

culture, especially the music. Interestingly, one of the people most proud of being

indigenous was a young man, approximately 30 years old, whose father's opinion of the

indigenous culture was that it was "low".


Figure 3-7. Panero basket made from motacu palm fronds (Attaleaphalerata) and used
to carry harvested products in Piso Firme (K. Conway).









Several older residents expressed a desire to teach the younger generation the

Chiquitano language and other elements of the culture, stating that it is important to keep

it alive in spite of all the migration and acculturation. In Piso Firme, only one older

woman and a regular visitor from Santa Cruz, an older man, spoke Chiquitano. The man,

a retired schoolteacher, visited annually and there was talk of the community contracting

him to teach schoolchildren the Chiquitano language. The different views and level of

cultural integration illustrate that this is clearly a community with one foot in the

indigenous world and another foot in the mestizo world and possibly on the brink of

making a choice to be in just one of these worlds. My impression of this community is

that it is more likely to become further integrated with western culture and economy than

to maintain its indigenous cultural ties and turtle consumption is likely to be affected by

ensuing changes that result from this transition.

The Community of Remanso

The census I conducted in the community of Remanso from June to July 2000

indicated that there were 97 households with 591 permanent residents (present in the

community for at least six months of the year). Remanso lies along the Itenez/Guapore

River at 13.520 S latitude and 61.870 W longitude, on the border with the Department of

Rond6nia, Brazil. Like their neighbors in Piso Firme, the lives ofRemansefios are

dominated by the Itenez River that runs parallel to one edge of the community, but apart

from this commonality, the communities have quite distinct demeanors. The distribution

of income sources in Remanso is more diverse than that of Piso Firme and while the

percentage of people who claim agriculture as their first source of income is lower, it is

still the most prevalent form of household income. Many residents have small livestock

and a few families have cattle, but few of these families have large herds and only two









claim cattle as their primary source of income. In Remanso the majority of households

obtain their income from agriculture, but unlike in Piso Firme, a substantial proportion of

households-35%-obtain most of their income through day labor or storekeeping

(versus 20% in Piso Firme) (Figure 3-8). Remanso is truly a frontier town-Brazil lies

across the Itenez River. The heavier volume of river traffic on the Itenez River, which

connects Remanso to Pimenteiras, Brazil, contributes to the stronger identification with

this Brazilian city and the culture. Piso Firme, in contrast, has more of a rural

agricultural settlement with stronger ties to Bolivia.

Salaried Sawmill P
13% 3%
18%0%

Mining



17%
33%

Fishing
Cattle
8%

Day Labor S
18%
17%


Figure 3-8. Distribution of jobs that provide the major source of income for households
interviewed in the community of Remanso, located along the Itenez River
(N=71 households).

Diverse Cultural Links to Remanso

Remanso, which was established in 1906, a year after Piso Firme, does not share

the same indigenous roots. It is a colonist settlement, begun in conjunction with the

rubber boom in the area. This history leads to a different sense of community in this case

from Piso Firme. I have several thoughts on how the colonist influences the









community's identity. Possibly because residents hail from more varied backgrounds, in

this community the varied backgrounds impede a sense of common identity, or people

want to fit in and therefore shy from pointing out their different cultural backgrounds.

Whatever the reasons) for this, I do not report on the diverse cultural antecedents of

Remanso.

Note on the History of Natural Resource Use in the Area from the 1940's to 1980's

Communities in this region have experienced several shifts in employment sources

in the 20th century. These shifts have affected lifestyles and economies in the region

through gradually increasing market integration.

There were four stages of exploitation of natural resources, beginning with rubber

in the early 20th century. Rubber barons came to the region and had a strong economic

influence in the Departments of Santa Cruz and Beni (Ministerio de Desarollo Sostenible

y Medio Ambiente 1996). Residents of the area, including Piso Firmefios and

Remansefios, collected rubber, which transformed the labor structure by introducing a

cash economy (Ministerio de Desarollo Sostenible y Medio Ambiente 1996). The

Bolivian market collapsed when the rubber trade dropped off in the early 20th century due

to the fall of prices on the international market, the growth of improved rubber (Hevea

brasiliensis) plantations in Malaysia and the introduction of synthetics.

Harvest of a medicinal plant used to cure headaches, locally known as ipecacuana

orpoalla (Cephaelis spp.) was initiated following the rubber boom, although this plant

did not develop into a large-scale trade item (Caballero 1994). The next stage of natural

resource exploitation, which spanned from the 1940s to '80s, involved several types of

wildlife, killed for their skins and feathers, for sale on the international market. The

black caiman (Melanosuchus niger) and caiman yacare (Caiman yacare) were harvested









from the 1940's through the 1970's in Bolivia (B. Godschalk personal communication

January 2002). Spotted cats-jaguar (Panthera onca), ocelot (Leoparduspardalis) and

margay (Felis wiedii)-were killed for their skins until the 1980's (Ministerio de

Desarollo Sostenible y Medio Ambiente 1996). In the 1980's psittacine birds, most

notably the hyacinth macaws (Ara hyacinthinus) and scarlet macaws (Ara macao), were

captured for the pet trade in North America and Europe (Luxmoore et al. 1988;

Ministerio de Desarollo Sostenible y Medio Ambiente 1996). With the imposition of

stricter laws on the trade and exploitation of wildlife through the Convention on

International Trade in Endangered Species (CITES) in the 1980's, trade of these species

has largely been reduced (Luxmoore et al. 1988).

In most cases today, felines are killed to protect cattle or when they are thought to

pose a threat to humans. During the year this research took place, at least two large cats

were killed, including an adult male jaguar killed in a cattle pasture by a resident of Piso

Firme who was en route to the forest to hunt. The hunter reported that the skin and skull

would obtain a good price when sold to a buyer in San Ignacio, indicating that commerce

in these items continues. More common today is resident involvement in the exploitation

of forests for timber and other nontimber forest products-there is a functioning palm

heart processing center in Porvenir, another community in the zone of influence of the

Park that lies 50 km south of Piso Firme, and a defunct one in Piso Firme.

A natural resource whose exploitation is of concern today is the river turtle. A

wildlife resource important at the regional scale, Podocnemis turtle populations have

been diminished by trade for food and as pets since the 17th century (Johns 1987;

Luxmoore et al. 1988; Pritchard and Trebbau 1984; Cavalcanti 1999; Santos Brito and









Ferreira 1978; Licata 1992). After the expulsion of the Jesuits in the late 18th century and

the introduction of substitutes for turtle oil, hunting pressure on P. unifilis and P. expansa

dropped dramatically (Castro de Le6n 1986). While these turtles are not exploited as

much for oil today as they were in the 18th century, they still face hunting pressure at

local and regional levels as subsistence and market food resources (Johns 1987; Klemens

and Thorbjarnarson 1995; Ministerio de Agricultura y Cria 1972; Mittermeier 1975;

Santos Brito and Ferreira 1978).

The continued use of river turtles in Piso Firme and Remanso in the shadow of a

protected area provided an opportunity to compare local socioeconomic and cultural

contexts surrounding wildlife resource use near a conservation area. Because markets

have been implicated in affecting wildlife use globally and because people in Piso Firme

and Remanso have access to and are involved in markets, I have chosen to emphasize

markets and factors affecting natural resource use as a result of market integration in my

research. Deeper understandings of socioeconomic factors that directly and indirectly

influence wildlife use go a long way in explaining communities' needs, which are

ultimately necessary for the development of community based conservation and

management programs, an applied objective of my research. Thus, by collecting general

socioeconomic data (e.g. household size, education level of head of household, primary

income source, number of annual trips to the city, connections to markets through

sales/trade, level of agricultural practice and wealth) in addition to specific turtle

consumption data I hope to provide a detailed picture of two communities (see Appendix

C for specific questions).














