Citation
Effects of the Samuel Hydroelectric Dam on mammal and bird communities in a heterogeneous Amazonian lowland forest

Material Information

Title:
Effects of the Samuel Hydroelectric Dam on mammal and bird communities in a heterogeneous Amazonian lowland forest
Creator:
Lemos de Sá, Rosa Maria
Publication Date:
Language:
English
Physical Description:
ix, 140 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Animals ( jstor )
Biomass ( jstor )
Birds ( jstor )
Dams ( jstor )
Ecology ( jstor )
Forests ( jstor )
Mammals ( jstor )
Natural reservoirs ( jstor )
Primates ( jstor )
Species ( jstor )
Birds -- Effect of dams on -- Brazil ( lcsh )
Dams -- Environmental aspects -- Brazil ( lcsh )
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF ( lcsh )
Hydroelectric power plants -- Environmental aspects ( lcsh )
Mammals -- Effect of dams on -- Brazil ( lcsh )
Wildlife Ecology and Conservation thesis, Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 134-139).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Rosa Maria Lemos de Sá.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022272618 ( ALEPH )
34360609 ( OCLC )

Downloads

This item has the following downloads:


Full Text













EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST


















By


ROSA MARIA LEMOS DE SA


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


1995































Copyright 1995

by

Rosa Maria Lemos de SA
































To Joio Paulo and Daniela My sources of strength and inspiration























ACKNOWLEDGMENTS


Most students consider themselves lucky if they find the right

academic advisor; I consider myself blessed for having found not one but two! John Robinson was my first advisor, who helped define my work and gave support during the initial field work. John Eisenberg took over when Robinson left the University, and he has provided valuable advice and moral support throughout this lengthy portion of my work. I thank both of them for their help and guidance! Kent Redford was greatly responsible for my interest in the University of Florida, and together with his family made life in Gainesville even more enjoyable. His knowledge and keen interest in the Neotropics has often motivated me, and for that I am thankful. I am also extremely thankful to Ron Labisky, Jay Malcolm, Richard Bodmer, Doug Levey, and Mel Sunquist for their comments, advice, and editorial help. Special thanks go to Richard Bodmer for agreeing to substitute for John Robinson at such short notice.

Field work in Rond6nia would not have been possible without the technical support from ELETRONORTE. Several people have greatly helped with the bureaucracy, allowing me to do field work: Edgar Menezes Cardoso, Bruno Payolla, Rubens Guilhardi, and Carlos Fabbris. I also had


iv











excellent field assistance from Carry Ann Cadmam, Rodrigo Mariano, Barroso, Antonio, and Chico. David Oren, Jos6 Maria Cardoso, and Chris Canaday identified bird specimens and provided exceptional friendship.

This study was generously supported by the World Wildlife Fund WWF/US, Conservation International CI, Lincoln Park Zoo "Scott Neotropical Fund", Tropical Conservation and Development Program TCD, Tinker Foundation, and the Program for Study in Tropical Conservation PSTC. Various individuals from these institutions simplified my life by providing easy flow between finances and field work: Gustavo Fonseca, and Sonia Rigueira from CI, Cleber Alho, and LouAnn Dietz from WWF-US, Steven Thompson from the Lincoln Zoo, and Kent Redford, Steven Sanderson, and Peter Polshek from TCD. The Conselho Nacional de Desenvolvimento Cientifico e Tecnol6gico CNPq granted me a scholarship for my academic work. The Empresa Brasileira de Pesquisa Agro-Pecudria EMBRAPA analyzed my soil samples. The Centro de Pesquisas para Conservagao das Aves Silvestres CEMAVE/IBAMA supplied bird banding permits and aluminum bands.

Life in Gainesville has been enriched by the ephemeral presence of friends such as Jay Malcolm, Justina Ray, Joe Fragoso, Chris Canaday, Wendy Townsend, John Payne, Ann Edwards, Miriam Marmontel, Andres Navarro, Susan Walker, Denise Imbroise, Dener Martins, Peter Crawshaw, Damian Rumiz, Cldudio and Suzana P~dua, Rajanathan Rajaratnam, and many others.

I am much grateful to Jodo Paulo Viana for his endless help

throughout all phases of this work. His love, patience, and assistance during the final weeks were specially cherished.


v


















TABLE OF CONTENTS




ACKNOWLEDGMENTS.......................................................iv

ABSTRACT............................................................viii

CHAPTERS


1 INTRODUCTION.....................................
Background....................................
The Samuel Hydroelectric Power Plant.......
Rescue Operations... a History................
The Samuel Rescue Operation................
Research Design...............................


