Hunting of birds in the Peruvian Amazon

MISSING IMAGE

Material Information

Title:
Hunting of birds in the Peruvian Amazon
Physical Description:
x, 192 leaves : ill. ; 29 cm.
Language:
English
Creator:
Begazo, Alfredo J., 1963-
Publication Date:

Subjects

Subjects / Keywords:
Fowling -- Peru -- Amazonas (Region)   ( lcsh )
Bird populations -- Peru -- Amazonas (Region)   ( lcsh )
Forest ecology -- Peru -- Amazonas (Region)   ( lcsh )
Wildlife Ecology and Conservation thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 174-191).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Alfredo J. Begazo.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030471666
oclc - 42658172
System ID:
AA00020976:00001

Table of Contents
    Title Page
        Page i
    Dedication
        Page ii
    Acknowledgement
        Page iii
        Page iv
        Page v
    Table of Contents
        Page vi
        Page vii
    Abstract
        Page viii
        Page ix
        Page x
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
    Chapter 2. Ecological variability between and within upland and Varzea forest
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
    Chapter 3. Determinants of bird harvest in the Peruvian Amazon
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
    Chapter 4. Hunting and population abundance of Amazonian birds
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
    Chapter 5. Ecological correlates of population decline
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
    Chapter 6. Conclusions and management recommendations
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
    Appendix A. Hunting data form
        Page 169
    Appendix B. Sample triads of the large-sized wildlife category
        Page 170
    Appendix C. Sample triads of the small-sized wildlife category
        Page 171
    Appendix D. Population density (individuals/km2) of species in upland forest using the program distance
        Page 172
    Appendix E. Population density (individuals/km2) of species in Varzea forest using the program distance
        Page 173
    References
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
        Page 184
        Page 185
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
        Page 191
    Biographical sketch
        Page 192
        Page 193
        Page 194
Full Text











HUNTING OF BIRDS IN THE PERUVIAN AMAZON


By

Alfredo J. Begazo












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


1999



































TO CHRISTINE, RHEA, AND KATRINA













ACKNOWLEDGEMENTS

This study could not have been carried out without the help of numerous people.

I thank members of my committee Dr. Richard Bodmer, Dr. Michael Moulton, Dr.

Kathryn Sieving, Dr. George Tanner, and Dr. Douglas Levey for their advice and

guidance. Special thanks go to Dr. Richard Bodmer for his assistance in every aspect of

this dissertation.

The data collection was assisted by a number of people. I thank Rolando Aquino

for helping with the collection of census data. Julio Curinuqui, Gilberto Asipali, Pablo

Puertas, Cesar Reyes, Michael Valqui, and Richard Bodmer also contributed in

significant ways in conducting the censuses. Etersit Peso conducted surveys in the

markets in Iquitos. Jorge Beltran, Einer Vela, and Ramon Noa provided valuable field

assistance. Their humor and patience to endure long stops along the trails, as well as,

their ability to provide fresh fish was important.

I thank all households at the villages of San Pedro and Nueva Esperanza for their

dedication in recording information required for this study. The Chistama family in

Nueva Esperanza and Ruiz-Huanaquiri in San Pedro provided room and board while I

was at the villages and made sure that there was food or at least masato left for me,

regardless the time I returned to the household.

Other people also participated significantly in this dissertation. Juan Ruiz

Macedoneo of Universidad Nacional de la Amazonia Peruana assisted with the

identification of seeds in the stomach contents. Pablo Puertas shared his data and was








always willing to share his knowledge on hunting by riberefio people and other aspects of

riberefio culture.

Logistical support was crucial in conducting this study. Access to remote

unhunted areas was made possible through the invaluable support of Dr. Richard Bodmer

and the crew of the "Nutria". Dofia Vilma, Julio Curinuqui, and Gilberto Asipali helped

with food preparation, and trail cutting, and they shared their knowledge of the forest.

In Iquitos the home of Jose "Pepe" Alvarez and Elena Burga became the base of

my constant travels and provided a sense of belonging during long stays away from

home. The Hostal La Pascana and Dofia Petty Sanati also provided a nice environment

during my first trips to Iquitos.

I would like to extend my appreciation to people who influenced my decisions

over time: First my parents for believing that whatever I did -even chasing birds around-

was OK, as long as I was excited about it; my sister Suzy Begazo for her strong will and

determination; Dr. John Fitzpatrick who was not involved directly in this study, but

whose early influence played a significant role in my getting to this point.

My deepest gratitude goes to Christine Kelly for whom I could not count the ways

in which she was such a key component in "us" getting a doctoral degree. Christine

provided insight, maintained my focus, was patient during trying times, and continues to

be a source of inspiration. Her cheerfulness and ability to help out in periods of great

stress were essential to the completion of this study. Dawn Kelly was also very important

in relieving me from household chores, especially during the writing process. My

daughters Rhea, and Katrina wondered what I could possibly be doing in front of that








computer for so long when the sun was shining and when playing in the back yard was

more fun. I thank my family for being there.

I thank the Biodiversity support program and the Chicago Zoological Society for

their financial support. The Fundaci6n Peruana para la Conservaci6n de la Naturaleza

also provided financial support for the initial phase of this project.














TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS......................................................................I

ABSTRACT..................................................................................... VIII

CHAPTERS

I INTRODUCTION................................................................................ 1
Introduction ................................................................. ............... 1
O verview .......................................... ......................................... 5
The Varzea and Upland Forests.......................................................... 5
Study D esign.............................................................................. 17

2 ECOLOGICAL VARIABILITY BETWEEN AND WITHIN UPLAND AND
VARZEA FOREST....................................................................... 25

Introduction................................................................................ 25
Methods.................................................................................... 27
R esults...................................... ............................................... 31
D iscussion................................................................................. 38

3 DETERMINANTS OF BIRD HARVEST IN THE PERUVIAN AMAZON......... 45

Introduction........................................................................ .... 45
Methods ................................................................................. 46
R esults.................................................................... ................. 56
Discussion ................................................................................. 77

4 HUNTING AND POPULATION ABUNDANCE OF AMAZONIAN BIRDS...... 84

Introduction ..................................... ......................................... 84
Methods................................................................................. 85
R esults.................................................................................... 100
D discussion ............................................. ................................... 113








5 ECOLOGICAL CORRELATES OF POPULATION DECLINE..................... 121

Introduction.............................................................................. 121
M ethods .................................................................................. 123
R esults................................................................ .................... 130
D discussion ................................................................................ 150

6 CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS................. 161

Management recommendations......................................................... 164

APPENDICES

A HUNTING DATA FORM................................................................... 169

B SAMPLE TRIADS OF THE LARGE-SIZED WILDLIFE
CATEGORY.............................................................................. 170

C SAMPLE TRIADS OF THE SMALL-SIZED WILDLIFE CATEGORY............ 171

D POPULATION DENSITY (INDIVIDUALS/KM2) OF SPECIES IN UPLAND
FOREST USING THE PROGRAM DISTANCE................................... 172

E POPULATION DENSITY (INDIVIDUALS/KM2) OF SPECIES IN VARZEA
FOREST USING THE PROGRAM DISTANCE................................... 173

REFERENCES.................................................................................... 174

BIOGRAPHICAL SKETCH.................................................................... 192













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

HUNTING OF BIRDS IN THE PERUVIAN AMAZON

By

Alfredo J. Begazo

August 1999


Chairman: Dr. Richard E. Bodmer.
Major Department: Wildlife Ecology and Conservation

Bird hunting by riberefio people (rural Amazonians) in the Peruvian Amazon was

studied in two of the most important Amazonian ecosystems; the upland and the varzea

forests during May 1996 and February 1998. Upland and varzea forests are subject to

different water regimes, which influence the ecology of wildlife and the activities of

riberefio people. Birds are an important source of animal protein for riverefio people.

However, the influence of hunting on bird populations in upland and varzea forests is

unknown. The goal of this study were; to examine the association between hunting by

humans and changes in population abundance of Amazonian birds; and to examine

ecological traits associated with changes in population abundance from hunting. The

objectives included the following: (1) to assess ecological variability between and within

upland and varzea forests in terms of vegetation structure and availability of selected

food resources, (2) to determine the harvest of birds in two villages established in upland

and varzea forests, and discuss these harvests in terms of cultural and biological factors,


VIII








(3) to determine the influence of hunting on populations of Amazonian birds, and to

evaluate the sustainability of current hunting in the two villages, and (4) to examine

ecological traits associated with the ratio of change in population abundance between

unhunted and hunted populations. Upland and varzea forests appear to differ

ecologically between forest and within forest type. Results suggest that canopy and

understory structure, and leaf litter cover differed between forest type, but were similar

within forest types. Fruit abundance differed between sites sampled within upland and

varzea forests. Village residents at San Pedro (SP) in the upland forest harvested a

greater proportion of birds (669 birds of 29 species hunted by 115 people) than those at

Nueva Esperanza (NE) in the varzea (955 birds of 24 hunted by 380 people). Hunters

preferred large birds that had good quality of meat, such as the Cracidae (cracids),

Tinamidae (tinamids), and Psophiidae (trumpeters), but the actual harvest indicated that

hunters took what was available to them. Hunters took small and less preferred birds

from the impacted fauna surrounding the settlements, and realized their preferences in

areas less impacted, away from the settlements. Hunting appeared not to influence the

populations Amazonian birds equally. Large cracids; Mitu tuberosa, Penelope jacquacu,

Pipile cumanensis, and trumpeters; Psophia leucoptera showed the greatest ratio of

population change between unhunted and hunted populations. Species of tinamous and

0. guttata did not show significant changes in population abundance between unhunted

and hunted populations. Changes in population abundance observed in all species were

consistent in the upland and the varzea forests. Species of the Cracidae and the trumpeter

appeared to be exposed to disproportional hunting pressure because of human

preferences, naive behaviors, and ease of detection. Hunters also prefer species of the








tinamous, but tinamous show secretive habits and efficient antipredator behaviors that

appear to prevent overhunting of their populations. Changes in population abundance

observed among species appear to have a biological explanation. Large cracids are

preferred and easier to hunt, but are more vulnerable to population decline because of low

intrinsic rates of population increases (rmax), long-lived individuals, and long generation

times. In contrast the tinamous have high rmax, short-lived individuals and short

generation times. Amazonian birds can be categorized in terms of their vulnerability to

hunting using the correlations between cultural and biological factors described in this

study. Recognizing bird populations' potential for harvest could help plan management

strategies toward their sustainable use and their conservation.













CHAPTER 1
INTRODUCTION

Introduction

In Amazonia, hunting has taken place since prehistoric times (Perea 1995) and

rarely caused local extinction because native Amazonians practiced nonselective and

nomadic hunting in small groups (Dufour 1990). Recently, local extinctions from

overhunting have become more common because forest-dwelling people have

concentrated in sedentary settlements, adopted modem hunting technology and become

involved in market economies (Robinson and Redford 1991, Padoch 1988, Stearman

1990). The association between hunting and patterns of population decline has been

documented on primates (Peres 1990, Peres 1996b, Coltrane 1998), birds (Silva and

Strahl 1991, O'Neil 1991, Mitchell and Raez-Luna 1991, Gonzales 1998), and ungulates

(Peres 1996a, Bodmer et al. 1997) by comparing population abundance in areas of

differing hunting intensities (Peres 1990, Glanz 1991). While most studies have

concentrated on the immediate influence of hunting on population abundance, the reasons

that some populations remain stable while others experience decline have been largely

neglected.

After habitat destruction, overhunting is the second most common cause of recent

extinctions (Atkinson 1989, Reid 1992). But extinction is rarely cataclysmic; rather, it is

a progressive process with total extinction preceded by local extinction and local

extinction preceded by population decline (Caughley and Gunn 1996). Hunting pressure








often leads to population decline because it removes reproducing individuals that result

in lower productivity at the population level (Caughley 1977). Hunting pressure is

potentially influenced by the abundance of wildlife available to hunters (Robinson and

Redford 1986, Bodmer 1995b), hunters preference, spiritual and physiological

prohibitions on the taking of certain species (Ross 1978, Ayres and Ayres 1991), hunting

technology (Hames 1989), and ease of hunting (Fitzgibbon et al. 1996, Bengstson 1984).

However, a given hunting pressure does not have similar implications on change in

population abundance across species because species differ in their ability to compensate

for individuals removed from the population (Robinson and Redford 1991, Bodmer et al.

1997, Fitzgibbon et al. 1996).

Few studies have addressed the implications of avian wildlife utilization among

rural Amazonians. Mitchell and Raez-Luna (1991) and Gonzales (1998) determined

associations between hunting and population abundance of birds by comparing

abundances observed in sites with differing intensities of hunting. Other studies

quantified avian utilization by rural Amazonians to address the hypotheses of protein

limitation (Gross 1975), hunting effort (Vickers 1980), ideas of taboos (Ross 1978),

changes in hunting attitudes (Ayres et al. 1991), importance of wildlife for rural people

(Gaviria 1981, Saravia 1992, Townsend 1995), and the effect of forest fragmentation

(Cullen 1997). Most of these studies showed a clear bias towards including few of the

large species (e.g., cracids and tinamous) while ignoring most of the smaller and less

frequently hunted taxa in their analyses.

Most Amazonian ecosystems are inhabited by rural people known in the Peruvian

Amazon as riberefios or river people (FPCN 1994). Two of these ecosystems are upland








and varzea forests that make up most of western Amazonia (Malleux 1975) and are

subject to differing seasonal riverine floods. Upland forests stand on areas of high

ground and are not subject to riverine floods, whereas, varzea forests stand on low and

flat terrain that is seasonally flooded for 2-4 months (Pires and Prance 1987). Differing

flood regimes have significant implications on the plant community (Gentry 1988),

population characteristics of mammalian wildlife (Emmons 1984, Peres 1996b, Bodmer

et al. 1997) and the activities ofriberefo people (FPCN 1994).

Addressing the implications of avian wildlife utilization by rural Amazonians is

potentially important in promoting forest conservation. After small-scale agriculture,

hunting is among the most important activities for riberefio people. But given the

increasing human populations in sedentary settlements (Egoavil 1992), coupled with

economic pressures (Padoch 1988), it is less and less likely that uncontrolled harvesting

of wildlife is sustainable (Shaw 1991). To understand the influence that hunting has on

avian wildlife, it is essential to determine how cultural and biological factors interact with

ecological characteristics of Amazonian ecosystems. Identifying the reasons that lead to

decline and those that lead to population stability can assist in developing general

principles through which managers can direct their effort when addressing the use and

conservation of Amazonian ecosystems.

In this dissertation, I examined the hunting of Amazonian birds and its association

with their population abundances, as well as the proximate factors of population decline

from hunting. I tested the hypotheses that hunting influence population of Amazonian

birds and that species whose populations are influenced by hunting are ecologically

different from others. The association between hunting and population abundance of








Amazonian birds was examined by comparing abundances observed in hunted and

unhunted sites selected in varzea and upland forests. Habitat in each site within upland

and varzea forest was assumed to be constant and population abundances similar among

sites prior to hunting by humans. If hunting influences population abundance, I would

predict that; one, rural Amazonians hunt birds in considerable number; two, population

abundance in hunted sites be lower than those in unhunted sites. I first test the

assumption that habitat within forest type are similar by quantifying the availability of

food and forest structure in hunted and unhunted areas within upland and varzea forests.

Then, I quantified the harvest of birds. Subsequently, I compared population abundances

observed in sites with differing intensities of hunting and populations in unhunted or

control sites.