CHAPTER 4
EFFECTS OF HUMAN SETTLEMENT ON THE NUMBER AND SIZE
DISTRIBUTION OF TURTLES

Introduction

Assessment of turtle abundance and population structure has implications for the

future of these wildlife resources and the human communities that depend on them. In

areas where wildlife is harvested, monitoring harvest impacts is critical. In eastern

lowland Bolivia such impact monitoring is necessary to determine the effects of harvest

on abundance and population structure. This information can help determine whether

hunting pressure poses a threat to turtle populations and it is an important precursor to

conservation and management programs. This information was collected at three sites-

Piso Firme, Remanso and Mangabalito-to compare hunted and less-hunted sites.

Study Area

Three 40-km transects were established along two rivers in lowland eastern

Bolivia. Two of these transects, Piso Firme and Remanso, straddle human communities.

The third transect, Mangabalito, is located adjacent to an abandoned human community.

Piso Firme and Remanso are located in the buffer zone of the Park. Piso Firme lies on

the Paragua River in the Department of Santa Cruz while Remanso lies on the Itenez

River in the Department of Beni, approximately 30 km northwest of Piso Firme.

Mangabalito, also on the Itenez River, is approximately 150 km upriver from Remanso in

the Department of Beni. The central docks in Piso Firme and Remanso served as the

middle points in the 40-km stretch of river used to collect the turtle abundance data.









Piso Firme and Remanso were selected because: 1) both are on rivers within the

ranges ofP. unifilis and P. expansa; 2) P. expansa is presently found in the Itenez River

and, fishers report that the species was formerly abundant in the Paragua River; 3) I

determined in preliminary fieldwork in August 1999 that residents in these communities

eat turtle meat; and, 4) the leaders of each community were willing to allow me to

conduct my research in their respective community.

Mangabalito, with no village within 50 km, was selected as the less-hunted site in

the Park. When the Park was expanded to its present extent in 1996 the community

living at Mangabalito abandoned the site (Juan de Dios Pefia, pers. comm. 2000). While

the community of Mangabalito was smaller than that of either Piso Firme or Remanso,

approximately 12 families lived here (Juan de Dios Pefia, personal communication 2001),

the fact that people lived there four years prior to my study make this an interesting site

for observing the potential response of a turtle population to reduced hunting pressure.

Mangabalito is upriver from Remanso, on the Itenez River, at 13.780 S latitude and

60.55 W longitude. The only remaining signs of the community that disbanded in 1996

are four dilapidated houses and a few hectares of fruit orchards. No livestock was seen in

the area during the time of the surveys.

All three sites have some level of institutional presence dedicated to the

conservation of wildlife. There is a Park camp located in Piso Firme, which means there

is almost constant contact with the Park via Park guards who live there. Three of the

Park guards are from Piso Firme, although they are not always based in this camp.

Mangabalito is also the site of a Park camp, but it is not occupied on a regular basis

anymore. In addition to being occasionally patrolled by Park guards, the Brazilian









Institute for the Environment and Renewable Natural Resources (IBAMA) occasionally

patrols at Mangabalito and even less frequently at Remanso. IBAMA oversees the sport

fishers and local residents who base their activities out of Cabixe, Brasil, which is 10 km

downriver from the endpoint of the Mangabalito transect. Remanso lies on the

Itenez/Guapore River, which forms part of the natural border between Bolivia and Brazil,

and is also patrolled by IBAMA as well as Park guards. During my study, relations

between residents of Remanso and the Park were not as well developed nor accepted as

in Piso Firme. There are continued efforts on the part of the Park to involve residents of

Remanso in the implementation of the Park.

Both Bolivian and Brazilian fishermen report that IBAMA agents issue harsh

sanctions to people found with turtles or turtle eggs in their boats. Because the

Itenez/Guapore River is an international body of water, Brazilian and Bolivian agents

share authority. At all three sites there is patrolling by Park guards, although it was

highly erratic during the year of my study due to various factors-lack of guards (the

Park employs approximately 25 guards for its 1.5 million hectare extent), gasoline, or

working boats.

Survey Procedure

Turtles were counted using a basking-observation technique. This technique is

non-invasive, can yield large samples, and produces less variability in data compared to

hand capture and hoop net capture techniques (Koper and Brooks 1998). Among these

three techniques, results from basking surveys provide more accurate population

estimates although with more variability (from underestimation by 32% to overestimation

by 13%) in relation to true population size (Koper and Brooks 1998). Basking surveys

also result in lower heterogeneity of capture for the sexes than hand capture, but more









accurate estimates of female populations than male, suggesting that basking surveys may

offer a relatively unbiased form of estimating population size and obtaining large samples

over a short time (Koper and Brooks 1998). The basking technique is limited by curves

in the river and angles of basking logs that prevent all turtles from being spotted before

they dispersed as the boat approached. I conducted all surveys and alternated the start

side of the river on each run, so bias should be low or consistent and the abundance

counts for the three areas comparable to each other (as suggested in Caughley 1977).

Basking counts were chosen as the most appropriate means of collecting abundance

counts. The technique as applied here is simple, requiring a boat, a global positioning

system (GPS) unit, observation and notation. With these tools and techniques, visible

turtles were counted along 40-km stretches of the Paragua and Itenez Rivers.

Transects were set up along 40 km of river length traveling down the middle of the

river. The 40-km length was used because fishermen indicated they would travel more

than 10 km for fishing/hunting trips of more than one day and 40 km was the maximum

physical distance that could be covered in the four and one half-hour period during which

surveys were conducted. The window of survey time (11:00 a.m. to 3:30 p.m.) was

delimited by the peak hours of sunshine, a significant factor influencing turtle basking

(Koper and Brooks 1998). See Figure 4-1 for a photograph of basking turtles.

I conducted the basking surveys during the dry season, between July and October 2000.

During the dry season there are more basking spots available along river edges because

the river is low and turtles are seen along the main river channel. During the wet season

turtles more commonly frequent bays and flooded forests adjacent to rivers (Soini and

C6ppula 1995) and are rarely seen along the main river channels. The two-way surveys









were conducted from an aluminum johnboat with a 25-horsepower outboard motor,

traveling at an average speed of 18 to 24 km/hour. An assistant drove the boat

approximately five meters from the riverbank. Waiting periods of at least 30 minutes

were observed a few kilometers beyond the end point of each survey to minimize

disturbance in the count on the opposite side of the river. Turtles were observed basking

in areas we had passed by on the first run after thirty minutes, so this was presumed a

sufficient waiting period. The side of the river and direction of travel were alternated

between each journey. Side of the river was alternated for surveys conducted at the less-

hunted site, Mangabalito, since all surveys were initiated from the furthest point

upstream. Figure 4-2 shows a schematic drawing of transects.




















Figure 4-1. Podocnemis unifilis turtles basking along the Itenez River (K. Conway).


All basking turtles and turtles seen in the water were counted. For each turtle

observed the one-kilometer segment in which it was seen, according to the GPS, and its

size class were recorded.









km 31-40 km 21-30 km 11-20 km 1-10
-, -, '" -, "' -, "' 11 community
Far Close commuClose Far
Downstream Upstream

Figure 4-2. Schematic diagram of the 40-km transects illustrating location of Piso Firme
and Remanso at the center of the 40-km transects and categorization of km
into distance categories.

I divided transects into two distance categories in relation to the communities-far

and close. "Far" segments are those more than 10 km upstream or downstream from the

center point while "close" segments are those within 10 km of the transect center. This

distance represents a cutoff point for one-day fishing trips according to local fishers and

was the main criterion in making this delimitation. In an Amazon Basin-wide study of

hunting pressure, Peres and Lake (2003) found a 9-km threshold for hunting activity

around villages, which further suggests this approximate distance as a division.

I conducted 30 surveys in the period from August 14 to October 3, 2000. The sum

of these surveys provided eight complete counts of turtles in 40-km transects at each site.

It was not always possible to survey the full 40-km transect on each river within each

assigned survey day due to weather, logistic constraints, equipment failure, or assistant

availability. Therefore survey coverage was not identical in each transect but section

results were combined among surveys conducted on successive days to provide

effectively equal effort in all sections over the course of the study. Counts of turtles

observed in each kilometer section were summed over 10-km intervals and the resultant

turtle counts were compared among the different rivers and between the close (within 10

km of community/center point) and far (>10 km from community/center point) distance

categories.









Size classes for my study were determined based on a combination of research on

both P. unifilis and P. expansa done on the Pacaya River in Peru by Soini (1995a and

1995b); the Maniqui River in Bolivia by Quiroga Vera (2000) and Knothe and Muihlteich

(1996); near Iquitos, Peru by Fachin Teran et al. (1992); and, life history data from

Pritchard and Trebbau (1984).

I set the size classes for my study to reflect the larger measurements of P. unifilis

cited by Soini (1995a) and Pritchard and Trebbau (1984) because I anticipated seeing

more of these turtles due to their greater abundance compared to P. expansa (Pritchard

and Trebbau 1984). Turtles in my study were classified into three size classes. These

classes are: < 20 cm (small), 20 to 35 cm (medium), and > 35 cm (large). The turtles for

which size could not be distinguished were classified as unknown. For P. unifilis, the

size classes assigned in my study apply as follows: small includes juveniles of both sexes;

medium includes mature males and immature females; and, large includes mature

females. For P. expansa, the size classes apply as follows: small includes juveniles of

both sexes; medium includes immature males and females; and, large includes adult

males, immature and mature females. In addition to the size class data collected on

turtles in transects, there were several opportunities to take measurements of captured

specimens. These measurements combined with information from the literature helped in

determining the size classification used here.

Habitat Description

In addition to the abundance and size class data collected, each km of transect was

described using a 13-category classification developed in the field (Appendix B). This

classification system was determined based on visual observations made over the course

of abundance counts and quantified near the end of the basking season (September 2000).









Once the categories were defined, I traveled along transects in my boat, driven by my

assistant, and quantified the meters of shoreline of the categories in each km.

Results

Basic descriptive statistics of the numbers of turtles counted at the three study sites

are presented in Table 4-1 and Figures 4-3 to 4-5. The numbers indicate a disparity in

numbers of turtles observed along the 40-km transects, with fewer turtles seen at the two

sites where there are human communities, Piso Firme and Remanso. This disparity is

echoed visually in the graphical presentation of this data, Figures 4-6 and 4-7, which

illustrate the average number of turtles seen in each kilometer at each of the three sites

over the sampling period with one standard deviation bars indicated. There is a clear

decline in the number of turtles around the center point of the two human-inhabited

transects.

Table 4-1. Descriptive statistics for the turtles counted along the three 40-km transects
(Piso Firme, Remanso and Mangabalito).
Minimum # Maximum # Mean Std
Site N
turtles/transect turtles/transect Turtles/km Dev
Piso Firme 387 12 77 1 21.08
Remanso 469 18 98 2 21.93
Mangabalito 2889 181 501 9 102.45

Mean Numbers of Turtles at Different Distances from Transect Center

I compared the mean number of turtles observed in 10-km segments between and

within sites. The 10 km immediately upstream and downstream of the center point of the

transects were designated "close" while the 10 km furthest upstream and furthest

downstream from the communities were designated "far" based on local turtling and

fishing practices as well as data from other studies suggesting this is a threshold near

which hunting intensity changes. Turtle counts in the 10-km segments used for analyses









are graphed in Figures 4-3 to 4.5. Descriptive statistics and results of the analyses are

found in Tables 4-2 to 4-5 below.

Table 4-2. Kruskal-Wallis test statistics comparing median number of turtles in segments
within (Close) and beyond (Far) 10 km of the center point between the three
sites. These results indicate that the difference is not random.
Far Close

N 24 24
Median 47.50 27.0
x2 16.99 16.90
df 2 2
p-value <<0.001 <<0.001

The results in Table 4-2 indicate that the larger median number of turtles seen in far

km is not due to chance.

Further examination of the data included looking at each site individually. The

results of the Mann-Whitney comparison of medians within sites (Table 4-3) indicate that

the differences seen in close vs. far segments are significant at Remanso only. The

median values indicate that more turtles were observed at distances beyond 10 km from

the center. On a site by site basis, the difference in number of turtles seen in close km is

significantly smaller at Remanso only. This finding suggests that hunting pressure is

stronger at Remanso than Piso Firme. Mangabalito, which has no constant human

presence, has more turtles overall and does not exhibit a pattern like Remanso. This may

be indicative of a more robust turtle population.

Table 4-3. Mann-Whitney test statistics for the comparison of close vs.far segments
within sites. Results indicate a significantly higher median number of turtles
farther from the center of the Remanso transect only.
Site Distance Mean Std Dev U p-value
Piso Firme Far 47.75
11.75 30 0.88
Close 49.00
Remanso Far 95.50 5
86.00 5 0.003
Close 21.75
Mangabalito Far 375.00
52.21 28 0.72
Close 347.25









Table 4-4. Frequencies of counts greater than or less than/equal to median values within
10 km (Close) or beyond (Far) transect center points of the three sites.
Median # Turtles Piso Firme Remanso Mangabalito
Far 47.50 > Median 1 3 8
<= Median 7 5 0
Close 27.00 > Median 3 0 8
<= Median 5 8 0

The results of the median frequency test show patterns at the three sites. At Piso

Firme only one median count was above the median value of 47.50 for far kilometers

while seven were below it and three of eight sample counts were greater than the median

calculated value of 27.00 turtles for close segments (Table 4-4). These values indicate

that in 12 of 16 cases the number of turtles observed was less than expected, suggesting a

small population. From the graph of turtles counted at Piso Firme it is evident that there

were more turtles downstream than upstream (Figure 4-3). This is possibly a combined

result of hunting pressure and ecological factors.

The observations are similar at Remanso, where the number of turtles counted in

km farther from the community was above the median value in three of the five samples

but never at the close segments. Again, this relationship is illustrated in the graph of

turtles counted, where there is a clear dip in the numbers around the center of the transect

(Figure 4-4).

In stark contrast, at all distances for all samples at Mangabalito the mean number of

turtles observed was above the median. The graphs of numbers of turtles observed

(Figure 4-5) indicate that more turtles were observed downstream at this site. Clearly the

turtle population at Mangabalito is greater than those at Piso Firme and Remanso.

Finally, to observe differences between hunted and less hunted sites, Mann-

Whitney tests were used (Tables 4-5 and 4-6). It is worth noting again that Mangabalito






57


and Remanso are on the same river, the Itenez, which makes a stronger comparison

between these two sites and the basis for testing research hypothesis two.


Piso Firme


KM 1-10 KM 11-20
KM 1-10 KM 11-20


7


KM 21-30


KM 31-40


KM Section

Figure 4-3. Number of turtles counted at Piso Firme, along the Paragua River, shown in
10-km segments. Each line represents a complete survey along the 40-km
transect.