Vulnerability to human hunters was tested by quantifying antipredator behavior

assuming that the reactions performed by birds in response to my presence was similar to

the reaction that birds would perform in the presence of an actual hunter. Species

potential for population recovery from hunting was tested by obtaining species'

reproductive parameters. Reproductive parameters obtained from captive-bred

populations were assumed to be similar to those of wild populations. If there were traits

that influence the vulnerability of some species to human hunters, I would predict that

species that show these traits support a greater harvest pressure than others. If species

differ in their potential for population recovery from hunting, I would predict that species

that show greater potential for population recovery show a lower change in abundance

between unhunted and hunted populations. I first tested whether species differ in their

vulnerability to hunting by examining the harvest of birds in terms of the traits expected








to influence the species vulnerability to human hunters. Subsequently, I tested whether

species differ in their potential for population recovery from hunting by examining

reproductive parameters in terms of changes in population abundance between unhunted

and hunted populations. Finally, I synthesized results and provided management

recommendations.



Overview

The Varzea and Upland Forests

Varzea forest Varzea forest is defined as Amazonian forest subject to seasonal

riverine floods. Varzea forests in Western Amazonia and the Peruvian Amazon are

subject to periodic inundation for 2-4 months of the year (Kalliola and Puhakka 1993).

Varzea forests occupy vast surfaces of mostly flat terrain that parallels major Amazonian

rivers (floodplains). Varzea forest includes lakes, lagoons, levees, backswamps and,

channel bars that vary in the length of inundation from permanent to temporary,

depending on the geography of the terrain (Figure 1-1). The length of floods largely

depends upon changes in water levels of major Amazonian rivers (Encamrnaci6n 1993).

Upland forest Upland forest is defined as forest not subject to periodic

inundation. Upland forest stands on high ground (Pires and Prance 1987). Upland forest

has been classified according to its physiography as a) terrace forest, with rolling hills up

to 30 m of altitude, and slopes of 15-30%, and b) hilly forest, with hills of up to 180 m

and slopes of 70-80% (Malleux 1975, Encamrnaci6n 1993).
























Maximum Level


Figure 1-1. Typical cross section of (A) upland forest and (B) varzea forest.








Climate

The climate in the Peruvian Amazon is relatively uniform. The annual median

temperature varies between 25.9C and 26.8 C (Table 1-1). The fluctuation in

temperature between day and night is greater than that between summer (March, the

coolest month: 25C) and winter (September-December, the warmer months:27.4 -

26.9) (Hueck 1978). On occasion, during the months of June and July, cold fronts from

the south move to the region, causing the temperature to drop to 10-5C (Marengo

1984). Atmospheric humidity is nearly constant, varying from 80% to 100%.



Table 1-1. Data from weather stations in 5 cities in the Peruvian Amazon.
Altitude Annual
City A Location Temperature (C) Precipitation
____ (mm)
Median Max. Min.
Iquitos 117 0345"S, 7312"W 25.9 15 35.5 2600

Yurimaguas 180 0545"S, 7605"W 26.8 11.3 37.8 1774

Pucallpa 148 0825"S, 7437"W 26.7 15.1 39.2 1399
Puerto
Plerto 256 1235"S, 69012"W 26 10 28 1423
Maldonado
Tarapoto 426 0632"S, 7619"W 25.9 11 38 1330
Source: Kalliola et al. 1993.



Changes in water level in the rivers and ultimately rainfall determine seasonality

in the region. Months with high precipitation are called winter, while periods of low

precipitation are called summer (Marengo 1984). However, rainfall at the regional level

has little influence on the water level of rivers. Instead, climates in the Andean Mountain

chain largely determine the fluctuation of water discharge in Western Amazonia (Kalliola








and Puhakka 1993). The maximum water discharge of rivers in the state of Loreto is

recorded during the months of October-April (Kalliola and Puhakka 1993), and water

levels in the rivers fluctuate between 2 andl5 meters (Dumont et al. 1990); (Figure 1-2).



Water Properties

Amazonian rivers are classified according to the physical and chemicalproperties

of their waters. This classification is based on differences in suspension loads and tannin

content. White-water rivers originate in the Andean Mountain chain (Sioli 1984), water

has a large load of suspended sediments and is circum neutral in pH. Black-water rivers

originates in water catchments in the lowland forest, water is low in suspended sediments

and nutrients, but rich in humic substances and tannins. Black-water has an acid pH.



Vegetation

Rainfall and periodic floods influence nutrient availability and plant diversity of

upland and varzea forests (Gentry 1988, Kapos et al. 1990). In upland forest, rainfall

washes nutrients away, resulting in nutrient deficient soils (Encarnaci6n 1993). On the

other hand, seasonal floods from white water rivers replenish nutrient in varzea forest

where nutrient availability appears not to be a limiting factor (Ayres 1993).

While soils in varzea forest are richer in nutrients, upland forest has greater

botanical diversity. Gentry (1988) concluded that in a 0.1 ha plot surveyed near Iquitos

(Peril), the number of plant species common to varzea and upland forest was very low. In

the Ecuadorian Amazon, Balslev et al. (1987) listed 60 species (18%) common to

adjacent upland and varzea forest of a total of 333 species recorded. Pires and Prance















120.5

118.5

116.5 1983

I1982
S114.5 .'
14 *... 1987

S 112.5 . .-
I \ x'-. ..- /:S
110.5 '- s 19
108.5 1

106.5 "/1

104.5 ,-- ,,, -,
Jan Fcb Mar Apr May Jun Jul Aug Scp Oct Nov Dec



Figure 1-2. Seasonal fluctuation of the water level of the Amazon river. Data
taken at the city of Iquitos, Istituto Nacional de Meteorologia e Hidrologia
(from Kalliola et al. 1993).







(1987) also found similar results in a comparison of upland and varzea forest in the

Brazilian Amazon. In overall terms, the richest ecosystems in plant diversity

undoubtedly are in upland forest (Gentry 1992).

Plant distribution is more homogeneous in varzea than in upland forests. Because

most plants in varzea forest rely on water and fish for dispersion (Goulding 1980), plants

are usually more evenly dispersed. In contrast, topography and microclimate are

associated with uneven distribution of plant species in upland forest (Hubbell and Foster

1986). Finally, topography, edaphic conditions, microclimate and hydrology were used

to distinguish 7 plant communities in upland forest, and 9 in varzea (Encamrnaci6n 1993).



The People

The human population in rural Amazonia is composed of indigenous and mestizo

people (mixed ethnic background) that are known in Peru as riberefios or river people

(Egoavil 1992). The origin of the riberefios can be traced to the exploitation of the

Peruvian Amazon during the 19th century. The transformation of Amazonian people

from tribal groups to riberefios began with the earliest European immigration, and

continued with the detribalization process imposed by missionaries, expansion of the

slave trade, and influx of immigrants during the rubber boom. Due to the need for labor

and the promise of prosperity, rubber exploitation attracted people from the Peruvian

coast, the Andean region, Brazil, Europe, North America, Asia, and detribalized

Amerindians (Dourojeanni 1990). Hence, the non-native population increased from

18,000 in 1876 to 120,000 in 1920 (San Roman 1975). After the abrupt ending of the

rubber exploitation, many immigrants returned to their place of origin, but others








established themselves in the Peruvian Amazon where they turned their attention to

other extractive activities (San Roman 1975).

Currently, riberefios greatly exceed indigenous people. Data from the 1981

census revealed that 280,000 nontribal people live in the rural sector of the state of

Loreto, comprising 85 percent of the entire rural population. The remainder 5% are

mainly indigenous people with a population of 50,000 (Egoavil 1992). Ribereiio villages

are located at the margins of major rivers and some are more than a century old (FPCN

1994). Indigenous groups and riberefios practice agriculture, fishing, hunting, small-scale

lumber extraction, and collection of minor forest products. However, riberefios are more

involved in market economies, in both the regional and international level, than are

indigenous groups (Padoch 1988).

In the political arena, ribereflos have been frequently forgotten or mistaken for

colonists (rural Amazonians that emigrated from urban areas); (Padoch 1988). Human

settlements in Amazonia are governed by local authorities and are, for the most part,

recognized by the Peruvian government (FPCN 1994). Although riberefilo villages

outnumber those of indigenous people, studies on riberefio culture and their relationship

with their natural environments have received little attention (Padoch 1988). Riberefios

are the traditional, and in many cases, the exclusive users of many Amazonian

environments, particularly the floodplains. Thus, the study and understanding of

riberefios, and their use of resources, will enhance the possibilities of sound conservation

policies for Amazonia.








The Birds

Twenty two species of birds from 9 orders and 13 families are treated in this study

(Table 1-2). Species included in this study were reported taken by hunters in exploratory

surveys in both human settlements. Four avian families occur only in the New World,

and 89% of species are endemic to the humid lowland forest of Amazonia (Stotz et al.

1996). Species were classified in three size categories: small < 0.6 kg, medium > 0.6 -

1.5 kg, and large > 1.5 kg.



Hunting

Hunting is defined as the customary and traditional practice of procurement of

wild animals for direct personal or family consumption as food and other non-edible by-

products.



Background

The hunting of wild animals is important to the livelihood of rural Amazonians.

Food (Dufour 1990, Alvard 1995), non-edible products (e.g., skins, hides, feathers) (Cajal

1988), and live animals (e. g., pets and biomedical research); (Inigo-Elias and Ramos

1991; Thomsen and Brautigan 1991) rank among the most important uses. Hunting of

animals also is important for social reasons. For example, studies have shown that the

type and amount of animals hunted is important to a hunter's prestige and the social

cohesion of the family and village (Balee 1989, Stearman 1990).

Many studies on wildlife use have documented the harvest of birds among

indigenous people in Amazonia. For the most part, anthropologists have included birds.







Table 1-2. Bird species included in this study.

English name Fam s Species Body Altitude Foraging tat
English name Order Family Genus Species mass (gr) (m)a strata b Habitat


White-throated Tinamou

Gray Tinamou

Bartlet's Tinamou

Cinereus Tinamou

Undulated Tinamou

Variegated Tinamou

Speckled Chachalaca

Spix's Guan

Blue-throated Piping Guan

Wattled Curassow

Razor-billed Curassow

Starred Woodquail


Moscovy Duck


Tinamiformes

Tinamiformes

Tinamiformes

Tinamiformes

Tinamiformes

Tinamiformes

Galliformes

Galliformes

Galliformes

Galliformes

Galliformes

Galliformes

Anseriformes


Anatidae


Tinamidae

Tinamidae

Tinamidae

Tinamidae

Tinamidae

Tinamidae

Cracidae

Cracidae

Cracidae

Cracidae

Cracidae

Phasianidae


Cairina


Tinamus

Tinamus

Crypturellus

Crypturellus

Crypturellus

Crypturellus

Ortalis

Penelope

Pipile

Crax

Mitu

Odontophorus


guttatus

major

bartletti

cinereus

undulatus

variegatus

guttata

jacquacu

cumanensis

globulosa

tuberosa

stellatus

moschata


600 L- 850

1050 L-1000

241 L-500

450 L 700

540 L -900

384 L- 900

500 L-1800

1282 L-1800

1425 L-900

2500 0

3060 L 900

310 L-1050


2468 L -1000 W Open habitats


Sources: Delacour and Amadon (1973), Hilty and Brown 1986, Dunning (1992), Stotz et al. (1996).


Interior forest

Interior forest.

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest

Interior forest










Table I1-2 continued

English name Order Family Genus Species Body Altitude Foraging Habitat
Engishnam Orer amiy GnusSpeies mass (gr) (m)' strata b H

Toucans Piciformes Ramphastidae Ramphastos spp d 400 L -1250 C Open habitats

Amazons Psittaciformes Psittacidae Amazona sppt 486 L 1200 C Open habitats

Macaws Psittaciformes Psittacidae Ara sppf 1130 L 900 C Open habitats

Gray-Fronted Dove Columbiformes Columbidae Leptotila rufaxilla 157 L 1900 T Interior forest

Pale-winged Trumpeter Gruiformes Psophiidae Psophia leucoptera 990 L -1050 T Interior forest

Neotropical Cormorant Pelecaniformes Phalacrocoridae Phalacrocorax brasilianus 2010 L 3500 W Open habitats

Anhinga Pelecaniformes Anhingidae Anhinga anhinga 1235 L 900 W Open habitats

White-necked Heron Ciconiiformes Ardeidae Ardea cocoi 3200 L 900 T/W Open habitats

Boat-billed Heron Ciconiiformes Ardeidae Cochlearius cochlearius 656 L 800 W Open habitats
Altitudinal distribution where L= Lowland humid forest. Upper limits denote distribution into the Andean mountain chain.
b Foraging strata where: C=forest canopy, T=--terrestrial, U=forest understory, and W=water.
SSpecies of open areas are those that typically inhabit the upper strata of the forest canopy or bodies of water.
d Includes Ramphastos cuvieri and R. culminatus.
C Includes Amazonafarinosa and A. ochrocephala
f Includes Ara chloroptera, A. macao, and A. Ararauna.








in their analyses of wildlife use to test hypotheses of protein limitation (Gross 1975),

efficiency of different weapons (Hames 1979), optimal foraging by humans (Hawkes et

al. 1997), harvest per unit effort (Safirio and Hames 1983, Stearman 1990), ideas on

taboos (Ross 1978), and various aspects of the use of wildlife (Rudle 1970, Yost and

Kelley 1983, Townsend 1995). While riberefio people constitute the majority in the

Amazonian rural sector, relatively little effort has been dedicated to the study of their use

of wildlife (Smith 1976, Ayres et al. 1991).In the Peruvian Amazon, studies on uses of

wildlife have focused on mammals (Bodmer 1990, Coltrane 1998), or have included birds

only as a marginal component of their analyses (Saravia 1992, Gaviria 1981). Birds

undoubtedly are important to the livelihood of rural people in Amazonia. Birds,

especially the cracids (Cracidae) and tinamids (Tinamidae) ranged from 9.7% to 56% of

all animals harvested in the most representative human groups in Amazonia (Ojasti

1993). However, little attention has been paid to the influence of hunting on populations

of Amazonian birds and the factors that influence population decline from hunting.

Hunting of Amazonian wildlife typically causes a decrease in their population

densities. This has been documented in mammals (Glanz 1991, Bodmer 1994a, Peres

1996b, Alvard et al. 1997, Coltrane 1998) and birds (Mitchell and Raez-Luna 1991).

Hunting influences densities of wild populations located closer to human settlements

more than than it does in areas distant from them (Mitchell and Raez-Luna 1991, Hill and

Hawkes 1983, Stearman and Redford 1995). Furthermore, large primates (Peres 1990,

Bodmer 1995b) and large birds (Cracidae and Tinamidae); (Dobson and Ortiz 1988,

Mitchell and Raez-Luna 1991, Silva and Strahl 1991) seem to experience relatively

greater rates of population reduction than other species.








While hunting typically reduces population size, natural variation in population

density may influence the magnitude of the impact and survivorship of populations

(Harper and Begon 1990). This is true because hunting would take a greater proportion

of individuals from a small population than from a large population (Caughley 1977).

Wildlife densities differ throughout Amazonia (Emmons 1984). Bodmer (1990)

suggested that riverine floods in varzea forest explain the lower densities of ungulate

species not adapted to semi-aquatic life. In contrast, riverine floods deposits nutrient rich

sediments on soils in varzea forest, and rich soils in varzea forest have been used to

explain greater mammalian (ground and arboreal dwelling species) biomass than in

upland forest (Emmons 1984, Peres 1997).