KM 1-10 KM 11-20


KM 21-30 KM 31-40


KM Section


Figure 4-4. Number of turtles counted at Remanso, along the Itenez River, shown in 10-
km segments. Each line represents a complete survey along the 40-km
transect.

The results indicate that the difference in mean numbers of turtles counted at Piso

Firme is less than that at Mangabalito and the difference is not due to chance alone. The









difference between Remanso and Mangabalito also indicate that the difference in mean

numbers of turtles counted at Remanso is less than that at Mangabalito. Hypothesis two

states that there will be more turtles at Mangabalito (suggesting less hunting pressure)

than at Remanso (suggesting more hunting pressure). Based on the results of the Mann-

Whitney test of difference, this is the case.


200
180
160
140
120
100
80
60
40
20
0


KM 1-10 KM 11-20 KM 21-30 KM 31-40
Km Section


Figure 4-5. Number of turtles counted at Mangabalito, along the Itenez River, shown in
10-km segments. Each line represents a complete survey along the 40-km
transect.


Table 4-5. Test statistics for test of difference between numbers of turtles counted at
close vs. far segments from center point of three transects.
Far Close
Mann-Whitney U 0.00 0.00
p-value <<0.001 <<0.001


Table 4-6. Comparison of differences between sites in numbers of turtles counted at
close vs. far segments from center points.
Far Close
Comparison
Mann-Whitney U p-value Mann-Whitney U p-value
Piso Firme vs. Mangabalito 0.00 <<0.001 0.00 <<0.001
Remanso vs. Mangabalito 1.0 <<0.001 0.00 <<0.001
Piso Firme vs. Remanso 12 0.04 14.5 0.07


IcenterI




























KM


* Piso Firme 0 Remanso


Figure 4-6. Average number of turtles seen in each km over eight samples at populated sites (Piso Firme and Remanso) with one
standard deviation indicated by bars on each kilometer and the communities at KM 20. Note the different Y-axis scale
from Figure 4-5, which illustrates the unpopulated site.


15

13

11











65

55

S45

35 -

25 25

15



-5
1 6 11 16 21 26 31 36
KM

Figure 4-7. Average number of turtles seen in each km over eight samples at the unpopulated site (Mangabalito) with one standard
deviation indicated by bars on each kilometer. Note the different Y-axis scale on Figure 4-4, which illustrates the
populated sites.









River Edge Habitat Characteristics

I describe the similarities and differences among the three sites based on the habitat

type classification I developed because it provides additional comparative information

between sites (Figure 4-8). I provide this information in addition to the comparison of

turtle numbers between sites based mainly on human presence as an important

consideration in factors leading to differences seen among the sites.

There are more steep barren cliffs at Remanso and Mangabalito than at Piso Firme.

This is likely due to the higher order of these rivers as well as geological differences.

These physical characteristics may cause a concentration of turtles in the main river

channel, which may contribute to the larger number of turtles seen in the Itenez vs. the

Paragua Rivers. A related characteristic is depth. In some parts of the Paragua River the

bottom at center is less than 1 m, while on the Itenez River the bottom was rarely to never

less than 1 m deep. This characteristic may also contribute to the Itenez River being

more suitable habitat for the turtles and could partially explain greater numbers seen on

this river.

There are more steep cliffs with aquatic vegetation (that constitutes some of the

turtles' diets (Bruno Coca 1999)) and sticks (important basking platforms) at

Mangabalito than either Remanso or Piso Firme. This characteristic may partially

explain greater abundance at Mangabalito due to the concentration of turtles in the main

channel, more foraging opportunities and basking platforms.

There are a few more beaches at Remanso and Mangabalito than at Piso Firme.

One possibility is that this may reflect more nesting habitat, particularly for P. expansa,

which nests exclusively on sandy beaches. This characteristic may have less effect












(U
S3000
2500
2000
S 1500
1000
500
E 0

O


3000
2500
2000
1500
1000
500
0


3000
2500
2000
1500
1000
500
0


rn El


lI [-I I-I rI--


II


Habitat Type


Figure 4-8. Meters of each habitat type at the three sites, shown individually by site. A)
Piso Firme. B) Remanso. C) Mangabalito.


H H-n


I I









on the abundance of P. unifilis. The possibility of more nesting habitat may contribute to

Mangabalito being a preferable site in comparison to Piso Firme or Remanso.

In summary, while there are subtle differences in river edge habitat between sites,

nothing stands out as dramatically different. The proportions of each vegetation type are

relatively similar across transects. This lends further strength to the suggestion that the

main factor leading to difference in abundance between the sites is human presence.

Size Differences among Sites

The mean carapace length, measured along the straight line from anterior to

posterior edges, of 10 P. unifilis females measured from the Paragua and Itenez Rivers

was 38 cm, with a range of 19 to 46 cm (sd=7.5 cm). Of these females, at least four had

eggs (the smallest egg-bearing turtle measured 36 cm length), which indicates that at least

four were reproductively mature. The mean carapace length of P. expansa females

measured was 59 cm (sd=10.7 cm, n=8) and 39 cm for the males (sd=4.9 cm, n=7). Of

the P. expansa specimens measured, one (length=70 cm) had egg yolks.

Soini (1995) gives both the straight line and curved measurements of one female P.

unifilis carapace as 47.7 cm and 50.5 cm, respectively, which yields a ratio of 1.05.

Extrapolating from this datum, the range of curved length measurements for the turtles

measured in my study may be calculated as 20 cm to 48 cm, which suggests that the start

of the large size class category at > 35 cm is a conservative estimate. Based on

measurements taken in the field that included a 36-cm long P. unifilis female with eggs,

combined with data from the literature (Pritchard and Trebbau 1984; Soini 1995a), the

large size class should consist solely of reproductively mature females and males. This

may negatively affect the number of nests and the number of eggs laid since smaller size

in P. unifilis is correlated with smaller clutch size (Vanzolini 1977). Figure 4-9 illustrates














300

250

S200
I-
0 150

E 100
z
50

0


Small Medium Large Unknown
<20 cm 20-35cm >35cm


Small Medium Large Unknown
<20 cm 20-35 cm >35 cm


Small Medium Large Unknown
<20 cm 20-35 cm >35 cm


Figure 4-9. Number of turtles in each size class observed with one standard deviation of
the number of turtles in each size class at each site and relative percent of the
turtles at the site in each of the four size classes noted. Note the different
scales between sites. A) Piso Firme B) Remanso C) Mangabalito









that the medium size class is the largest of the four size classes, at all three sites. Also,

the large size class is considerably greater at Mangabalito and is a greater proportion of

the overall sample at this site in comparison with the others. The small class is very

abundant at Piso Firme. The standard deviation for the medium size class is the greatest

(sd=74.45) (the other standard deviations are: small (sd=34.02), large (sd=58.8) and

unknown (sd=1.67)).

There is a trend towards smaller turtles near Piso Firme, but this does not exclude

the presence ofP. expansa in the Paragua River, because fishermen reported catching

two individuals, and I observed at least one adult female in transect counts. Two possible

explanations for the lack of sightings of more P. expansa are: 1) they are not as

commonly seen basking (local fishermen and R. Vogt pers. comm. 2000); or, 2) they are

not necessarily resident in this river but are there seasonally in search of nesting beaches.

Of these explanations, the former has been cited in the literature (Pritchard and Trebbau

1984) and noted by biologists working in the Amazon (R. Vogt pers. comm. 2000).

According to local residents, in the past P. expansa was commonly found on the Paragua

River. Again, this information suggests reduced populations near humans.