Environmental characteristics of upland and varzea forest also has an influence on

local people. Varzea forest has rich fisheries, and fish is the most important source of

animal protein to local people (Gonzales 1998). In contrast, upland forest has small

rivers that typically have smaller and less available fish (FPCN 1994). Moreover, soils in

varzea forest have greater potential for agriculture than soils in upland forest

(Dourojeanni 1990).

While some species experience population decline in response to anthropogenic

disturbance, others maintain stable populations (Balmford 1996). Some species are

preferred by hunters and vary in the extent that they are detected (Fitzgibbon et al 1996),

and in the ease of hunting because they forage at different elevations within the vertical

height of the forest, and perform antipredator behaviors that differ in efficiency to reduce

the risk of human predation. These factors influence how often individuals of a species

are taken by hunters. While greater hunting pressure often leads to population decline








(Caughley and Gunn 1996), species differ in their reproductive biology and the capacity

to compensate for individuals removed from the population.

Birds have important ecological roles in Amazonia. Most species hunted by

riberefios are frugivores and are important seed dispersers and seed predators (Janzen

1974, Terborgh 1986, Redford et al. 1992, Strahl and Grajal 1991). In addition,

frugivorous birds have been found to contribute 60 to 70% of the total biomass of avian

seed dispersers in most tropical bird communities (Janzen 1974, Terborgh 1986).

In the Peruvian Amazon, the government regulates the use of wildlife.

Overhunting of species with valuable hides (e.g., some ungulates, otters and spotted cats),

during 1950-1970, led to their dramatic population declines (Dourojeanni 1990).

Consequently, in 1973 the Peruvian government prohibited commercial hunting in the

Peruvian Amazon, and allowed hunting of a limited number of species for local

consumption in the rural sector only (Decree: D. S. 934-73-AG). Legislation also

curbed the harvest of commercially used species. However, hunting for local

consumption remains largely unmanaged. While some groups in the Neotropics use

wildlife at sustainable levels (Dufour 1990, Vickers 1991), there is evidence that other

people hunt wildlife at levels that may not be sustainable (Alvard et al. 1997).

Amazonian wildlife must be managed, but it poses a challenge given the complexity of

Amazonian ecosystems (Freese 1997).



Study Design

A case study design was used to address the relationship comprising hunting and

its influence on Amazonian birds. This natural experiment embodied two hunting








systems located in the major landscape characteristics (upland and varzea forest) and

socio-economic features (riberefio culture) of the Peruvian Amazon (Padoch 1988). This

study examined the association between hunting and population abundance of

Amazonian birds through a comparative observational design (Cook and Campbell 1979)

involving populations subject to differing intensities of hunting and unhunted or control

populations.



Research objectives

Hunting systems in upland and varzea forest offer an appropriate context in which

to examine the hunting of Amazonian birds and its association with their population

abundances, as well as the proximate factors of population decline from hunting. Bird

species included in this study are mostly frugivorous, occur in upland and varzea forest,

and use all dimensions of the forest (from the forest floor throughout the canopy).

Additionally, upland and varzea forests are subject to differing environmental factors that

influence characteristics of the forest (Hubbell and Foster 1983), ecology of wildlife

(Emmons 1984, Peres 1996b, Bodmer 1990) and activities of local people (FPCN 1994).

The current study was designed specifically to test the relationships comprising

hunting of birds and its association with population abundance, and the proximate

reasons that influence population decline from hunting.

The overall research questions of this study are the following:

1.) Is hunting associated with changes in population abundance of Amazonian
birds?
2.) Are species vulnerable to hunting ecologically different from others?

To test the hypothesis that hunting pressure influences populations of Amazonian birds, I








compared population abundances observed in sites with differing intensity of hunting

with abundances in unhunted or control sites. If hunting influences populations of

Amazonian birds, I would predict the following: one, riberefio people would harvest birds

in considerable numbers; two, population size would covary with the intensities of

hunting; three, population size of Amazonian birds would be lower in hunted than in

unhunted sites.



Assumptions:

In testing these hypotheses I assumed that I controlled for habitat differences by

choosing sites in upland and varzea forests. I also assumed that I controlled for intensity

of hunting within upland and varzea forests by choosing sites that differed in intensity of

hunting. Farther assumptions are that; one, the intensities of hunting in sites selected

does vary as thought; two, population abundances in the sites with the differing

intensities of hunting were similar prior to hunting by humans.

To test the hypothesis that species vulnerable to hunting are ecologically different

from others, I examined whether species differed in traits that are associated to

disproportional hunting pressure, and traits that influence their capacity to recover from

individuals removed from the population. If hunting is associated to changes in

population abundance, I would predict that; one, some species show traits that result in

differential exposure to human hunters, and are hunted more often than others; and two,

species differ in their capacity of recovery of individuals removed from the population,

and show lower variation in population abundance between unhunted and hunted

populations.








Assumptions

In testing whether some traits are associated with disproportional hunting pressure

of some species, I quantified species antipredator behavior assuming that antipredator

behavior performed by a bird in reaction to my presence would be the same as those they

would perform in response to a true hunter. In testing whether species differ in their

potential for recovery from harvest, I assumed that parameters obtained from captive-

bred populations resemble those of wild populations.



Alternative Hypotheses

Hunting by humans may influence the populations of different species in different

ways because of; one, differences in human population density, age of the human

settlement, and settlement patterns between the hunting systems chosen in upland and

varzea forest; two, preference of hunters; three, availability of alternative sources of

animal protein that influence hunting pressure on birds; four, habitat modification in the

surrounding of the human settlements that influence food base of certain species; and

five, differences in population abundance between forest type.

I conducted fieldwork during 1996-1998. The results formed the basis for the

chapters in this dissertation. In chapter 2, I discussed characteristics of upland and varzea

forest ecosystems. I tested the hypothesis that:

- Upland and varzea forests differ ecologically between forest type, but are similar
within forest type.

In this chapter, I discussed forest structure and abundance of food resources, as these

ecological traits are likely to influence the occurrence and abundance of species. This

discussion formed the basis for comparisons between harvest patterns, population








abundances, and the associations of hunting with population abundances.

In chapter 3, I discussed the harvest of birds, and the factors that influenced the

harvest of birds. The specific questions answered were:

Do hunters prefer some species of birds to others?

Do hunters show similar hunting patterns between forest type?

Are species hunted at the same intensity within and between forest type?

Do search patterns influence the harvest patterns of birds?

Are characteristics of the human settlement associated with harvest patterns of
Amazonian birds?

This information on wildlife use formed the basis for examining the association between

hunting with abundance of bird populations presented in chapter 4.

In chapter 41 discussed the association between hunting and changes in population

abundance. Subsequently, I evaluated the sustainability of hunting of Amazonian birds.

The specific questions answered were:

- Is hunting associated with variation in population abundance of Amazonian birds?

- Is current harvest of birds sustainable?

This information provided a measure of the association between hunting with populations

of Amazonian birds.

In chapter 5, I discussed the reasons that some species were more susceptible to

hunting than others. The specific questions answered were:

- Do species show traits that are associated with a disproportional vulnerability to
human hunting?

- Do species differ in their capacity of population recovery from individuals removed
from the population?

In chapter 6 1 summarized the result of this study with respect to the association








between hunting and population abundance of Amazonian birds, and the factors that

influence population decline from hunting. Finally, I provided management

recommendations based on the results obtained in this study.



Study Area

The comparative method required adjacent sites that have areas hunted (=hunted

areas), as well as areas with little or no hunting unhunteded areas). The protected areas of

Tamshiyacu-Tahuayo Communal Reserve (TTCR) located in upland forest, and Pacaya-

Samiria National Reserve (PSNR) located in varzea forest, both in the state of Loreto,

permitted this comparison. Areas within TTCR and PSNR were compared to hunted

areas located in their periphery. Village residents of the village of San Pedro (SP) in

TTCR hunted in areas within 25 km radius from the village whereas resident in Nueva

Esperanza (NE) in PSNR hunted in areas within 18 km radius from the village. Areas

within TTCR and PSNR were assumed to be subject to little or no hunting. This

assumption is based on these sites being located in remote areas, and on their protected

status. In upland forest, the center of the sampling activities at the unhunted area was

located at S0421'6.2', W7151'04.9". The center of sampling activities at the hunted

area was located at S04020'14.3", W73I '53". The centers of the hunted and the

unhunted areas are separated by a distance of 135 km (Figure 1-3).

Site configuration in the varzea forest shows similar characteristics. The center of

sampling activities in the unhunted area was located at S0415'4.5" W7549'30.9", and












































Legend
E Pacaya-Samiria National Reserve
7* (Varzea forest)
E Tamshiyacu-Tahuayo Communal Reserve
(Upland Forest)
I Hunted areas, A=EI Pinche, B=Sta Elena.


1 Unhunted areas, C=Q. Blanco, D=Yavari.
0 City [
760 75 74 *



Figure 1-3. Map of the Pacaya-Samiria National Reserve and Tamshiyacu-Tahuayo
Communal Reserve showing the location of the study sites.








the center of sampling activities at the hunted area was located at S0551'13.3"

W7505'17.1". The centers of hunted and unhunted areas in varzea forest are separated

by a distance of 43.4 km. Finally, there is a straight-line distance of approximately 230

km between the general areas sampled in upland and varzea forest. (Distances were

estimated with a Geographic Positioning System (GPS).

PSNR encompasses 2,080,000 hectares of mostly flood prone forest (91.6%) and

levees and terraces that escape inundation (6.5%); (Malleux 1975). PSNR lies within the

Ucamara depression. The Ucamara depression encompasses 60,000,000 to 70,000,000

hectares of predominantly flat terrain that includes the confluence of the Marafion

andUcayali rivers at the point of origin of the Amazon River (Villarejo 1988). The

Ucamara depression is seasonally flooded by nutrient-rich white water that flows from

the Andean mountain chain (Sioli 1984). The geological history of the Ucamara

depression reveals an active lateral movement of river courses that are still a prominent

feature of the area (Kalliola and Puhakka 1993).

Upland forests in Western Amazonia are areas located in the lowland plateau of

the lakebed that covered much of the Amazonian region in the Tertiary Period (Furch

1984). TTCR encompasses 322,500 hectares of mostly upland forest (71%) located on

the terrace forest that divides the valleys of the Amazon from the valley of the Yavari

rivers. Overall, the area corresponds to a type "a" (terrace) upland forest. In contrast to

the varzea forest, upland forest represents an older and more static ecosystem. Soil

formation has been dated to the Miocene epoch and shows deposits of elements of river

systems that prevailed ir past geologic periods. Thus, soils in upland forest are more

weathered and relatively poor in nutrients (Sioli 1984).














CHAPTER 2
ECOLOGICAL VARIABILITY BETWEEN AND WITHIN UPLAND AND VARZEA
FOREST


Introduction

Rich soils in varzea forest (Sioli 1984), have explained the greater biomass of

nonvovolant mammals (Emmons 1984) and primates (Peres 1997) than in upland forest.

Conversely, the lack of tolerance to riverine floods have also explained the lower plant

diversity in varzea than in upland forest (Gentry 1988). While upland and varzea forests

show variation in some ecological traits, upland and varzea forest have been assumed to

be ecologically similar within forest type (Bodmer et al. 1997, Coltrane 1998).

Animal biomass is largely dependent upon resources in the population's

environment (Harper and Begon 1990). Variation in fruit abundance apparently

influences frugivore abundance at the patch and local level (Levey 1988, Loiselle and

Blake 1991, Blake and Loiselle 1991) and ecosystem level (Whitmore and Prance 1987).

Fruit and invertebrates are the most important food resources for most Amazonian birds

(Terborgh 1985, 1986).

Several factors may influence food abundance. Studies have documented that

fruit abundance appears to be greater in lighted environments (Levey 1988, Loiselle and

Blake 1991). Moreover, in Neotropical forest light passage through the canopy

influences understory structure (Hubbel and Foster 1984). Edaphic characteristics

(Kalliola et al. 1993) and plant community (Encamrnacion 1993) may also influence the








variation in fruit abundance, patterns of fruit production, such as spatial distribution,

size of fruit patches (by virtue of differing number of plant species producing fruit), and

fruit size (Ayres 1993). Invertebrate biomass in the forest floor has been documented to

be positively correlated with measures of leaf litter accumulations (Vitt and Zani 1998,

Sweede 1998). However, the extent to which leaf litter accumulates may be influenced

by differences in water regimes (Nadkamrni 1994) between and within upland and varzea

forest. Examining the mechanisms that are likely to influence the occurrence and

abundance of animals is essential to interpreting life history traits, and, on the applied

level, to interpret the implications of hunting by humans on population of Amazonian

wildlife.

Most birds regularly hunted by rural Amazonians are frugivores (Redford et al.

1992). Up to 77.3% of species included in this study feed on fruit to varying degrees

(Delacour and Amadon 1973, Strahl and Grajal 1991, this study). In addition, 71% of

species that include fruit in their diets either forage exclusively (e.g., tinamous and

trumpeters) or partially on the ground (e.g., cracids).

In this study, I tested the assumption that upland and varzea forests differ between

forest type, but are ecologically similar within forest type. I tested this hypothesis in

terms of variation in vegetation structure and food abundance between and within forest

type. I first compared the canopy structure observed in selected sites within upland and

varzea forests, as canopy structure determines light passage, which influences the

structure (Hubbell and Foster 1986), and fruit production in the understory (Levey 1988,

Blake et al. 1990). Sites sampled within forest type correspond to the unhunted and

hunted sites from which most data for the chapters of this dissertation were obtained.








Then, I compared fruit abundance observed in sites within upland and varzea forests.

Due to sampling constraints, I restricted the analysis of fruit abundance to a within forest

comparison. Subsequently, I compared leaf litter accumulation observed in unhunted and

hunted sites in both forest types as an indirect measure of abundance of invertebrates on

the forest floor (Vitt 1998, Sweede 1998). I assumed the sites sampled had pristine plant

communities and that variation in canopy structure, understory density and food

abundance represented ecological differences that may influence the occurrence and

abundance of animals (Pastor et al. 1997).



Methods

Sampling was conducted in "Q. blanco" (=hunted sites) and "Yavari" unhunteded

sites) areas in upland forest, and "El Pinche" (=hunted sites) and "Sta. Elena" unhunteded

sites) areas in varzea forest (Chapter 1). Vegetation structure, leaf litter cover, and fruit

abundance were measured in 10 trails in upland forest and 12 trails in varzea forest

(Table 2-1). To measure vegetation structure and leaf litter cover, I established sampling

stations 500 m apart along freshly cut trails. To avoid bias from cutting, each sampling

station was located at 5 m off the trail on the right hand-side regardless of the direction

followed by the surveyor. If sampling on the right hand side was impossible due to

physical circumstances, I proceeded on the left of the trail. Long and freshly cut trails

were necessary because older trails tend to overestimate understory density and fruit

production due to greater light availability. Furthermore, short trails tend to overestimate

canopy fruit production due to the scale of the sample (See Blake et al. 1990).








Table 2-1. Sampling effort in sites in upland and varzea forest. Data on vegetation
structure, leaf litter cover, and fruit abundance were obtained along the trails established
in upland and varzea forests sites. Trails were surveyed only once.
Number of Sampling Readings
stations taken at all S schedule
Sampling sites trails per Sampling schedule
site a per site sampling
stations
Upland Yavari 5 25 100 June 9 June 30 1997
120
Upland Q. Blanco 5 30 120 February 12 23 1997

Varzea Sta. Elena 7 65 260 June 6 June 17 -1996

Varzea El Pinche 5 28 112 June29 July 14 1996
a Each trail measured 4 km on average.