The turtles observed at Remanso are on the whole larger than those on the Paragua,

with a 16:63:19 ratio of percentages of small:medium:large. Based on the large size

class, which includes reproductively mature P. unifilis females, there is a larger portion of

potentially reproducing females at Remanso than Piso Firme. The large turtles observed

at Mangabalito (32% n=926) probably include reproductively mature females of both

species and is by far the largest percentage among the three sites. Assuming the large

category is the only one that contains reproductively mature females, less than seven









percent of the turtles seen at Piso Firme, less than 19% at Remanso and less than 32% of

the turtles at Mangabalito are reproductively mature. For long-lived species with delayed

sexual maturity, the ratio of mature to immature members is low, and these numbers may

be representative of this. But further counts must be conducted and compared over time

to discern trends.

Due to differences in sizes of the two species and the size class categories applied,

the majority of the P. expansa specimens seen are immature, so discussion of population

structure will be restricted to P. unifilis except when stated otherwise.

Comparison of Size of Turtles at Different Distances from Transect Center

Table 4-7 indicates the proportions of turtles of all sizes observed closer and farther

from transect center points. These numbers suggest trends towards larger turtles farther

from transect center points.

Table 4-7. Proportions of turtle size classes seen within sites at far and close categories
with averages over all three sites and at Mangabalito, the less-hunted site.
Proportions of Size Classes

Distance Small Medium Large Unknown
Piso Firme far 37% 47% 10% 6%
close 38% 56% 3% 3%
Remanso far 15% 63% 19% 2%
close 17% 62% 18% 2%
Mangabalito far 17% 44% 39% 0%
close 23% 52% 25% 0%
Size Class Average 25% 54% 19% 2%
Mangabalito Average 20% 48% 32% 0%

Visual inspection of the proportions of the size categories observed at each site

(Figure 4-10) confirms the differences in size class distributions among the three sites

shown in Figure 4-9. Examining the size class proportions between close and far sections

at each site, it appears that there is a difference in the number of large turtles at close







67


categories at both Piso and Mangabalito, but not at Remanso. The proportion of small

individuals also differs somewhat between close and far categories at Mangabalito.





U 100% 3%
S90% --- 17%
S80% 0-0
N 70% Ounknown
S 60% small
50% -
50 *medium
C 40%
0 30% Large
o 20%
.
O 10%
C, 0% -- -




/ C/1,
Site and Distance cla



Figure 4-10. Proportions of turtles of different size classes by distance categories at the
three sites.

The median number of turtles was greater at further segments when the three sites

were combined for analysis (Table 4-8). Analyses of these differences by Kruskal Wallis

test (Tables 4-8 and 4-9) indicate that these differences in median size are significantly

different (p=0.07 for far and p=0.09 for close). If the Mangabalito site, which has

experienced less hunting pressure since 1996, can be considered as a "recovered" or

"partly recovered" location with normal size class distributions, then it is apparent that

turtle size class can be quite variable in the absence of human predation. While the

pattern at Piso Firme conforms to expectations that sites nearer to the community would

have fewer large specimens, results at Remanso are not as dramatically low. The very

low proportion of large turtles near Piso Firme (3%) compared to the levels at

Mangabalito (25 to 39%) suggests depletion of this size class while the proportion of









large turtles near Remanso (18%) suggests moderate depletion. However additional data

collection is needed to confirm whether these interpretations are robust.

Table 4-8. Kruskal-Wallis test statistics comparing all four size classes at close vs. far
segments from center point at all three sites. Results indicate significant
differences between size classes in close vs. far segments overall.
Far Close
N 12 12
Median 72.50 35
x2 7.21 6.57
df 3 3
p-value 0.07 0.09

Table 4-9. Frequencies of size classes counted in relationship to medians for far vs. close
transect segments when all three sites are combined.
Size Class
Median # Turtles Small Medium Large Unknown
Far 72.50 > Median 1 3 2 0
<= Median 2 0 1 3
Close 35.00 > Median 2 3 1 0
<= Median 1 0 2 3


Examining the median values (Table 4-9) reveals that: 1) small turtles are more

abundant in closer km; 2) there are more medium turtles at both distances; 3) large turtles

are more common in far km; and, 4) unknown size turtles are less frequent at both

distances. These trends suggest that the medium size is the most common size class

among all three sites. This size class is not indicative of sexually mature populations

because the medium size class only includes mature males. My data suggest that hunting

pressure is negatively affecting the longevity of the populations as posited in hypothesis

three.

A Gamma test was applied to compare the association between size and distance

at hunted and less hunted sites for turtles of known size classes (Table 4-10). The greater









number of discordant pairs indicates that there is a negative association between size and

distance.

Table 4-10. Gamma test for known size classes at different distances compared between
hunted and less hunted sites indicating a negative association between turtle
size at close vs. far distances in hunted and less-hunted areas.
Size Class and Number of
Turtles
Site and Distance Small Medium Large
Piso Firme & Remanso close 89 164 22
Mangabalito close 322 718 347
Piso Firme & Remanso close 130 331 93
Mangabalito far 258 662 579
Concordant Pairs 1,098,451
Discordant Pairs 1,781,691
Gamma -0.24


When a comparison of turtles is made on the basis of increasing hunting pressure

and known size, there is a negative association (Table 4-10). There is a 76% probability

that a decrease in turtle size is associated with proximity to humans. Humans appear to

have a negative impact on the sizes of turtles seen.

Discussion

Differences in Turtle Abundance in Relationship to Humans

The abundance data presented in this chapter suggest that human activities,

including predation, may play roles in reduction of turtle numbers. For research

hypothesis one (HI: turtles will be less abundant closer to human communities) there was

statistically significant evidence that indicated differences in turtle abundance closer to

humans vs. farther away from humans with fewer turtles counted closer to humans at the

hunted sites. For research hypothesis two (H2: there will be more turtles at Mangabalito

(less hunting pressure) than at Remanso (more hunting pressure)) there was also evidence

for differences being due to proximity to humans at Remanso.









If hunting is the main reason for the distance-abundance relationships, my data

suggest that current practices of turtle hunting depress turtle abundance. The fewer

turtles seen close to the community at Piso Firme and Remanso may be a result of

hunting pressure. The lower turtle counts on the Paragua River, around Piso Firme,

suggest that the effect of humans on turtle abundance may be greater at this site. This

may be a result of the smaller river and hunting pressure. The contrast of the two human-

inhabited sites with uninhabited Mangabalito, which has greater numbers of turtles at all

distances in 100% of the samples, suggests even more clearly that the cause of the

difference is human activity. A comparison of the two sites on the Itenez River,

Remanso and Mangabalito, provides further support for the hypothesis that hunting

pressure is the main difference between these sites that leads to fewer turtles close to

humans because the confounding factor of different rivers is not present.

Redford and Robinson (1985) showed that depletion of game species within the

catchment area of villages led to decline of game. Peres and Lake (2003) quantified a

correlation between proximity to humans and reduced prey abundance in the Amazon

Basin, with 9 km being the threshold beyond which hunting did not negatively affect prey

abundance. My use of a 10-km threshold from humans was based on information from

fishers that they do not typically travel more than that distance on day-long fishing trips.

Thus, my results concur with Peres and Lake's large-scale work and the conclusion of

hunting pressure within a catchment area leading to degradation of natural resources

closer to humans that agrees with Redford and Robinson. Like Bodmer et al. (1997) my

results show that human activities, which may include hunting pressure, can negatively

influence game species, in this case river turtles.