Light passage through the forest canopy influences understory structure (Hubbel

and Foster 1986), and greater fruit production (Loiselle and Blake 1991, Levey 1988).

To test whether this process differs between and within upland and varzea forest, I

examined the structures of the forest canopy and the understory foliage in both forest

types. To assess the percentage of canopy cover, I used a spherical densiometer

(Lemmon 1956). I performed four readings rotating the densiometer in four opposite

directions at each sampling station. Readings from equi-spaced dots on a concave grid

correspond to the percentage of overhead area not occupied by canopy. The average of

the four readings was multiplied by 1.04 and subtracted from 100 to obtain an estimate of

canopy density in percent (Lemmon 1956).

To assess the understory foliage structure, I used a 2.13 m long pole graduated

along its length. This pole was positioned vertically through the understory profile, and

the number of foliage contacts within each 30.5 cm graduated section was recorded. This

was repeated 4 times in four opposite directions at each point and an average foliage








contact-30.5cm was obtained. Foliage densities, that includes all plant species, was

obtained for each of the 7, 30.5 cm height bands of the graduated pole.

To assess the leaf litter cover, I used a variation of the method used to measure

foliage density (Bibby et al. 1992). The method consisted of looking at the ground

through a 7.6 cm diameter and 20.3 cm long PVC tube. Then, I visually estimated the

percentage that was covered by leaf litter. I made 4 measurements at each point and

calculated the average leaf litter cover.

To quantify variation in fruit abundance within upland and varzea forests I used

absolute counts of fruits. Because I attempted to document ecological variability within

forest type, I also examined whether spatial distribution, path size (=aggregation of fruit

of the same type) and fruit size varied between sites within forest type. The spatial

distribution of fruit uses the time between encounters with a single fruit or fruit patch.

This analysis assumed that variation in time elapsed between encounters with fruit

indicated variation in spatial distribution of fruit.

To measure fruit abundance, I used a modification of an area-based survey (Blake

et al. 1990) used by Hilty (1980). I counted all ripe and unripe fruit on the forest floor

within 3 m on both sides of a trail, and those on shrubs and treelets within 5 m of height.

All fresh fruit was counted regardless of condition (fresh, rotten, damaged or partially

eaten). I did not distinguish between fruit produced by shrubs and treelets from those

produced by the canopy (Janson 1983).

I obtained the time (minutes) elapsed between encounters with fruit at a speed of

1 km/hour. The size of fruit patch was obtained by counting the number of fruit of the

same type at every encounter. When fruit of different types were found mixed in a single








aggregation, I recorded the number of each type of fruit and considered them as separate

patches. An index of fruit size was obtained by taking two cross section measurements

using a ruler. The approximation of fruit size uses the area (cm2) obtained by multiplying

the two cross-section measurements. To obtain the frequency distribution of fruit patch

size and fruit size in sites within forest type, I indexed measurements obtained for patch

and fruit size into size categories. The breath of categories were determined on the basis

of the range of measurements obtained for data on patch and fruit sizes.



Limitations of the Data Sets

There are several problems that must be considered in interpreting the data

obtained during this study. Perhaps the most serious problem was the protocol used to

obtain fruit data. In the varzea forest sites data on fruit was gathered in a period of 50

days allowing for comparison between sites. However, in the upland forest sites fruit

data was gathered during June (least fruit production in north eastern Peru [Garber 1993],

and February out of the period of least fruit production). This sampling problem likely

bias the results toward greater abundance of fruit in the sites sampled out of the period of

least fruit production. Characteristics of fruit (e.g., frequency distribution of fruit patch

size and fruit size) are likely to be influenced by variation in sampling chronology.

Variation in sampling protocol was largely due to logistic constraints.

A second problem was that fruit data from hunted sites may be influenced by

both, the rarity or absence of large frugivores (e.g., large primates, ungulates, and large

birds), and extractive activities by village residents. The unhunted sites are likely to have

pristine animal communities and were free from human extractive activities (e. g., fruit








extraction, building material and lumber).

The results of the analysis of leaf litter cover should be interpreted considering the

influence of seasonal floods on leaf litter cover. The analysis of leaf litter cover

represents a point in time and observed differences may vary in function of the frequency

and duration of seasonal floods in varzea forest, which are likely to influence leaf litter

cover.



Results

Canopy cover and understorv density. I examined the canopy structure as it

influenced light passage to the understory and, thus, influenced understory foliage density

and fruit production (Hubbel and Foster 1986, Levey 1988, Loiselle and Blake 1991).

Percentage of canopy cover did not differ within sites in each forest type, and was nearly

significant in favor of greater percentage of canopy cover in upland forest than in varzea

forest (P=0.1) (Table 2-2). In upland forest, readings of percentage of canopy cover

ranged within 10 points, while in varzea forest readings ranged within 21 points (each

point represents approximately one percent of canopy cover [Lemmon 1956]). Thus,

suggesting that the canopy in varzea forest is more variable in structure, and perhaps

allows a greater passage of light to the understory layers than in upland forest (Table 2-

3).

Foliage density, as measured by contact/30.5 cm, was significantly (P=0.00)

higher in varzea (0.40.46) than upland forest (0.280.3) (Table 2-4). Layers differed in

foliage density, but foliage density across the understory profile (=foliage density at each








Table 2-2. Summary statistics comparing canopy cover in upland and varzea forest.
Data are the arc-sine transformed readings of percentage of canopy cover.

Mean square df F P
Between forest 0.02 1 2.75 0.1

Within sites 0.02 1 2.09 0.15

Forest type*Site 0.01 1 1.69 0.2


Table 2-3. Average percentage of canopy cover and minimum and maximum values
obtained in sites within upland and varzea forests. Each measurement represents the
average of four measurements taken at each sampling station along trails. Each trail
measured 4 km on average and was sample one time.
Sampling sites Canopy cover (%) Range (%)

Upland Yavari 87.3 81 92

Upland Q. Blanco 87.4 83 92

Varzea Sta. Elena 84.59 68 -94

Varzea El Pinche 75.83 77 -93


Table 2-4. Summary statistics for the factorial ANOVA comparing understory profiles in
upland and varzea forest. Data are number of foliage contacts/foot on a graduated pole.
Mean
Square df F P
Square
Between upland and varzea 2.51 1 17.80 0.00
Among layers 2.62 6 18.60 0.00
Between sites (within forest type) 0.01 1 0.05 0.82
Forest type Layers 0.34 6 2.38 0.03
Forest type Sites 0.00 1 0.03 0.86
Layers* Sites 0.09 6 0.62 0.71
Forest Layers Sites 0.17 6 1.20 0.30








of the seven height bands) did not differ within forest type (P=0.82) (Figure 2-1). The

greater variability in canopy cover in varzea than in upland forest appears to support the

contention that light passage influences the foliage density in the understory and perhaps

influenced the greater understory foliage density observed in varzea forest.

Variation of fruit abundance within upland and varzea orest. I quantified fruit

abundance within sites in upland and varzea forest by counting fallen fruit on the forest

floor and fruit on shrubs and treelets. In varzea forest, fruit counted in "Sta. Elena" was

greater than in "El Pinche" (X2l 1.3, df=l, P=0.000). Likewise, in upland forest fruit

counted in "Q. Blanco" site was greater than in "Yavari" (x=7.7, df=-l, P=0.005) (Figure

2-2). These results suggest that fruit abundance showed spatial variation within upland

and varzea forest (Denslow et al. 1986). However, this results should be interpreted in

terms of sampling schedules in upland and varzea forest.

Time (minutes) elapsed between encounters with fruit (single fruit or patch) was

similar between "Yuracyacu" (3.13.7) and "El Pinche" (3.44.2) sites within varzea

forest (t=1.96, df=467, P=0.39), but differed in "Yavari" (44.3) and "Q. Blanco"

(2.52.5) in upland forest (t=1.97, df=-217, P=0.000). While time elapsed within sites in

varzea forest did not differ, variation in time elapsed between encounters in upland forest

might be an artifact of the sampling during different months in "Q. Blanco" (Table 2-5).

I also examined whether fruit patch size differed within forest type. Variation in

fruit patch size was assumed to reflect fruiting patterns within forest type. Small-sized

fruit patches were predominant in sites within both forest types, and their frequency

distribution did not differ between "Yavari" and "Q. Blanco" (Kolmogorov-Smimrnov

Z=0.64, P=0.8), and between "Yuracyacu" and "El Pinche" in varzea forest

















78-

i2 r, --/

-o
20
0

r~- 4-
"i i- ^ -i





w5 3 - |-----
Ui








> 2 -
0







-0.4 0.0 0.4 0.8 1.2

Average number of foliage contact/30.5 cm 1SD



Figure 2-1. Understory foliage profile for upland (broken line) and varzea forest
(solid line). The profile represents the average number of foliage
contacts/30.5 cm on a graduated pole along the vertical section of the forest
understory.




















500


400-


300-


200-


0 --


E Pinche Sta ena


SVazea forest
Upland best


Yavari Qda. Blanco


Figure 2-2. Variation of fruit abundance (Number of fruit counted/km). Bars
(with error bars) represent fruit abundance recorded in 9 trail in varzea forest
and 8 trails in upland forest. Trails averaged 4 km in length.















Table 2-5. Sampling effort and encounter rate of fruit in upland and varzea forest.
Distance T Mean encounter
.. Total fruit .. .a
Location surveyed counted rate (minutes)
(Km) counted

Upland-Yavari 04 31'01.3"S 73 24'44"W 26.8 4014 4.02 4.31*

Upland-Qda. Blanco 04 23'59.5"S 73 03'47.4"W 17 3439 2.45 2.54*

Varzea-Sta. Elena 05 14'04.5"S 75 49'30.9"W 35.1 10,073 3.1 3.71

Varzea-El Pinche 0451'13.3"S 7505'17.1"W 25.1 5,312 3.4 4.16
Significant at the 0.05 level.
a Individual fruit or fruit patch encountered at an approximate pace of 30'/km








(Kolmogorov-Smimov Z=0.63, P=0.82) (Figure 2-3).

I obtained fruit measurements to examine whether fruit size varied within upland

and varzea forest. Small-sized fruit was predominant in both forest types and frequency

distribution of fruit size categories was similar in the two sites within varzea forest

(Kolmogorov-Smimov Z= 0.42, P=0.99), the two sites within upland forest

(Kolmogorov-Smimov, Z= 0.49, P=0.97) (Figure 2-4).

Leaf litter cover. I examined the availability of invertebrates on the forest floor of

both forest types by measuring leaf litter cover. This analysis assumes that greater

concentrations of leaf litter reflect a greater invertebrate abundance on the forest floor

(Sweede 1998). Upland forest had a greater leaf litter cover than varzea forest (Table 2-

6). Leaf litter accumulation did not differ within forest type and ranges were similar

(Table 2-7). This suggests that invertebrate abundance on the ground of upland forest

might be greater than in varzea forest, but similar within forest type.



Table 2-6. Summary statistics comparing leaf litter cover. Data are the arc-sine
transformed reading of percentage of leaf litter cover.

Mean square df F P

Between Forest 1.24 1 8.21 0.000

Within Sites 0.00 1 0.003 0.96

Forest*Site 0.00 1 0.002 0.96








Table 2-7. Average percentage of leaf litter cover and minimum and maximum values
obtained in sites within upland and varzea forests. Each measurement represents the
average of four measurements taken at each sampling station along trails. Each trail
measured 4 km on average and was sample one time.
Sampling sites Litter Cover (%) Range (%)

Upland Yavari 85.6 35 -100

Upland Q. Blanco 86 36-100

Varzea- Sta. Elena 75.83 21-100

Varzea El Pinche 76.5 35 100





Discussion

Productivity of ecosystems dictates the potential to support animal biomass (May

1974). High productivity in the lowland Amazonian forest has often been used to explain

its high biological diversity (Whitmore and Prance 1987). Nevertheless, the lowland

Amazonian forest is composed of a myriad of ecosystems, some of which clearly differ in

biological diversity and animal biomass (Tergorgh 1992). Results of this study indicated

that while vegetation structure and leaf litter cover varied between upland and varzea

forest, fruit abundance also varied between sites sampled within forest type.

Varzea forest showed a more variable percentage of canopy cover. Wider range

in percentage in canopy cover in varzea forest coupled with greater understory density

suggests a greater passage of light than in upland forest. Klinge et al. (1990) suggested

that seasonal floods in varzea forest are responsible for the low plant diversity and less

complex canopy structure. Moreover, Encamacion (1993), in a more descriptive study

points out that the canopy of varzea forest is less dense and more heterogeneous than that





















'--" 9 O G Sn Z ;e Z a



Q. Blanco






oS . . .- .=


Sta. Elena


El Pinche


*^ S:~~~~~~~~~ ~ 8 =VK S K W U V ;7. AM ff A Z 4

Fruit-patch size categories

Figure 2-3. Frequency distribution of fruit patch/km. Data on fruit patch size was
collected from a total of 9 trails in Yavari and Q. Blanco (upland forest) sites and
a total of 8 trails in Yuracyacu and El Pinche sites (varzea forest). Trails measured
4 km on average.


-7 -7













105
90-
75
60
45 1
301
1.1


Yavari


- CC 0 N 0 CC 0 N 0 CC 0 N P CC 0 N 0 CC 0 N 0
4 4 .... .0 ~ AM AM AM .0.0.0 C CC * 4 4 4 0


mm' "d ' 4 A* .0 .4 .i . .


Sta. Elena


o A .4 AM a o AM .4 A a e .4 AM A N .4 ? .? 1


105 -
90-
75
60-
45
30


0 -


El Pinche


0 0f N 0 CC 0 fN 0 CC N> 0 Ci 0f Ni 0 CC 0 N> 0>
'6' 'O 4. -^ ^ -* .*- AM .0 A< M AM* AM* -t f .0 .0 .0 0 *0 4 .* .
2C CS MS CS tS CC zC CC CC C*


Fruit size categories (cm2)




Figure 2-4. Frequency distribution of fruit patch/km. Data on fruit size was
collected from a total of 9 trails in Yavari and Q. Blanco (upland forest) sites and
a total of 8 trails in Yuracyacu and El Pinche sites (varzea forest). Trails measured
4 km on average.


Q. Blanco








of upland forest. Therefore, this suggests that upland and varzea forest differ in terms of

canopy structure that may be partly responsible for greater foliage density in the

understory (Hubbel and Foster 1986).

Failure to detect clear differences in canopy cover may have been influenced by

sampling constraints in varzea forest. Due to access constraints, forest with permanent or

near permanent water (e.g., swampy forest), that is often found in varzea forest, was

circumvented. This type of forest has the lowest plant diversity and typically a broken

canopy (Klinge et al. 1990). Most sampling in varzea forest took place in areas free from

permanent water that supports plant communities that most resemble those in upland

forest (Encamacion 1993).

The spherical densiometer (Lemmon 1956) may not be an adequate devise to

compare canopy covers of high complexity. The quantification of the canopy cover relies

on openings in the canopy that are mirrored in a concave grid. A common problem was

that openings in the canopy were numerous but too small to meet the criteria needed to

define a point of a percent, hence, leading to subjective approximations (Lemmon 1956).

The lack of adequacy to assess cover in complex vegetation structures may derive from

the fact that the spherical densiometer was developed and tested in less complex

conditions of the tempered forest (for further discussion see Cook et al. 1995).

Fruit abundance varied between sites within forest type. Spatial variation in fruit

abundance observed can be interpreted by linking documented pattern of phenology of

fruit production in Neotropical forests and the sampling schedules used in this study.