Differences in Sizes in Relationship to Humans

The third research hypothesis addressed the size of turtles expected nearer to

humans vs. farther (H3: there will be fewer large turtles closer to human communities

than farther away). Distance in relationship to humans may play a role in the size of

turtles observed, most likely because hunters preferentially take larger turtles. My

findings indicate no significant differences in sizes of turtles between sites only a trend

towards smaller turtles more frequently observed closer to human communities and larger

turtles farther away. The difference in population structure suggested by the analyses of

all three sites indicates more mature populations farther from the communities than

closer, similar to the cases reported by Robinson and Bodmer (1999). When the human-

inhabited and presumably heavier hunted sites (Piso Firme and Remanso) are combined

and compared to the uninhabited and less-hunted site (Mangabalito) there is a negative

association (Y=-0.24). Larger turtles are less abundant closer to Piso Firme and Remanso

than Mangabalito. This suggests that the larger turtles are the first to be removed nearer

to humans as a result of hunting pressure.

The number of different-size turtles observed at the three sites indicates many more

small and medium than large turtles at Piso Firme and more medium than large turtles at

Remanso and Mangabalito. The most mature population among the three sites is

Mangabalito, where 32% of the turtles observed were large. According to Congdon et al.

(1993) juvenile loss rates must be extremely low in long-lived organisms for populations

to remain stable. If small turtles bask at the same rate as medium and large turtles, higher

proportions of this size class than were noted would be expected in populations with high

recruitment. In terms of juvenile loss rates, Piso Firme shows the lowest proportions

(37% of the turtles seen were small vs. 16% at Remanso and 20% at Mangabalito), which









may indicate that juveniles are surviving longer at this site or recruitment is greater.

However, it is not certain whether small turtles have the same basking behavior as larger

turtles, so no definite conclusions about survivorship may be reached based on this

basking count data alone. Trends in the medium size class (which includes juvenile

females) must be observed over time to determine what is happening with the juvenile

portion of these populations, because this will have an effect on the viability of these

populations ofP. unifilis.

My data show a high percentage of medium turtles at Remanso (63%) but not many

of these appear to reach maturity since only 19% of those seen were large. Signs of

hunting pressure are particularly strong at Remanso. Piso Firme also appears to have

hunting pressure, but it is possible, considering that abundance and size class data

indicate fewer differences and fishers report catching fewer turtles from this river, that

hunting pressure is lower than at Remanso and turtles are beginning to recuperate.

Noting the trend towards larger turtles seen at the Mangabalito site it is worth considering

that humans may have a negative effect on the size of turtles in demographic terms.

Conclusion

My results show that there are fewer and smaller turtles close to human

communities. This may be the result of hunting pressure. If this is the case, removal of

turtles and their eggs is harmful and is leading to decreased river turtle populations in this

area of Bolivia. The observation of more and larger turtles near the recently abandoned

site of Mangabalito compared to both of the two currently hunted sites may be the result

of turtle populations having recovered from hunting pressure here. If there has been a

recovery of the population at Mangabalito in the absence of humans for five years,









currently reduced hunting pressure may be sufficient to allow turtle recovery. The

implication that recovery is possible is exciting for conservation efforts.

Because turtle size is an indication of population maturity, my findings suggest

negative potential impacts of turtle hunting. If a population consists mainly of small and

medium turtles the implication is that large reproductively-mature turtles and long-term

survival of the populations may be threatened. This is especially true at Remanso, where

abundance appears to be strongly affected by hunting pressure.

My data show abundance, and the time to reproductive maturity is uncertain

(estimates range from 5-10 years (Pritchard and Trebbau 1984)), so the persistence of the

turtle populations studied at these sites in lowland eastern Bolivia cannot be estimated.

Both P. unifilis and P. expansa turtles are threatened and in danger of being hunted to

local extirpation in much of their ranges as a result of hunting pressure. My data suggest

that eastern lowland Bolivia is no exception to this trend. However, if hunting pressure

on the turtles is reduced, there is reason to believe abundance can increase, restoring two

valuable ecosystem and social system components and providing humans with continued

access to this important food resource.














CHAPTER 5
SOCIAL STRATIFICATION, MARKET INTEGRATION, AND HOUSEHOLD
CONSUMPTION

Objectives and Hypotheses

The objective of this portion of my research was to collect data on selected

socioeconomic characteristics and degree of market integration and analyze their effects

on Podocnemis turtle use in lowland Bolivia. Data were collected on assumed measures

of market integration and turtle consumption patterns (described below).

Von Thunen's intensity theory ([1826] 1966) described a relationship between

distance to market and intensity of production. Building from this theory I made a

comparison between two riberefo communities. My designation was that Remanso was

more market integrated because it is 1) closer to a market-Pimenteiras, Brazil; 2) lies on

an international, highly traveled river; and, 3) offers greater cash-generating job

possibilities for residents. Based on these qualities, the cost of participation in a market

is less for residents of Remanso than Piso Firme, leading to its apriori classification as

more market integrated.

Various socioeconomic data collected served as dependent variables for the market

integration analysis. The variables tested include: "trips"-the number of times per year

a household member traveled out of the community (an indication of outside ties, the

majority are at least partially market-related); "chaco"-how much land they cultivate

(an indication of subsistence livelihood and therefore not market-related); "education"-

years of education of the head of household (indication of greater access to cash-earning









jobs); "size"-number of people in the household (indicator of a greater need for access

to market or cash); and, "job"-type of job (certain employment offers greater cash-

earning possibilities).

The indirect characteristics reported were chosen because they appear to effect

natural resource use in the context of market integration as demonstrated through

linkages via trips out of the community. Increased harvest of natural resources attributed

to market integration has been described by Godoy (2001), Redford and Robinson

(1985), Sierra et al. (1999), and Von Thtinen ([1826] 1966). In contrast, other studies

cite positive relationships between market integration and natural resource conservation

(Agrawal 2001; Agrawal and Yadama 1997; Godoy 2001; Schweik et al. 2003; Tang and

Tang 2001). My objective in testing the variables I tested was to better understand the

relationships between market and subsistence livelihoods and turtle consumption in Piso

Firme and Remanso.

The two communities were also compared in terms of household turtle

consumption. Turtle consumption variables tested include: "eat"-number of times the

previous week turtle was consumed in the household; "send turtles"-whether someone

in the household sends turtles out on the weekly bus; "send eggs"-whether someone in

the household sends turtle eggs out on the weekly bus; "hours turtling"-hours per

household per week dedicated to turtling; and, "prefer"-preference between turtle, other

meat or cash. "Eat" and "hours turtling" were considered subsistence forms of

consumption because they do not require cash. "Prefer (turtle)" was considered

subsistence consumption because it suggests greater usefulness for the meat over the cash

equivalent. "Send turtles", "send eggs" were considered market consumption because









they involve consumption outside of the community, which has been designated a non-

subsistence use by the Park. "Prefer (cash)" was considered market consumption because

it was thought to indicate greater importance of cash over turtle, which would be more

likely in a more market-integrated household.

The hypotheses tested are:

H4: There is greater market consumption of turtle protein in Remanso than Piso

Firme and greater subsistence consumption of turtle protein in Piso Firme than in

Remanso.

H5: There is a negative relationship between wealth and turtle protein sale/trade

such that as personal wealth increases, the consumption of turtle protein decreases.

Personal wealth affects decisions about consumption. A deeper understanding of

the relationship of these potentially interactive factors has implications for the future of

the turtles as food sources, as incorporated into attitudes and a management plan. Tests

of difference on these data examine differences and similarities of effects of wealth on

turtle consumption. If community members are interested in developing formal rules of

use for the turtle resource, these socioeconomic analyses illustrate some confounding and

unrelated characteristics they may chose to consider further.