Sampling in varzea forest took place within a period of 50 days. Given the distance

between sites sampled (43 kin) it is likely that environmental factor influencing fruit








production (e.g., rain seasonality) in both sites were similar. Studies have documented

relationships between variation of fruit abundance and habitat (Levey 1988, Loiselle and

Blake 1991). While varzea forest is considered distinctive Amazonian forest ecosystem

(Gentry 1988), there appear to be a myriad of plant communities within varzea forest

(Encamrnacion 1993) which could be influencing the spatial variation of fruit abundance.

Moreover, fruit production may vary among individual trees of the same species

(Wheelwright 1986), and in function to edaphic conditions (Kalliola et al 1993).

Fruit also varied in abundance between sites in upland forest. While factors

discussed above also apply to upland forest, sampling schedule in upland forest may have

influenced the results observed. The months of June-July are the period of least fruit

production in northeastern Peru (Norconk 1986, Castro 1991, Garber 1993), and

sampling in the "Yavari" site took place during June. While April-May constitute the

peak of fruit production in the region, sampling in February may have constituted and

artifact for the greater fruit abundance observed in "Q. Blanco" site.

Leaf litter cover was greater in upland forest than in varzea forest and suggests a

greater abundance of invertebrates on the forest floor (Vitt and Zani 1998, Sweede 1998).

Differences in leaf litter cover and in invertebrate biomass can be better understood by

considering differences in seasonal floods between upland and varzea forest. Seasonal

flooding of 2-4 months speeds the decomposition process (Vitousek 1985, Nadkarni

1994), washes away leaves, and likely kills most ground invertebrate fauna. Leaf

accumulation in the flood-free upland forest is likely to provide better conditions to build

up greater invertebrate fauna that favors at least the most terrestrial birds. Support for the

influence of availability of insects on occurrence and abundance of wildlife stems from








the near absence of understory insectivore primates in varzea forest and greater

abundance in upland forest (Peres 1997).

Environmental heterogeneity results in variation in resource abundance and

influences the habitat's potential to support wildlife (Pastor et al. 1997). A greater light

penetration through the canopy of varzea forest was linked to a greater understory density

(Hubbel and Foster 1986) than in upland forest. However, fruit abundance was not

uniform within upland and varzea forests. Differences in flood regimes between upland

and varzea forest may also influence leaf litter accumulation and perhaps invertebrate

abundance on the forest floor. Long periods of floods may result in a lower invertebrate

biomass in the forest floor because water speeds leaf litter decomposition (Vitousek

1985), and washes away leaf litter cover. In addition, 2-4 months of inundation is likely

to prevent formation of large invertebrate biomass on the forest floor.

Data indicates that structural characteristics of the canopy and understory as well

as leaf litter cover appear to differ between forest type. However, fruit abundance

differed between sites within forest type. These results support the assumption that

upland and varzea forests differ ecologically. While structural characteristics of

vegetation and leaf litter cover were similar between sites within forest type, fruit

abundance varied. Variation in fruit abundance may influence frugivore abundance

(Howe 1983, Levey 1988, Loiselle and Blake 1991) upland and varzea forests appear to

vary ecologically within ecosystem and therefore population abundance, at least of the

most frugivores species, may be influenced by ecological variation within each forest

ecosystem. In conclusion, upland and varzea forests appear to show ecological variation

within forest type, which may influence the occurrence and abundance of Amazonian





44

wildlife. This conclusion is considered in the study design and interpretation of results

of subsequent chapters in this dissertation.












CHAPTER 3
DETERMINANTS OF BIRD HARVEST IN THE PERUVIAN AMAZON


Introduction

Anthropologists and wildlife biologists have collected data on hunting by humans

to test a number of different hypotheses. Quantification of harvest of animals have been

used to test hypotheses of sustainability of hunting (Townsend 1995, Fitzgibbon et al.

1996, Alvard et al. 1997, Bodmer 1997, Coltrane 1998, Begazo and Bodmer et al. 1997),

hunting in habitats modified by humans (Jorgenson 1993), the efficiency of different

weapons (Hames 1989, Alvard 1995), optimal foraging by humans (Hawkes et al. 1997),

protein limitation (Gross 1975), cultural beliefs (Ross 1978), and socioeconomic

importance of wildlife (Gaviria 1981, Saravia 1992, Pierret and Dourojeanni 1967).

These studies generally reveal that hunters take a wide variety of animals at varying

frequencies. However, the cultural and biological factors that determine the fauna

available to hunters, and what ultimately they take have received little attention.

Wildlife available to hunters is the most obvious factor that may account for the

type and quantity of animals hunted (Redford and Robinson 1987). What hunters take is

also influenced by cultural factors, namely, hunter's preferences (Silva and Strahl 1991),

cultural beliefs (Ayres et al. 1991, Ross 1978), economic incentive (Dourojeanni 1985,

Bodmer 1995) and access to modem hunting technology (Vickers 1980, Vickers 1991,

Yost and Kelley 1983). Cultural factors coupled with ecological characteristics of the

habitat are likely to influence the availability of wildlife to hunters as well. For example,








settlement patterns, age of the settlement, human population density (Stearman and

Redford 1995) and the extent of habitat modification in the surroundings of the

settlements (Jorgenson 1993) influence the availability of wildlife and choices accessible

to hunters.

Addressing the hunting pressure that species endure, and the factors that influence

patterns of harvest are the first step in developing resource management in Amazonia

(Posey et al. 1984). In an attempt to explore cultural and environmental correlates of the

harvest of Amazonian birds, I gathered information on hunters' preference, hunting

patterns, and harvest of birds from residents at four human settlements located in upland

and varzea forests in North eastern Peru. I first determined the type and quantity of birds

harvested in the human settlements. Then I examined the harvest in terms of hunter's

preference, hunting patterns, and characteristics of the human settlements to ask whether

the type and quantity of birds taken by hunters is associated with cultural factors.

Subsequently, I examined the availability of other sources of animal protein (i.e.,

domestic animals and mammalian wildlife) in the settlements to explore whether hunting

pressure on birds is associated with variation in the availability of alternative sources of

animal protein.


Methods

Study Site

This study took place in the villages of San Pedro and Nueva Esperanza located in

the buffer zones of PSNR and TTCR (Chapter 1). Data on hunting pressure also were

obtained from the villages of San Felipe in Upland forest (W7157'35",S418'06" ) and

Dos de Mayo in varzea forest (W7549'30.9", S05 14'04.5"). Characteristic of the four








villages are provided in Table (3-1). Data were collected during May 1996 and

February 1998.

San Pedro (SP) SP is located along the Blanco River, a tributary of the Tahuayo

River, outside the area of influence of TTCR. The Blanco River is a typical upland forest

watercourse that has high water levels only during the wet season (January-May), and

maintains regular low water levels during the dry season (June-November). Residents of

SP maintain commercial trade with the city of Iquitos. There is a straight-line distance of

74.8 km between Iquitos and SP and it takes approximately 8 hours by regular fluvial

transportation to go between the two. SP can be considered a recent human settlement.

Until the year 1990 only 7 households and approximately 35 people inhabited the area (P.

Puertas personal communication). Since then, the human settlement has experienced an

annual growth rate of 29% and population size has increased by 228%. The main

activities are shifting cultivation, hunting, and fishing.

Nueva Esperanza (NE) NE is located on the periphery of PSNR, on the south

bank of the Marafion River. The Marafion River constitutes a major commercial

waterway connecting the Amazonian cities of Iquitos and Yurimaguas. The village is

surrounded by a network of small waterways that are used by village residents to access

the interior of the reserve. NE is an older village. Since about 1937, NE has been

occupied with at least three households, whose main activity has been shifting

cultivation, fishing, hunting and gathering (FPCN 1994). NE experienced an overall

annual growth rate of 2.5% between the years of 1972-1992 (FPCN 1994).









Table 3-1. Composition of households and human population of villages sampled during 1996 1998 in upland
and varzea forests. The harvest of birds is evaluated in terms of household sampled, hunters in households sample and consumers in
household sampled.


Habitat
transformation


Household
s in village


Households
sampled


Hunters in
households
sampled 8


Consumers
in
households
sampled b


Upland 115 Approx. 15
years



Varzea 380 Approx. 45
years


Households
spread in 28 km
of meandering
river.

Households
surround a
central place.


Approximately
300 m radius


Approximately
1000 m radius


San Felipe


Dos de Mayo


Upland 91


Varzea 155


Households
c surround a
central place.
Households
c surround a
central place.


Village


Forest
type


Human
population


Age of
settlement


Settlement
patterns


San Pedro



Nueva
Esperanza


' Adult males that hunted regularly.
b Hunters and non-hunters in the households. Includes (children and adult men and women).
c Data not available for the village.








San Felipe (SF). SF is located on the Yavari-Mirim River which dissects vast

areas of upland forest. Because of its remote location and difficult access due to low

water in the river during half of the year (June-November) SF maintains limited

commercial trade.

Dos de Mayo (DM) DM is located in a remote location on the Yuracyacu River

and also maintains limited commercial trade with the city of Yurimaguas. DM lays in

predominantly varzea forest within the limits of Pacaya-Samiria National reserve in the

state of Loreto.



Recording Techniques

To obtain information about the specifics of bird hunting, I used a methodology

that included participatory data collection, direct observation, and interviews. A diverse

approach was necessary to corroborate information from village residents who were

hesitant to provide information on their hunting activities. Although village residents in

all four communities had previously worked with wildlife researchers, reluctance to

cooperate was particularly noticeable in NE because most hunting took place within a

legally protected area, and village residents showed some concern about the implications

of this study.

The participatory method involved household members in the data collection. I

interviewed all households in the villages and provided a booklet to households that

voluntarily expressed interest in participating in this study. The booklets were provided

during June 1996 and were kept by households for one year until July 1997. Booklets

had the portraits and local names of the birds most frequently hunted (Appendix A).







Household members (eg., adult men and women in a house) were asked to record;

species identification; quantity; kill site; and date. The booklet-household method was

supplemented with interviews with household members to account for birds not recorded

on the booklets. When household members verbally reported a bird not registered in the

booklets, I asked the same questions as for items listed on the booklets. Data on harvest

yields were obtained from the participating households in each community (Table 3-1).

Direct observation and interviews also were used to gather data. I spent a total of

130 days in the village of NE and a total of 53 days in the village of SP. During these

periods, I observed and accompanied hunting, fishing, and wild fruit gathering trips. I

accompanied a total of 6 trips in NE and 4 trips in SP. In NE, 4 trips lasted less than a

day and two up to 9 days in the hunting grounds. In SP, I was able to accompany only

one trip of 6 days and the rest lasted less than a day. Participation on hunting trips were

intended to observe the process of bird hunting. All hunting trips were part of hunter's

regular activities, and were not undertaken under any request. Trips that I accompanied

were with hunters from households that participated on the data collection through the

booklet-household method. A hunting-gathering trip was defined as one intended to

procure land animals, fish or forest products. In addition, household members were

interviewed on the frequency, the duration, and species and numbers of animals taken

during hunting-gathering trips, as well as, on the type and quantity of domestic animals

kept by the household.








Composition of the Harvest

I obtained annual harvest of birds in each of the four communities by means of

hunting records (booklet-household method) kept by household members. Information in

booklets from the villages of NE and SP was transcribed to a permanent field data book

approximately every 2 months. Booklets were transcribed and returned during

approximately 1.5 hour-visit to participating households. A line separating transcribed

from new records was drawn for each species after transcription. When space available

for a species was filled out with records, I provided a new booklet. During the study, I

provided 4 extra booklets to hunters that filled out the space of at least one species.

Annual harvest from the remote villages was obtained from 7-month records kept by

households in DM (November 1996 May 1997), and 5 months in SF (February-June

1997).

Because the number of households sampled differed between SP and NE, I used

harvest rates for comparative analyses (No. birds/consumer-year). Consumers are all

household members over 5 years of age. I used consumer of 5 years and older to

standardize the harvest per number of people that consumed to total harvest. This

assumption implies that children less than 5 years of age did not consume the annual

harvest at the same rate of consumers of 5 years and older. Consumption per consumer

was used because birds are procured not only by adult male hunters, but also by other

members of the households. In addition, households often supported multiple families.

The annual avian weight harvested was obtained from the number of birds harvested

multiplied by the species' body weight.







Hunting Technique and Access to Ammunition

Hunting technique is defined as the mean (e.g., hunting weapon, traps, fishing

nets) used to procure birds. While transcribing data from the booklets to the permanent

field notebook, I asked household members about the technique used to procure each of

the birds. Direct observation was also used while accompanying hunting-gathering trips.

Additionally, I asked about the price and mode to acquire ammunition. Information on

the cost and access to ammunition was corroborated by obtaining the price and

availability of ammunition at retailers established at the villages and from merchants

traveling along rivers that stopped to commercialize goods at the villages.



Habitat Modification and Population Abundance

In this study, I examine two forms of habitat modification by humans; a) forest

transformation to agricultural fields and their subsequent fallows, and b) palm trees and

palm fruit extraction. Population abundance of some species increases in disturbed

habitats (Posey et al 1984; but see Jorgenson 1993). However, the reduction of food

resources that are important for a species is likely to influence their population abundance

(Pastor et al. 1997). Based upon this premises, I analyzed stomach content samples of the

species most frequently hunted, which included; the Cracidae (cracids), Tinamidae

(tinamids) and Psophiidae (trumpeters) to examine whether habitat modification

influences these species' diets, and possibly their population abundance and availability

to hunters. Household members collected the stomach contents of birds that they hunted.

They were previously trained on the preservation of stomach samples and supplied with

material for their preservation. Stomach samples were preserved in a mix of 3:1








water/formaldehyde until the day of analysis.

Seeds found in the stomachs contents were identified to the species level when

possible. Experienced botanists from a Peruvian University, "Universidad de la

Amazonia Peruana", aided with the seed identification. Using the frequency of plant

species found in the stomachs, I obtained the percentage of occurrence of plant species.

Additionally, I assessed the importance (volume) of three broadly defined food types; a)

leaves, b) fruit (pulp and seed) and c) animal matter (invertebrates). I poured each

stomach content into a 1.8-cm diameter dish, identified and sorted items, and visually

estimated the relative wet volume of each of the three food types in each stomach sample.

This analysis included taxa for which 10 or more stomachs were collected. Due to the

small sample size, stomach samples collected in upland and varzea forest were pooled.



Importance of Birds

This analysis compares the annual avian harvest with the annual mammalian

harvest recorded in NE (Bodmer et al. In Press) and SP (Puertas In preparation).

Riberefio people typically sell part of the large species of ungulates and large rodents to

the market and use small animals for household consumption (Padoch 1988). I examined

the numerical importance and contribution in weight of birds in relation to that of

mammalian wildlife harvested for household consumption.

Because economic incentive influences the type and quantity of wildlife hunted

by riberefios (Dourojeanni 1985), I also examined if avian wildlife represented a source

of income in the village and at the city market. The city of Iquitos' two most important

market places were visited every Saturday (the day of most commercial activity)








(Bendaydn 1990). To account for the periods in which the commercialization of

mammalian wildlife varied in magnitude (Bedaydn 1980), sampling was conducted

during the dry season (May through August 1997) and wet season (January through

March 1997). Each visit consisted of inspecting the two market places for birds being

offered for sale. Upon encounter I recorded species; sale price; number of individuals;

and the condition (dead, alive, fresh or smoked). Because the commercialization of avian

wildlife did not vary between dry and wet seasons, annual estimations were extrapolated

from data obtained over 7 months.



Preference of Hunters

I examined hunters' preferences because hunters' perception of prey influences

the type and quantity of animals taken (Ross 1987). Small mammals and birds are

commonly consumed at households. To examine the possible influence of the

availability and preference for alternative sources of animal protein on hunting pressure

on birds, I quantify hunters' preferences of birds and mammals commonly used for

household consumption. The rationale being that preference and availability of mammals

commonly used for household consumption may influence hunting pressure on birds.

Hunters' preference was determined through interviews using the method of triads

(Bernard 1994). This method also has been called the "lie detector method", and it is an

accurate and precise mode to determine an individual's position before an array of

choices. The method consists of presenting the interviewee with "n" items randomly

arranged in triplets that are shown one at a time. All possible triplets were randomized

and presented to each interviewee one triplet at a time. The interviewees ranked the three








species presented on each triplet, and were asked to justify their choice. The number of

triplets and the number of times that each item appears on the triplets are a function of the

number of items included in the test. The results are based on the frequency that each

item is ranked first by all interviewees. The preference for each species was obtained by

dividing the frequency that each item was ranked first by the total number of

interviewees.

Hunters' preference was determined in terms of characteristics (e.g., taboos on the

consumption of certain species and palatability of the meat) of animals. Hunters prefer

larger than smaller animals (Alvard et al. 1996). To remove the effect of size on hunter's

choice, items were included in two separate tests on the basis of their similarity in body

size. One test included 16 items; 8 mammals (average weight: 41.7 kg) and 8 birds

(average weight: 20.8 kg), in 17 triads with each item appearing 3 times (see Bernard

1994) (Appendix B). The second test included smaller birds and mammals of similar size

and included 14 items; 6 mammals (average weight: 0.90.3 kg), and 8 birds (average

weight: 0.70.3 kg) (Appendix C).



Limitations of the Data Sets

There were several factors that must be considered when reviewing these results.

One, these results were based on a small number of human settlements. As a

consequence result could be biased toward specific characteristic of the human

settlements studied and may not necessarily represent general patterns at the regional

level. Two, the method of triads used to quantify preference tends to overestimate

preference for species that interviewees are not very familiar with (e.g., rare species in








their hunting grounds) (Ross 1978). These factors may bias the results on hunters'

preference because some species were not equally abundant in the hunting grounds of

upland and varzea forest. Three, the quantification of annual harvest was based on

voluntary participation of households. Because hunting is not equally important for all

households within a human settlement, participating households may represent

households for which hunting was an important activity. This sampling bias may lead to

over estimation of the annual harvest of birds in the human settlements studied. Another

factor the may lead to over representation of the annual harvest was that some household

members appeared to be under the impression that I expected them to have the booklet

with records at my bimonthly visits. This was noticeable in two households in NE that

showed inflated numbers of curassows hunted in a period of two months. These records

were not included in the study and I emphasized during my visits that the booklets were

intended to keep records of birds that they would normally hunt. Four, annual harvests at

the villages of SF and DM were extrapolated from 5 and 7 month records respectively.

While annual harvest appeared to be similar throughout the year in the villages with

complete annual records, there is the possibility that the harvest in the periods recorded

vary from the harvest during the periods unrecorded in the two villages.



Results

Bird harvest. I recorded the avian harvests in four riberefio communities to

determine preference, characteristics of the human settlements, hunting techniques,

habitat modification, and access to ammunition. In the village ofNE, a total of 955 birds

were reported taken by 33 households or 165 consumers in a year period. This made up







an annual per consumer harvest of 6 birds or 6.6 kg-year. In the village of SP, a total of

669 birds were reported taken by 17 households or 85 consumers in a year period. This

made up an annual per consumer harvest of 8 birds or 9.1 kg-year. Consumers in SP

harvested a greater number of birds than in NE (XI=6.1, df=l, P=0.013). Households in

upland forest took a greater number of species (29) than those in varzea forest (24),

(X2=7.9, df=-l, P=0.004) (Table 3-2).

Village residents at remote locations harvested a smaller number of species at

lower per consumer rates than those at SP and NE (Table 3-2). In the village of DM, 142

birds were reported taken by 26 households or 125 consumers in a year period. This

made up an annual per consumer harvest of 1.1 birds or 1.21 kg. In the village of SF, 112

birds were reported taken by 9 households or 35 consumers in a year period, which made

up an annual per consumer harvest of 3.2 birds or 3.5 kg.

Preference. I examined hunters' preference to see how it influenced the avian

harvest. In NE, 544 triads were presented to 33 hunters. Likewise, in SP 289 triads were

presented to 17 hunters. In SP, hunters preferred mammals over birds (Kendall

coefficient of concordance (KCC) W=0.25, P=0.046, n=16), whereas in NE, preferences

were similar (KCC, W= 0.14, P=0.13, n=16). Among wildlife of small size category,

mammals were also preferred over birds in SP (KCC, W=0.37, P=0.33, n=14), and

preferences were similar in NE forest (KCC, W=0.14, P=0.16, n=14). Hunters in NE

regarded mammals and birds equally, and those in SP preferred mammals over birds.

Silva and Strahl (1991), and Delacour and Amadon (1973) suggested that birds are

heavily hunted because their meat is preferred over that of mammals. However, I did not

find evidence to support this hypothesis among riberefto people in the Peruvian Amazon.








Preference of wildlife among riberefio people varied among individual

households. Riberefio people represent a diverse ethnic group that is generally composed

of village residents that have immigrated from other region within the Peruvian Amazon.

This mixture of ethnic diversity and variation in costumes and was reflected in

household' preference and the harvest of wildlife (Table 3-2). For example households

that immigrated from indigenous groups showed a less specific preferences over species

of wildlife. On the other hand, households of traditional riberefios showed a more

selective preference over species of wildlife.

The distinction between household of indigenous descent and those of more

traditional riberefio culture was also noticeable in terms of beliefs and taboos regarding

the consumption of meat of certain species (Ross 1978). Households of indigenous

descent considered beliefs and taboos in their hunting preferences, whereas modem

riberefios did not consider beliefs and taboos, but the quality of the meat was more

important in determining their preferences.

While hunters did not show differences in preferences between mammals and

birds the bird meat was generally considered one of better flavor and with medicinal

properties. Hunters indicated that meat from birds was often consumed in circumstances

when hunters wished to eat meat of good flavor. Bird meat was also provided to people

in the process of recovery from sickness or as specific medication for certain illness. For

example, cracid meat wc.s highly regarded as meat with medicinal purposes. Cracids

were often requested by village residents as a form of medication for simple diseases or

as a special nutritious food for women recovering from labor activities.






Table 3-2. Comparison of individuals hunted per species, percentage of total harvest, and per consumer harvest rate between the four
villages studied.

San Pedro Nueva Esperanza San Felipe Dos de Mayo
Percent Percent PercentPecn
Number Percent Harvest Number Percent Harvest Numberercent Harvest Number Percent Harvest
Species NumberHtotal rveatNumb of total a of total r oftotal t
Species of bird of total rate a of bird rate of bird havs rate of bird v rate
__harvest harvest harvest harvest


Tinamus gutattus

Tinamus major

Crypturellus cinereus

Crypturellus undulatus

Crypturellus variegatus

Crypturellus barttleti

Mitu tuberosa

Crax globulosa

Penelope jacquacu

Pipile pipile

Ortalis gutatta

Odontophorus stellatus

Cairina moschata


28 4.2

110 16.4


0.6

0.6

3.1



26.5

4.6

3.6

0.3

0.7


0.33

1.29




0.05

0.05

0.25



2.08

0.36

0.28

0.02

0.06


12.34

3.56

15.06




7.43

0.21

5.44

15.06

3.24


22 2.30 0.1


10.7 0.343


4 3.6 0.114


24



22

4

16



2


21.4



19.6

3.6

14.3



1.8


0.686



0.629

0.114

0.457



0.057


2.4

2.4

14.5


0.03

0.03

0.16


9.7 0.11


4.8

30.2




1.2


0.05

0.34




0.01


I I L






Table 3-2. Continued
San Pedro Nueva Esperanza San Felipe Dos de Mayo
Percent Percent Harvest Num erc Hanes
Number percent Harvest Number percent Harvest Number Percent Harvest Number percent Harvest
Species of bird oftota rate of bird of total rate of bird of total rate of total rate a
Paarcrxharvest harvest rate of bird harvest rate of bird harvest
Phalacrocorax
olivaceous 12 1.26 0.1 -
Ramphastos spp 33 4.9 0.39 25 2.62 0.2 2 1.8 0.057 -
Amazonaspp 27 4.0 0.32 31 3.24 0.2 -
Araspp 28 4.2 0.33 37 3.87 0.2 18 16.1 0.514 3 2.4 0.03
Leptotila rufaxila 12 1.8 0.14 39 4.08 0.2 39 27.8 0.32
Psophialeucoptera 81 12.1 0.95 13 1.36 0.1 6 5.4 0.171 2 1.2 0.01
Anhinga anhinga 2 0.3 0.02 39 4.08 0.2 2 1.8 0.057 -
Ardea cocoi 24 3.6 0.28 21 2.20 0.1 -
Cochlerius cochlerius 4 0.6 0.05 72 7.53 0.4 -
Aramanilata 10 1.5 0.12 37 3.87 0.2 -
Aratinga
leucophthalma 7 1.0 0.08 12 1.26 0.1 5 3.6 0.04

Helornisfulica 15 2.2 0.18 -
Pionus menstrus 17 2.5 0.20 -
Morphus gujanensis 1 0.1 0.01 -..

Micrastur ruficollis 2 0.3 0.02 -
Total number of birds harvested divided by number of consumers at each village (Table 3-1).








All village residents from NE and SP justified their choices of wildlife on the

basis of meat quality (smell and toughness) and the size of the animals (Figure 3-1).

While animals within the large wildlife category differed in size (1.2 kg to 3.3kg),

preference was independent of size in SP (r=0.2, P=0.4) or in NE (r=-0.004, P=0.9). It is

widely agreed that size is the most important factor determining hunter preference

(Redford et al. 1992, Alvard et al. 1997, Hawkes et al. 1997), but riberefio people

appeared to compromise size for quality of meat, at least among the range of animals

used in this test (Figure 3-1, 3-2).

Cracids, tinamous and trumpeters were consistently preferred in NE and SP, while

herons, anhingas and cormorants, were less preferred. Preference for some species

showed discrepancies between NE and SP. M tuberosa was ranked among the top 6

species in NE but ranked lower in SP. Likewise C. moschata and P. ]acquacu were

ranked among the top 6 species in SP, but they were ranked lower in NE.

Finally, village residents in both forest types did not associate any spiritual or

physiological prohibition on the hunt and consumption of any bird. They associated

Night monkeys Aotus nancymae with reincarnated spirits. However, they hunted and

consumed its meat, suggesting that traditional beliefs have little influence over the

harvest of wildlife among ribereno people.

Overall, hunters appeared to realize their preferences. Cracids, tinamids,

psittasids (parrots and macaws), and trumpeters (Psophiidae) were harvested at greater

rates than species that did not meet characteristics of being large and having good meat

(Figure 3-3). These results suggest that hunter preference influenced the type and

quantity of birds taken and this is consistent in both SP and NE communities.
















5% CGt


4% Md J

4% Pp
3% Wc-
A 3% Ms-'
2% Ah- 1
2%Wn-1
1% AnJ
1% CoJ






7% Rc

4% Gtt

4% Ar
4% Cm
3% Wn.

B 3% wcJ


9% Hm

11% Rc

11% Sg
66% 13% Bc

13% Ag

13% Ar











9% Pp
9% Ag
10% Hm

74% 11% Bc

12% sg

19% Md


Figure 3-1. Wilflife preference obtained from village residents at A:San Pedro and
B:Nueva Esperanza. Species belong to the large-sized category. Preference is given
as the frequency (in percent) that each item was ranked first among the choices
presented in the triads. Hunters' in upland and varzea forest show consistency in
preference for mammalian species over avian species and cracids, and tinamous among
avian species. Mammalian species are Wc=Cebus albifrons, Bc=Cebus
apella, Hm=Alluata seniculus, Ms=Pithecia monachus, Ag=Dasyprocta fuliginosa,
Ar=Dasypus novemcinctus, Cm=Nasua nasua, An=Tamandua tetradactyla.
Bird species are, Gt=Tinamus major, Rc=Mitu tuberosa, Sg=Penelope jacquacu,
Pp=Pipile cumanensis, Md=Cairina moschata, Co=Phalacrocorax olivaceus,
An=Anhinga anhinga, Wn=Ardea cocoi.















18% CU
6% Ss
-1I% Cp



5% Og 7 7e 0 13% Ar

A 4% O 13% Ps
4% CC-"
15% Mp

1% Oh!

0% SfLI




__8% CU
10% On
11% Ps
5% CC1
4% Rs92
B 2% Ar 18% Og

1% As-/
1% Oh!! 28% Ss
0% Sf



Figure 3-2. Wilflife preference obtained from village residents at A:San Pedro and
B:Nueva Esperanza. Species belong to the small-sized category. Preference is
given as the frequency (in percent) that each item was ranked first among the
choices presented in the triads. Hunters in both forest types consistently preferred
trumpeter over other avian species. Species are small birds and mammals where,
Sf=Saguinus fuscicollis, On=Aotus nancymae, Sc=Saimiri sp., Cp=Callicebus
cupreus, Mp=Myioprocta pratti, Ss=Sciurus spp. Cu=Crypturellus undulatus,
Cc=Crypturellus cinereus, Og=Ortalis guttata, Oh=Opistocomus hoatzin,
Rs=Ramphastos spp., As=Amazona spp., Ar=Ara spp., Ps=Psophia leucoptera.


















6% Ra


5% Ar
A 2%Co
1% An-/

0% Ah

0% Ph-


4% Co Z

4% Ah-
B 3%Ra-

2% An-
1% Pp-p/

1% Ph-


84%


.7% Ps
14% Pp

22% Ti




43% Cr


-10% Ar

11% Ps


32% Ti



32% Cr


Figure 3-3. Percentage of avian meat (kg) harvested per household-year in
terms of taxonomic group in A: upland forest, and B: varzea forest; where,
Ti= Tinamidae, Cr-Cracidae, An=Anatidae, Ph=Phalacrocoridae,
Ra=Ramphastidae, Ps=Psittacidae, Co=Columbidae, Ps=Psophiidae,
Ah=Anhingidae, Ar=Ardeidae.








Hunting patterns. I examined hunting patterns of riberefio people, and

characteristics of the human settlements in an attempt to find associations between the

bird fauna available to hunters and what hunters took. Hunting birds was opportunistic

and not mutually exclusive from other activities (e.g., fishing, gathering). Riberefios

typically carried a shotgun on any outing, regardless of the purpose. Hence, birds were

hunted upon casual encounters in the course of any activity and were regarded as the

meat for the day's meal.

Birds were hunted in both short and long trips. A short trip consisted of searching

for wildlife, fish, and forest products from the vicinity of the household to distances of

approximately 5 km, and hunters returned to their household the same day. In SP short

trips were undertaken by at least one member of the household (n=l17 households), 1-5

times a week, with a mode of 2 and a weekly average of 3.05 times or 156.4 per year. In

NE short trips were undertaken by at least one member of the household (n=33

households), 1-5 times a week, with a mode of 3, and a weekly average of 2.52 or 159

times per year. Frequency of short trips undertaken by hunters was similar in both

villages (t=2.02, df=40, ?=0.13). In SP short trips lasted 3.51.8 hours while in NE

4.32.6 hours (t=2, df=48, P=0.21). In both villages, short trips ranged wide in duration

from short forays of 0.5 to 11 hours.