Methods

Sample Size Calculation

The unit of analysis is the household. A household is defined as the group of

individuals who lives together, shares food resources and eats together. Individuals over

the age of sixteen are considered adults because it is common for children to move out of

their parents' home and start their own families at this age. A systematic random sample

of households was the basis for data collection and hypothesis testing.









To determine the number of households per community I needed to sample, I began

with censuses. In June and July 2000 I went door to door in each community,

accompanied by a member of the community, and asked an adult of the household how

many people lived in the house for more than six months of the year. I also asked for the

breakdown of individuals living in the household under and over 16 years of age. The

sample size was calculated based on these censuses. The following formula was used

with the census information to determine the number of households required in the

samples:

S2 N P(1-P)
C2 N-i) + 2 P(P) Sample population

Where X2 is the value for chi-square at the 95% confidence level (3.841) with one

degree of freedom, N is the population size, P is the population parameter (0.5 was used

in this calculation because it assumes the greatest amount of variability in a population

and results in the largest sample population being calculated) and C is the confidence

interval of five percent (Bernard 1995). The calculated required sample size of

households was 56 for Piso Firme and 77 for Remanso.

Interview Administration and Questions

Sampling design was random and based on the censuses, with household head

names alphabetized and interviews assigned using a random number chart. Data were

collected using structured interviews administered from August 2000 through May 2001.

The interview consisted of three sections-demographic, income and market access, and

turtle-related questions (Appendix C). For the majority of interviews a community

member accompanied me. In instances when an assistant was not available I conducted

the interviews alone. The adult head of household of either sex who was available at the









time of the visit was interviewed. Visits took place between 7:30 am and 6:30 pm, seven

days of the week. In many cases both adult heads of household were interviewed.

Wealth Ranking

The wealth ranking of households in each community was conducted following

Slocum et al. as "a very effective way to determine socio-economic groupings in a

community by measuring households" (1995: 236-7). This exercise provided a valid

indicator of relative wealth rankings in each community, from the perspective of

knowledgeable informants (Adams et al. 1997; Takasaki et al. 2000) rather than based on

my personal interpretation of wealth in Piso Firme and Remanso.

The wealth-ranking exercise was completed by the recommended five people in

each community (four individuals and one couple) who were chosen because they were

long-time residents who knew all residents in the community, represented both sexes, had

different educational backgrounds and were engaged in different livelihoods (Slocum et

al. 1995).

Responses were converted into one ranking scale for each community. Names of

household heads from the censuses conducted in each community (June and July 2000)

were written on cards. Participants classified all households in their community, relative

to the others, in terms of personal wealth. Working with the participants, I read the

names aloud and asked the participant to sort the cards into groups according to their

criteria. After the first sort, names in each pile were read aloud again and participants

made any changes they decided were necessary. Next, participants explained their

criteria for placement in the various groups and elaborated common characteristics

among group members (see Appendix D for a detailed description of the participants'

categorization criteria). Participants created three to six groups. The final score for each









household was generated based on the number of groups each participant created and the

total number of households in the community using the formula:


household's rank by
participant n total # of households
total # of groups X in community = household score
in community
assigned by
participant n

The five scores for each household were averaged by community. Ultimately, three

wealth groups were deemed sufficient indicators of differences among wealth groups

within each community. Each household was placed into a wealth group based on its

average score. The wealth group assignments were the basis for testing hypothesis five.

Definition of Market Goods and Market Integration

For my study, market goods are those grown, hunted or fished and sold or used as

cash. Goods that are grown, hunted or fished for household consumption or given to

other households as gifts and not as payment for services are considered subsistence

items. This definition is based mainly on the exchange of cash or goods with an

established cash value in an attempt to address the conditions of rural communities where

access to cash is limited and residents have developed systems of barter and exchange

that do not require cash for acquisition of goods or services.

Market influence, as it exists for residents of Piso Firme and Remanso, is based on

growing access to cities and ensuing dependence on cash to acquire goods and services

necessary to achieve a desired lifestyle coupled with more consistent and greater cash-

generating activities outside the communities (e.g. work on Brazilian fishing boats, in the

mine located south of Remanso, or skilled and unskilled labor in towns such as San

Ignacio, Bolivia and Pimenteiras, Brazil). Access to cash and a need to leave the









community to acquire it was noted by a number of people throughout my time in the

field. This is particularly relevant to turtles, which are a local commodity that brings in a

high-cash profit.

Access to cash depends directly upon income source. Job categories used in the

analyses are mutually exclusive except for "State", which is a combination of teachers,

doctors and nurses; and "Park", which is a combination of Park guards and those who

work for Fundaci6n Amigos de la Naturaleza, the nongovernmental organization that co-

manages the Park with the Servicio Nacional de Areas Protegidas. These jobs were

combined because of the overlap among entities that pay the salaries. The income-

earning potential forjobs was determined through a combination of interviews with

residents and personal observation of livelihood strategies that suggested greater outside

linkages and cash generation as well as more frequent cash exchange afforded by certain

jobs.

Socioeconomic Characteristics and Their Implications for Livelihoods

Some characteristics have built-in assumptions, which are described with the

variable. The "job" that supplies the household with its primary source of livelihood has

an influence on use of natural resources as well as access to cash. "Jobs" that are based

on natural resource extraction may directly involve turtle harvest in the communities of

Piso Firme and Remanso, e.g. fishers who dedicate a portion of their time to capturing

turtles for household consumption or sale for cash to obtain other goods. This is

especially important in Piso Firme, where job options that provide cash are limited but

demands for cash are decidedly present. One measure of a household's access to cash is

suggested by its primary income source. When "jobs" are graphed in order of income-

earning potential, it is evident that agriculture is the main source of income in both Piso









Firme and Remanso (Figure 5-1). The mean for "chaco" (hectares of cultivated land)

indicates that most households in both communities maintain a chaco (x =0.80 ha

sd=0.47 for Piso Firme and x =1.03 ha sd=3.17 for Remanso) (Table 5-2). It is also

worth noting that a greater percent of the households in Remanso (48%) are supported by

higher-income-earning jobs (day labor + park + state + store) than are in Piso Firme

(38%).





A 30
S.30 i-------------------I
25
320
I 15
o 10 --
5
0
E
z



Sv DO Piso Firme Remanso



Figure 5-1. Main source of household income, in increasing order of cash-generating
potential in Piso Firme (N=55) and Remanso (N=71).

"Trips" outside the communities is one measure of external linkages, which

typically involve markets. First, leaving the communities requires cash to pay for

transportation and lodging, although there are occasional opportunities for free logistics

(travel in Park or private vehicles and lodging with family or friends). Secondly, I

observed that most people returned to the community with goods from the city for

household use or sale. Thus, the assumption with trips is that it is an indication of market

integration due to the cash required for logistics and the opportunities presented to









participate in the market. The motive for the majority of trips outside the communities

was health-related (Table 5-1). Residents of Remanso travel more frequently and more

commonly to Pimenteiras, Brazil, which is five hours upriver from Remanso and six

hours upriver from Piso Firme. Piso Firmefios travel more often to San Ignacio, Bolivia,

which is a 12 to 24 hour journey by road. These data suggest closer ties to Pimenteiras

by residents of Remanso and to San Ignacio by residents of Piso Firme. The proximity of

Pimenteiras to Remanso implies lower transport costs and greater likelihood of

participation.

Table 5-1. Motivation for trips taken out of communities to San Ignacio, Bolivia and
Pimenteiras, Brazil by residents of Piso Firme and Remanso.
Trips to San Ignacio
Motivation for Trip Piso Firme Remanso
Health 19 2
Work 2 3
Visit Someone 4 6
Make Purchases 3 0
Total 28 11
Trips to Pimenteiras
Health 1 14
Work 0 2
Visit Someone 1 3
Make Purchases 0 3
Total 2 22

Mann-Whitney tests were used to compare socioeconomic characteristics between

the two communities. Pearson Chi-square tests for difference between distributions of

categorical characteristics that indirectly influence turtle consumption at the community

level were computed (Table 5-2). Results are discussed below.

"Education" exhibits a large variance among heads of households, with most

residents having at least two years of education and head of household in Remanso

averaging one additional year of education ( x =6.55) than in Piso Firme. "Education"

indirectly influences turtle consumption because it is likely to effect employment, which










Table 5-2. Descriptive and test statistics of socioeconomic characteristics that directly and indirectly influence turtle consumption
detailed by community. Mann-Whitney test results for H4 indicate more turtle consumption in Piso Firme for all variables
but "prefer", for which Remansefios indicated greater preference for cash equivalent.


Piso Firme
Min Max


Characteristic


Min


Remanso
Max


Std Dev


Test
Statistic


p-value


Indirect Influence
"Chaco" 0.80 0 3 0.466 53 1.03 0 25 3.17 71 U=1385.50 0.008
"Trips" 2.13 0 6 1.77 52 4.10 0 24 2.61 70 U=1761.00 0.58
"Education" 5.55 0 13 3.48 51 6.55 0 17 4.34 66 U=1464.50 0.23
"Hours fishing" 9.17 0 10 12.32 55 6.42 0 18 8.06 69 U=1640.50 0.19
"Size" 6.05 2 15 2.67 55 5.8 1 15 2.82 71 U=1887 0.75
Direct Influence
"Eat" (s) 1.38 0 7 1.52 50 0.87 0 4 0.88 71 U=1432.5 0.04
"Send turtle" (m) 0.65 0 6 1.36 48 0.13 0 4 0.62 69 U=1335.5 0.003
"Send eggs" (m) 8%=yes 34%=no 49 3%=yes 55%=no 69 x2=4.58 0.03
df=l
"Hours turtling" (s) 11.02 0 24 24.4 50 6.31 0 16 14.93 65 U=1321.50 0.03
"Sell" (m) 47%=yes 16%=no 20 16%=yes 22%=no 12 x2=3.54 0.06
df=l
"Prefer" (s, m) 20%=1 3%=2 18%=3 49 17%=1 6%=2 36%=3 70 x2=5.19 0.07
df=2
Note: Definitions of variables, c refers to characteristics that have been categorized: "Chaco" refers to number of hectares of cultivated land a
household maintains; "Trips" number of times someone from the household leaves the community for an urban destination during the year;
"Education" is years of education of head of household; "Hours fishing" if someone in the household fishes, this refers to the hours per week that
are dedicated to this activity; "Size" number of people in the household sharing food; "Eat" is the number of times the previous week that turtle
was eaten in the household; "Send turtle" the number of times someone in the household sent turtle out on the bus in one year; "Send eggs"
whether or not someone in the household sends turtle eggs out on the bus, 0=no, 1=yes; "Hours turtling" is the number of hours dedicated to
catching turtles per week by someone in the household; "Sell" whether or not someone in the household sells turtles/eggs, 0=no, 1=yes; "Prefer"
preference for turtle, or monetary equivalent of a P. unifilis (the sequence represents increasing gradations of market consumption) 1-turtle, 2=of
equal value, 3=cash equivalent. s indicates subsistence consumption variable, m refers to market consumption variable. Significant results at 95%
confidence level appear in boldface.









in turn influences time spent directly producing food or cash, and the complimentary

need for cash or food.

"Hours fishing" indirectly influences turtle consumption because it increases the

probability of dedicated or opportunistic turtle capture. Also, because most of the fishing

in these communities is a subsistence activity, it may indicate less reliance on or reduced

access to cash, prompting households to find other means of generating cash to meet their

needs, if the latter. The mean for fishing is greater in Piso Firme (9.17 hours/week) than

Remanso (6.42 hours/week).

"Size" of household also has indirect effects on turtle consumption because more

people to care for is likely to result in greater needs for food and cash or a higher

probability that someone in the household will dedicate themselves to activities that

indirectly or directly involve capture of turtles. On average, families in Piso Firme are

larger than in Remanso.

Turtle Meat Consumption in Piso Firme and Remanso

Heads of several households in each community reported daily meat consumption

in July and August (n=3 in Piso Firme, n=17 in Remanso). This data show that turtle

constitutes a large proportion of the meat consumed during this period, which is the

prime turtle consumption period because it coincides with nesting (Figure 5-2). Turtle

meat constituted a quarter to a third of all meat consumed, which illustrates its dietary

importance. Domestic meat (beef, chicken, pig and duck) also constituted large portions

of meat consumed.

All the socioeconomic characteristics directly related to turtle consumption ("eat",

"send turtle", "send eggs", "hours turtling", "sell" and "prefer") described in Table 5-2









have higher means in Piso Firme than Remanso. This indicates both greater subsistence

and market consumption of turtle in Piso Firme.

The variable "prefer" was created to determine how much market integration was

expressed by what was of more interest to household heads. The question asked was "if

you had the choice, which would you prefer-a P. unifilis turtle, the monetary equivalent

in Bolivianos, or they are the same?" The assumption made with this variable was that

greater interest in having the turtle was an indication of a more subsistence-based

household where turtles are worth more than their cash equivalent, while preference for

the monetary equivalent indicated a more market-based household where cash held a

higher value. Preference for neither the turtle nor its cash equivalent was an intermediate

state and was interpreted as not suggestive of one livelihood system over another.

In any case, turtle meat is an important part of peoples' diet in both communities-

it was reported as consumed an average of once a week per household over the course of

the year of data collection. With an average size of six in Remanso's 101 households,

this represents 31,500 meals, or 29% of the year's meals if three meals a day are

consumed. In Piso Firme, where average size is six among the 70 households and the

average number of times turtle was eaten throughout the year was 1.25 times a week,

turtle may represent the basis for 27,300 meals over the course of a year or 36% of a

year's meals. Based on the responses to the questions about "prefer" and "eat", more

turtle meat is consumed in Piso Firme.












A



wild duck
2 beef peccary
2% 15% 4%
turtle
24% chicken
tapi 4% deer
6% duck 0%
fish%
pig/ fish 2%
6% paca 31%
6%


B

wid duck
0% beef
16% peccary
turtle 1%
34% chicken
9%
deer
tapir 1%
0% duck
pig fish 3%
6% 30%
paca
0%

Figure 5-2. Proportions of meat consumed in a purposive sample of households from
July to August 2000. Note that turtle meat is the largest proportion of meat
consumed in these months. A) Sample of three households (47 meals) in Piso
Firme B) Sample of 17 households (77 meals) in Remanso.

Differences Between Subsistence and Market Consumption of Turtle in Piso Firme
and Remanso

To test hypothesis four (H4: there is more market consumption of turtle in


Remanso than in Piso Firme and more subsistence consumption of turtle protein in Piso


Firme than in Remanso) Mann-Whitney tests and Pearson chi-square tests of difference


were used to compare the variables for turtle consumption between Piso Firme and


Remanso (Table 5-2).


"Sell", "prefer" and "send eggs" were tested with a Pearson chi-square test because


they are categorical response variables. The difference between communities for "sell" is


significant (p=0.06) and indicates greater incidence of sale in Piso Firme. The test for