Long trips were undertaken less frequently. Long trips in NE took 2-15 days

away from the household, with an average of 5 days. Long trips in SP took 3-11 days,

with a longer average (7 days) than NE (t=1.7, df=46, P=0.03). Long trips in SP covered

distances of up to 33 km with a mode of approximately 25 km, whereas in NE, long trips

covered distances of up to 29 km with a mode of approximately 18 km. Hunters in NE








undertook a greater number of long trips per year (9.36.1) than hunters in SP (4.61.6)

(t=2, df=33, P=0.000). Long trips in NE had fishing and hunting as the main objective

whereas in SP hunting was the main objective.

Areas near the settlements were subject of a greater hunting pressure than areas

away from the settlements. Residents in SP took more birds from areas near the

settlement (short trips) (505) than from areas away from the settlements (long trips)

(132), (x2=-16, df=l, P=0.000). Residents in NE took a similar number of birds from

short trips (521) and from long trips (434) (x2O--0.51, df=l, P=0.47). Comparing short

trips only, residents in SP took a greater number of birds than residents in NE (X2=4.4,

df=- 1, P=0.04), but both took a similar number of birds in long trips (X2= 3, df= 1, P=0.1);

(Table 3-3). Small and less preferred species predominated the take from short trips in

both villages, and help explain their presence in harvest yields in spite of their low

preference. In contrast, birds taken in long trips were predominantly large and preferred

species, but account for a lower proportion of the total harvest (Figure 3-4).

The bird fauna surrounding NE differed in composition from that surrounding SP

(Chapter 4). Variation in the composition of the bird community appears to be associated

to differences in settlement patterns, a greater human population density and older age of

NE than SP (Stearman and Redford 1993). The bird fauna available to hunters clearly

influenced the take of birds from short trips in NE where hunters did not report the large

cracids M tuberosa, P. jacquacu, the trumpeter Sophia leucoptera, and only small

numbers of P. cumanensis. The bird fauna in the surroundings of SP showed a more

complete bird community (Chapter 4) and hunters reported taking all species from short

hunting trips. This appears to suggest that characteristics of the human settlements,








Table 3-3. Number and percentage of birds hunted in short trips and long trips by
hunters in San Pedro (upland forest) and in Nueva Esperanza (Varzea forest) during
February 1997 through February 1998._________
Upland Forest Varzea Forest
Short trips Long trips Short trips Long trips
V Number Percent x Percent
cisNumber Percent Number Percent Number Percent Number Percent
Species of birds of total of birds of the br otal ofbirds the
ttlof birds of total of birds tofthl
total total
Tinamus gutattus 21 75 7 25 0 0 0 0
Tinamus major 94 85 16 15 74 63 44 37
Crypturellus cinereus 0 0 0 0 34 100 0 0
Crypturellus undulatus 0 0 0 0 138 96 6 4
Crypturellus variegatus 4 100 0 0 0 0 0 0
Crypturellus barttleti 4 100 0 0 0 0 0 0
Mitutuberosa 6 29 15 71 0 0 71 100
Crax globulosa 0 0 0 0 0 0 2 100
Penelopejacquacu 131 74 46 26 0 0 52 100
Pipile cumanensis 9 29 22 71 9 6 135 94
Ortalis gutatta 23 96 1 4 27 87 4 13
Odontophorus stellatus 2 100 0 0 0 0 0 0
Cairina moschata 2 40 3 60 5 23 17 77
Phalacrocorax 0 0 0 0 6 55 5 45
olivaceous
Ramphastos spp 33 100 0 0 21 84 4 16
Amazonaspp 27 100 0 0 30 97 1 3
Araspp 28 100 0 0 24 65 13 35
Leptotila rufaxila 12 100 0 0 39 100 0 0
Psophia leucoptera 54 67 27 33 0 0 13 100
Anhinga anhinga 2 100 0 0 22 56 17 44
Ardeacocoi 19 79 5 21 16 76 5 24
Cochlerius cochlerius 4 100 0 0 25 35 47 65
Aramanilata 10 100 0 0 37 100 0 0
Aratinga leucophthalma 7 100 0 0 12 100 0 0
Helornisfulica 15 100 0 0 0 0 0 0
Pionus menstrus 17 100 0 0 0 0 0 0
Morphus gujanensis 1 100 0 0 0 0 0 0
Micrastur ruficollis 2 100 0 0 0 0 0 0





























1000-1499 >1500


0
0-499


500-999


Figure 3-4. Size categories of birds taken in A short trips and B long trips. Small birds
were taken at greater rates in short trips than in long trips. The cracid Penelope jacquacu
still occur in the areas near SP and is hunted at high rates in SP and explains the high
value in the 1000-1499 gr size category.


S Long trips SP
-- Long trips NE


1000-1499


>1500








coupled with village residents' hunting pattern, may be associated to variation in the

bird fauna available to hunters, which in turn appears to be associated to what hunters

took.

The fauna available to hunters appeared to be influenced also by natural

differences in abundance between upland and varzea forest. This was particularly

evident in the hunting returns from long trips. P. jacquacu and S. leucoptera were more

abundant in upland forest (Chapter 4) and hunters in SP took a greater number of

individuals than in NE (x-26.6, df=-l, P=0.000), (x-=36.1, df=-l, P=0.000). Likewise, P.

cumanensis was more abundant in varzea forest and it was taken in greater numbers than

in SP (X2=6, df=l, P=.0-1).

Access to ammunition and hunting technique. Access and cost of ammunition

and the use of fishing nets influenced what hunters took. Hunter procured most birds

with 16 gauge shotguns, and the harvest appeared to be influenced by the cost and access

to ammunition. The price of a cartridge was greater (by 64%) in the remote villages of

SF and DM (range $ 1.6-1.9) than in NE and SP (range $ 0.95-1.19). Households at SF

and DM took half the number of species at lower average rates than those in NE and SP

(df=29, t=1.75, P=0.04). Hunters unanimously declared that birds were hunted at low

rates due to the high cost and difficult access to ammunition. These results concur with

the notion that hunters in villages with easy access to hunting equipment are less

concerned about the cost of ammunition and more likely to shoot a wider type and size of

prey at greater rates (Vickers 1980).

The prevalence of fishing among residents in NE combined with high abundance

of piscivorous birds resulted in a greater harvest of these species in NE than in SP. Only








2 anhingas and no cormorants were taken in SP. Although these species were not

preferred 95% of anhingas taken in a year period (n=39 birds), and 40% of cormorants

(n=12 birds) were accidentally trapped in fishing nets.

Baited cages, snares and fall traps were used in NE and SP to trap generally small

birds (< 400 gr). Species captured with traps were the same in NE and SP (doves, wood-

quails and small tinamous) and accounted for (7%) in NE and (5%) in SP of the total

annual harvest.

Habitat modification. In this study, I examined two forms of habitat modification

by humans; a) forest transformation to agricultural fields and their subsequent fallow, and

b) palm trees and palm fruit extraction. Based on the premise that population abundance

of some species increases in disturbed habitats (Posey et al 1984) and the reduction of

food resources that are important for a species are likely to influence their population

abundance (Pastor et al. 1997), I investigated the implications of habitat modification on

the food base of Amazonian birds hunted by humans. Stomach samples of the Tinamus

spp., Crypturellus spp., 0. guttata contained mostly fruit, including plant species that

typically grow in disturbed forest, human farming areas, and crops planted by humans

(Figure 3-5). Stomach samples of P. jacquacu, and P. cumanensis showed mostly fruit of

palms. The diet of M tuberosa, and P. leucoptera consisted of fruit and insects, but did

not show evidence of items that would be linked to habitat modification.

The frequency at which palm fruits occurred in the stomach samples of P.

cumanensis and P. jacquacu was similar (XO-0.05, df=l, P=0.8). The occurrence of palm




















100 I


~' 80



S60
E


~.40-


20



20
0 x .. .x



em Fit- Isct -- 0eve
ID 35 Co
E I. E P l
60 X a- X











Fruit insects Leaves




Figure 3-5. Proportion (percent volume) of food items identified from stomach
contents of birds taken by riberefio people in upland and varzea forest. Stomach
samples from upland and varzea forest were pooled to increase sample sizes as
follows; Tinamus spp= 29, Crypxurellus sp=3 1, P. jacquacu=36,
P. cumanensis=-20, 0. guttata=- 10, P. leucoptera-- 19, and M tuberosa-- 10.








species varied in the samples between species (Figure 3-6). However, all 6 palm species

were heavily extracted for building material, food, and as a source of income (Vazquez

and Gentry 1989).

Palm extraction is more prevalent near the settlements (personal observation), and

it is likely that the older and more densely populated NE exerted a greater impact on palm

trees than SP. In NE P. jacquacu has been locally extirpated while P. cumanensis is rare

in the grounds used for short trips (Chapter 4). Conversely, the two species were

regularly reported taken from the hunting grounds used for short trips in SP. While still

exploratory, it is likely that differences in the availability of palm trees and palm fruit

between the two human settlements influence the occurrence and abundance of species

that depend on palm fruit in areas surrounding the human settlements.

In contrast, stomach samples of the Tinamidae revealed that their diet included

species varied in the samples between species (Figure 3-6). However, all 6 palm species

were heavily extracted for building material, food, and as a source of income (Vazquez

and Gentry 1989).

Palm extraction is more prevalent near the settlements (personal observation), and

it is likely that the older and more densely populated NE exerted a greater impact on palm

trees than SP. In NE P. jacquacu has been locally extirpated while P. cumanensis is rare

in the grounds used for short trips (Chapter 4). Conversely, the two species were

regularly reported taken from the hunting grounds used for short trips in SP. While still

exploratory, it is likely that differences in the availability of palm trees and palm fruit

between the two human settlements influence the occurrence and abundance of species

that depend on palm fruit in areas surrounding the human settlements.







73










80 -

E" 70
U


70 Penelope jacquacu
Q Pipile cumanensis
60
6
0
50

40

r. 30 1
0
20

10) 10
0

0 1_
Euterpe Iriartea/Socrat Jessenia Mauritia Oenocarpus







Figure 3-6. Occurrence of palm fruit in stomach samples of P. jacquacu and P.
cumanensis. The percentage of occurrence of palm types varied between the two
bird species but palms as a whole were found at similar frequencies in both
species.








In contrast, stomach samples of the Tinamidae revealed that their diet included

fruits and seeds of plant species that grow in secondary forest, as well as crops planted by

human (Leopold 1959, Jorgenson 1993). Twenty nine percent (n=58) of stomach

samples of Crypturellus spp. and Tinamus spp. combined, contained seeds of plants that

are often found in secondary forest. Furthermore, 16% of the stomachs contained crops

grown by people at the villages. Agricultural practices appear to provide supplemental

food for these species of the tinamids. Greater availability of food often results in greater

abundance of animals (Howe 1983, Loiselle and Blake 1993). Greater availability of

food for Crypturellus spp. and Tinamus spp. in surrounding areas of the human

settlements is likely to result in the greater population abundances observed (this study).

Greater abundance of these species observed also reflects a greater representation in the

harvest yields obtained by hunters.

Hunting pressure on birds. I examined the influence of alternative sources of

animal protein on the hunting pressure of birds by measuring the availability of

mammalian wildlife and domestic animals in SP and NE. I also tested the hypothesis that

hunting pressure of birds is influenced by economic incentives by measuring the

monetary value of birds.

Availability of mammalian wildlife. I examined the harvest of birds and

mammals used for household consumption to see if differences in availability of

mammalian wildlife influenced the hunting pressure on birds. In this analysis I assumed

that SP and NE had equal access to ammunition, hunters did not prefer birds over

mammals and that the harvest of small mammals represented availability. The later

assumption is based on the hunting of animals for household consumption (e.g., birds and








small mammals) being largely opportunistic, reflecting the frequency of encounters

between hunters and prey (Hawkes et al. 1987, this study). In SP, village residents

harvested a greater number of birds than mammals used for household consumption

(X=133.6, df=-l, P=0.2). But avian weight in kg harvested was similar to the weight

harvested for mammals (x2O.2, df=-l, P=0.9). Likewise, in NE, hunters took a greater

number of birds than mammals (X2=24.1, df--=1, P=0.000) and avian weight harvested was

similar to the weight harvested from mammals (X2=0.14, df=-1, P=0.71) (Figure 3-7).

Abundance of animals in the hunting grounds often reflects the harvest of wildlife among

rural Amazonians (Bodmer 1996). Birds were hunted at similar rates as small mammals,

suggesting that there is not variation in hunting pressure over one of the two groups, and

this was consistent in both forest types.

Availability of domestic animals. People in SP harvested birds at greater rates

than those in NE. I examined possible influence of meat from domestic animals on the

hunting pressure on birds by measuring the prevalence of domestic animal in both

villages. Animal husbandry was more prevalent in NE than in SP. In NE most

households (92%) owned chickens, 38.9% ducks and 16% pigs. In SP 88.7% owned

chickens, 2% ducks and 14% pigs. Domestic animals, especially chickens, were

considered assets, and were frequently used as a source of income, as well as for

household consumption (Gonzales 1988). In both NE and SP the proportion of

households that owned chickens was similar. However, households in NE kept a greater

number of chickens (16.412.7) than household in SP (96.1); (t=2.63, df=-39, P=0.006).

A greater potential for agriculture in the rich soils of varzea forest (Dourojeanni 1990)

possibly influenced feed production and the potential












30 -


25


20


15


10_


5


0






50


40


30


20


10


0


Upland


Varzea


Upland Varzea


N Mammals


- Birds


Figure 3-7. Comparison of A weight (kg) and B number of individuals between
mammals used for household consumption and birds. Mammalian harvest was obtained
from Bodmer et al (In Press) and P. Puertas (In preparation).








for domestic animal husbandry. This help explain the greater prevalence of domestic

animals in NE than in SP. These results suggest that meat from domestic animals was

more available in NE than in SP and may partly explain differences in hunting pressure

on birds (Ayres et al. 1991).

Monetary value. I examined the magnitude of commercial trade of birds to see if

monetary incentive was associated with hunting pressure on Amazonian birds. The city

market study suggested that commercial use of species regularly hunted in rural areas

was negligible (Table 3-4). A total of 21 species and 2791 individual birds was recorded

in the two most important market places in Iquitos during a year period. Nearly half of

the species commercialized in the Iquitos market (48%) were regularly used for

consumption in SP and NE. However, species that were regularly hunted in the villages

studied accounted for orly 24 individuals or (0.9%) of the birds annually sold in the

Iquitos market. Furthermore, most species consumed in the rural sector were sold as

pets, and only 2 carcasses were offered for consumption.



Discussion

Studies on wildlife harvest reveal the wide range of species taken by Neotropical

human hunters (Redford and Robinson 1987). Hunters usually prefer large-bodied

animals to small-bodied animals (Redford 1992). This hunting preference makes

economic sense since large-bodied animals provide more meat per cartridge. However,

birds are an item used for household consumption often needed to obtain the meat for the

day's meal, and cost-benefit relationships are not the most important reasons explaining





78


Table 3-4. Number of individual birds sold in the Iquitos market in a year period. Prices
are in $ US.
c Condition Number Price Total
Species of sale a of birds unit value


Tinamus major *

Crypturellus undulatus *
Mitu tuberosa *

Penelope jacquacu *

Amazona amazonica *

Amazona ochrocephala *

Ara ararauna *

Ramphastos sp *

Psophia leucoptera *

Ara several *
Aratinga leucophthalmus

Brotogeris versicolorus

Pionus menstrus
Veniliornis sp

Ardea cocoi *
Bubulcus ibis

Cacicus cela

Tyrannus tyrannus

Tyrannus savana
Agelaius icterocephalus

Gymnomistax mexicanus


Great tinamou

Undualted tinamou

Razor-billed Curassow

Penelope jacquacu

Orange-winged Amazon
Yellow-crowned
Amazon
Blue and Yellow
Macaw
Toucans
Pale-winged Trumpeter

Red-bellied Macaw

White-eyed Parakeet

Canary-winged Parakeet

Blue-headed Parrot
Woodpeckers

White-necked Heron
Cattle Egret

Yellow-rumped Cacique

Eastern Kingbird
Fork-tailed Flycatcher
Yellow-hooded
Blackbird
Oriole Blackbird
Total


a Condition of sale C=dead fresh or smoked, L= Live animal, sold as pet.
* Species regularly hunted for subsistence purposes in the rural sector.


C

C,L
L
C,L

L

L

L

L

L
L
C,L

C,L

L
C

C
C
C

C

C

C

C


2

2
2

3

3

3


3.8

2.4
7.1

6.4
26.2

33.3

23.8

7.1

14.3
7.1
0.5

1.5

5.1

0.2
1.0
1.4

0.6

0.1

0.1

0.1

0.2


2

2

2
3

346

1018

5
9

3
5
17

298
242

360

463
2791


7

4
12
22

90

114

41

12

24
24
160

1520

26
2

3
7
10

43

35

34

85
2276








why hunters take what they do. Hunters preferred large birds with good quality meat,

but their preferences were in proportion to the extent that birds were available to them.

Bird availability is influenced by human population densities, settlement patterns, and

age of the villages coupled with environmental pressures on bird populations.

Additionally, the take of birds may be influenced by constraints in access to ammunition

and by differential availability of alternative sources of animal protein. I have shown

associations between factors that influence what hunters take and why they take what

they do. However, factors are closely interrelated perhaps in an inextricable manner.

Characteristics of human settlements influence the composition of the fauna

available to hunters. Large species of wildlife tend to become increasingly rare near

villages, and the extent of ratification and local extinction is linked to the age of the

settlement, number of hunters, and settlement patterns (Stearman and Redford 1995).

These characteristics fit differences in the historic background of NE and SP, and explain

the harvest composition obtained from short trips in both villages.

Differences in population abundance between upland and varzea forest influenced

what hunters took. Species abundance varies among Amazonian ecosystems (Emmons

1984), perhaps in response to differences in availability of resources (Chapter 2). Among

preferred species, greater abundance clearly reflected greater harvest in SP and NE.

Greater abundance of less preferred species in varzea forest (herons, cormorants and

anhingas) also resulted in greater representation in the harvest yields.

Habitat modification and extractive activities may influence the availability of

birds to hunters. Agricultural gardens and disturbed forest have been associated with

greater abundance of certain species of wildlife (Balee 1984, Denevan et al. 1984).








Moreover, Posey et al. (1984) argued that certain species would not occur in forest that

were not modified by humans (but see Jorgenson (1993). Tinamous for example seem to

take advantage of productive second growth and agricultural crops (Leopold 1959,

Jorgenson 1993, this study), and were abundant and frequently hunted in areas

surrounding of the human settlements.

Palm extraction is likely to have the opposite effect on species that depend on

palm fruit. Palm fruit appears to be an important food resource for P. jacquacu and P.

cumanensis given the wide spatial (upland and varzea forest) and temporal (18 months)

sampling and the high percentage of occurrence of palm fruit in their stomach samples.

Hence, the low availability of food areas surrounding the settlements, may help explain

these species' absence and rarity and their little representation in harvest yields from

short hunting trips.

The association between habitat transformation and palm extraction with variation

in population abundance was consistent in SP and NE, and appeared to vary with the

magnitude of the causal factors. NE is an older settlement and has a greater human

population density than SP. Because palms used by cracids represent and important

building material and source of income to riberefio people (Vazquez and Gentry 1989), it

is likely that greater human population density in NE have exerted a greater impact on

palm trees. While still exploratory given the nature of the evidence, bird populations

appeared to correspond to hypothesized differences in habitat modification in NE and SP.

Populations of the tinamous were greater in the hunted grounds of NE (Chapter 4) and

were hunted more frequently than in SP. Conversely, P. jacquacu has been locally

extirpated and P. cumanensis is rare in the hunting grounds near NE, while they still









persist and are taken from the grounds near the settlement in SP. Some village residents

recognized the importance of palm trees for the cracids and deliberately left standing

trees to attract them.

Hunting technique. Specific hunting techniques often determine the species and

number of individuals taken by hunters (Hames 1980, Alvard 1995). Riberefo people in

both forest types mostly used the same hunting technique but differed in the use of

fishing nets. The use of fishing nets clearly influenced the greater harvest of diving birds

in NE. However, it was more an ecological consequence rather than the use of a specific

hunting technique given the opportunistic nature of taking birds with fishing nets (birds

were accidentally entangled in fishing nets). Permanent bodies of water, fish and fish-

eating birds are ubiquitous in varzea forest and people fish with nets more than in SP.

Data showed that high cost and limited access to ammunition led to low harvest

rates of preferred species. Hunters in communities with access to ammunition near large

cities are less concerned about the cost of ammunition (Vickers 1980) and more likely to

take a wider range of birds including less preferred species and even small passerines. In

remote communities where the cost and access to ammunition becomes more difficult

birds were not consider a prey due to disproportional low return for the value of

ammunition (personal observation). Hence, it can be generalized that communities

located in areas where ammunition is readily available are more likely to exert a larger

impact on bird populations. In addition the harvest of birds is characterized by the

predominance of small and less preferred species. Other factors are also involved.

Villages located in more accessible areas typically have greater human population

densities and seem to have a greater impact on on populations of Amazonian birds.







Intensity of the harvest. The harvest of birds was greater in SP than in NE.

Gross (1975) suggested that the access of animal protein is perhaps the most important

determinant of the number of individuals in a group, and nomadic movements among

indigenous people in Amazonia. Other studies (Smith 1976, Ayres et al. 1991) have

indicated that pressure over wildlife was relaxed in the presence of alternative sources of

food (e. g., domestic animals, canned food). Birds and mammals used for household

consumption were hunted at similar intensities and assumed to be equally available in

both villages. But, domestic animals were more available in NE. Another, and perhaps

the most important, difference that help explain the greater pressure over birds in SP is

the access to fish. Varzea forest has rich fisheries (Bayley et al. 1992) and fish is a staple

food in NE (Gonzales 1998). Basso (1973) pointed out that fish played an important role

in the diet of indigenous people with access to this resource, to the extent that land

animals were rarely hunted or considered low quality meat. In contrast lower availability

of domestic animals and fish in upland forest ecosystems (Brack 1990) helped explain a

greater hunting pressure on birds in SP.

Birds commonly hunted in both villages had little monetary value. In the

Peruvian Amazon the sale of wildlife represents an importance source of income to

riberefio people. However, birds did not represent a source of income and were entirely

consumed at the household level indicating that unlike Amazonian ungulates, monetary

incentive is not a factor influencing the harvest of Amazonian birds.








Conclusions

There are several factors that influence the type and amount of birds that riberefio

hunters take. Overhunting associated with the age of the human settlement, human

population density, settlement pattern and habitat degradation appear to influence the

source fauna available to hunters. Hunters preferred some species to others, but their

preference was realized to the extent of the availability of species largely due to

consequences of history and characteristics of human settlements. The availability of

species was also influenced by habitat differences between forest type, therefore, what

hunters took. Cost and access to ammunition was also an important determinant of the

type and quantity of birds taken. Differences in the type and quantity of birds taken in

short and long trips provide an index of the impact that hunting has on populations of

Amazonian birds. Finally, a greater access to fish, potential for agriculture and animal

husbandry appear to reduce the need for animal protein for household consumption. This

was observed in NE, whereas in SP limited access to these resources appears to result in a

greater pressure over birds.













CHAPTER 4
HUNTING AND POPULATION ABUNDANCE OF AMAZONIAN BIRDS

Introduction

Wildlife in the humid lowland rain forest has traditionally been an essential

source of protein for people (Dufour 1990). Despite significant cultural and socio-

economic changes, hunting of forest wildlife is still an important activity that provides

meat for household consumption and for the market (Dourojeanni 1985, Padoch 1988).

However, wildlife resources are susceptible to overexploitation and species can be driven

to local extinction.

Determining whether hunting in tropical forests is sustainable is difficult because

one needs to understand the wildlife population status, the productivity of populations,

the response of populations to hunting, and the extent of variation in patterns of hunting

(Robinson and Redford 1991). These data are seldom available for rural communities or

for tropical forest species. However, the need to evaluate the sustainability of hunting

has resulted in the use of indices and models that provide a first evaluation of

sustainability (Safirio and Hames 1983, Stearman 1990, Vicker 1991, Robinson and

Redford 1991, Bodmer 1994). Models commonly used to evaluate sustainability of

hunting rely on several assumptions sometimes derived from temperate species. Hence,

evaluating sustainability of hunting have used several models intended to embrace the

intricacies of natural systems and predict sustainabiltiy of hunting based on the consensus

of results from the various models (Fitzgibbon et al. 1996, Coltrane 1998).








In this study, I examined the association between hunting and population

abundance of Amazonian birds. The influence of hunting was inferred from comparisons

of population abundances in hunted and unhunted (control) sites. I evaluated

sustainability of hunting using models that rely on some knowledge of the biology of the

harvested species, population densities, hunting pressure, and assumptions derived from

studies of temperate species.



Methods

Sampling Design

This study evaluates the association between hunting and population abundance

of Amazonian birds using a comparative observational design (Cook and Campbell 1979)

involving two hunting systems. Each hunting system was composed of a human

settlement, their hunting grounds, and distant areas within upland and varzea forest

subject to no human hunting. One hunting system was located in and around the upland

forest of the Tamshiyacu-Tahuayo Community Reserve (TTCR); (Figure 4-1), and the

other in the varzea forest ofPacaya Samiria National Reserve (PSNR) (Figure 4-2).

Within each hunting system, I compared population abundance observed in sites with

differing hunting intensities; heavy, moderate, and light hunting with sites subject to no

hunting (control). A site was defined as an area that represented an intensity of hunting.

Sites were constituted by several sampling trails or line transects. Habitat was kept as

constant as possible within the upland forest sites, and within the varzea forest sites. This

assumption is based on the sites being located in continuous upland and varzea forest not

separated by any major barrier (e.g., a major Amazonian river), similar plant













































Figure 4-1. Map of the area sampled in upland forest showing the sites with the
three intensities of hunting, l=Heavy, 2=Moderate, 3=Light and the control 4.
The study area in upland forest is located within Tamshiyacu-Tahuayo Communal
Reserve in the state of Loreto, Northeastern Peru.













































Figure 4-2. Map of the area sampled in varzea forest showing the sites with the
three intensities of hunting, 1 =Heavy, 2=Moderate, 3=Light and the control 4.
The study area in varzea forest is located within Pacaya-Samiria National
Reserve in the state of Loreto, Northeastern Peru.







(Encamrnaci6n 1993) and wildlife communities. Sites with differing intensities of

hunting were chosen on the basis of both, the distance between sites and human

settlements, and hunting intensity at these sites. Hunting intensity was determined from

interviews with hunters on the frequency of hunting trips to these sites (Table 4-1).



Table 4-1. Sites with differing intensities of hunting in upland and varzea forest.
Distances are straight lines connecting the village with the sites sampled. Annual
frequency is the percentage of trips to the area sampled. Village residents provided
annual frequency of trips.

Nueva Esperanza (Varzea) San Pedro (Upland)'
Distance Annual Distance Annual frequency
(km) frequency (%) (Km) (%)
Heavy (1)* 4.5 41 3 50
Moderate (2) 14 18 18 25
Light (3) 22 9 26 25
Unhunted (4) 43.3 0 135 0
(*) Number in parenthesis identifies the location of sites on the map.
(1) (P. Puertas In preparation).



The comparison between hunted and unhunted sites was conducted in the

protected areas of TTCR in upland forest, and PSNR in varzea forest, both in the state of

Loreto. Areas within TTCR and PSNR were compared to hunted areas located in their

periphery, near the human settlements of Nueva Esperanza (NE) and San Pedro (SP). In

upland forest, the centers of the hunted sites and the unhunted control are separated by a

distance of 135 km. In varzea forest, the centers of hunted and unhunted areas were

separated by a distance of 43.4 km. Finally, the areas that encompassed all sites within

upland forest and all sites within varzea forest was separated by a distance of 230 km.

Distances were estimated with a geographic positioning system (GPS, Garmin 12XL).









After determining the association between the intensity of hunting and patterns

of population change from site to site, I determined the sustainability of hunting. The

evaluation of the sustainability of hunting uses population densities obtained in unhunted

sites within each forest type and pooled data from hunted sites with the differing

intensities of hunting within forest type.



Relative Abundance and Population Density

I measured abundance as the number of individuals sighted per 100 km of line

transects surveyed. Trails were cut in a predetermined direction using a compass, and

were located along rivers, usually within 1.5-2 km from each other. Birds were censused

by walking a trail at an approximate speed of 1 km/hour. I first recorded the species and

the number of individuals per sighting, then proceeded to tape measure the perpendicular

distance between the trail and the initial position of the animal. When a group of animals

was sighted, I measured the perpendicular distance between the trail and the center of the

group at its initial position.

A total of 28 line transects that made up 113 km were opened in upland forest.

Likewise, a total of 24 line transects that made up 102 km were opened in varzea forest.

Line transects averaged 4 km in length. Censuses were conducted between the 0700-

1530 hours. Each line transect was censused once per day and up to 6 times after

opening.

To census water birds, I also conducted censuses along sections of rivers and

small tributaries by means of dugout canoes powered by paddles and/or small outboard

motors. A total of 9 separate sections of rivers and tributaries were censused in upland








forest. Likewise, a total of 11 separate sections of rivers and tributaries were censused

in varzea forest. Each section averaged 7 km in length and was censused up to 3 times.

Censuses in the unhunted sites in varzea forest were conducted during the month of June

1996. Censuses in the hunted sites in varzea forest were conducted in the months of

April through June 1995 and June through July of 1996. Censuses in the unhunted sites

of upland forest were conducted during June 1997, and in the hunted sites during

February, March, and April 1997.

Not all species were recorded during all of the censuses. I differentiated between

interior forest dwelling species, tinamouss, cracids, trumpeters, wood-quails and doves),

and species of open habitats (anhingas, herons cormorants, ducks, parrots, macaws,

toucans). Species of open habitats are those that inhabit lakes, rivers and the top of the

forest canopy. Interior forest species were recorded in all of the sites with the differing

intensities of hunting. Species of open habitats were recorded in sites with heavy and

moderate hunting and the control sites. Censuses were conducted by multiple parties.

Each surveying party was composed of one biologist and a field assistant.

Population densities for all species except water birds were estimated using the

program "Distance" (Laake et al. 1994). The program uses the probabilities of detection

of randomly distributed objects (animals) to estimate density within the sampled area

(Buckland et al. 1993, Pp. 2). To obtain the density of water birds, I used Overton's

(1971) procedure. Because distances between the surveyor and the position of water birds

were not obtained during water surveys, I used a fixed width (width of the watercourse

surveyed).


D= N/W.L