~2 ~4 ~9
........10


2 CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION..............11


Introduction..............................................


Methods..............................
Study Site........................
Temperature and Precipitation.....
Nutrients in Soil.................
Vegetation........................
Results..............................
Temperature and Precipitation.....
Nutrients in Soil.................
Vegetation........................
Canopy.........................
Understory.....................
Discussion and Conclusions...........


.......11
.......12
.......12
.......14
.......14
.......15
.......15
.......15
.......18
.......21
.......21
.......24
.......27


3 INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A


CONSEQUENCE OF DAMMING................


Int
Met





Res


...........................30


roduction.....................................................30
hods..........................................................32
Study Site....................................................32
Data Collection...............................................32
Data Analysis.................................................33
Biomass estimation.........................................35
ults..........................................................35
Sightings per Kilometer Walked................................35
Species Observed..............................................37
Primates..............................................................38
Primate density estimates prior to damming..................38
Individuals captured and released..........................39
Primate density estimates after damming....................39
Density and body weight of primates........................42
Total primate density estimates............................42
Density changes between years at the Reserve...............45
Biomass....................................................48
Terrestrial Diurnal Mammals...................................52
Density estimates prior to damming.........................52


vi


........
........
........
........
........
........


........
........
........
........
........
........
........
........
........
........
........
........


........
........
........
........
........
........
........
........
........
........
........
........










Individuals captured and released..........................52
Density estimates after damming............................53
Total terrestrial mammal density estimates..................54
Biomass ....................................................56
Discussion and Conclusions.......................................58
Primates......................................................58
Density estimates..........................................58
Density changes between years at the Reserve...............66
Density comparisons with other western Amazonian sites.....67 Total density..............................................67
Biomass....................................................69
Terrestrial Diurnal Mammal....................................74
Density............. ....................................... 74
Total density..............................................75
Biomass......... ........................................... 77

4 UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL


DAM......................................................


Introduction......................
Methods...........................
Study Site.....................
Study Design...................
Data Collection................
Data Analysis..................
Results...........................
Inter-Year Comparison..........
Overall Site Comparison........
Species Richness...............


Guild Structure...........................
Faunal Similarity Between and Within Sites Ecological Similarities Among Sites....... Discussion and Conclusions...................
Overall Site Comparison...................
Species richness.......................
Species abundance......................
Guild Structure...........................
Faunal Similarities.......................
Ecological Similarities...................


-.. .. ........ 78
........ 78
........79
........79
........79
........80
........81
........83
........83
........84
........86
........90
........93
........97
........97
........97
........97
.......101
.......101
.......102
.......103


5 RECOMMENDATIONS FOR FUTURE HYDROELECTRIC DAM CONSERVATION PROGRAMS.105
Overview........................................................105
Considerations..................................................107
Rescue Operations............................................107
Creation of Conservation Units...............................111
Preliminary Studies..........................................112
Follow-up Studies............................................113
Conclusions.....................................................114

APPENDIX A MAMMALIAN SPECIES RECORDED AT THE SAMUEL DAM SITE......... .115

APPENDIX B BIRD SPECIES CAPTURED DURING THIS STUDY...................118

APPENDIX C BIRD SPECIES RECORDED AT THE SAMUEL DAM SITE..............124

LIST OF REFERENCES...................................................134

BIOGRAPHICAL SKETCH..................................................140


vii


.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
.........


.........
.........
.........
.........
.........
.........
.........
.........
.........
.........
















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


EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD
COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST By

ROSA MARIA LEMOS DE sA

December 1995

Chairman: Dr. John F. Eisenberg Major Department: Forest Resources and Conservation (Wildlife Ecology and Conservation)


Two sites located near a hydroelectric dam reservoir, in the northwestern region of Brazil, were monitored in order to document changes in mammal and bird communities brought about by: a) the creation of the reservoir, and b) by the release of rescued animals into one of the study areas. Mammals were sampled with transect surveys, and understory bird communities with mist nets.

Primate densities and biomass at the Samuel Ecological Station (located at the southeast border of the Samuel Dam reservoir and hereafter referred to as reserve) increased after the Samuel Dam flood gates closed. Primate biomass in the reserve was 154 kg km-2 in 1988,

-2
prior to damming, and increased to 165 ( 65) kg kMC in 1989, immediately after flooding, due to the release of animals captured in the area flooded by the reservoir. However, because of shallow water, most primates remained in the flooded forest for the first year since the vegetation in the area was still alive. By 1990, all woody


viii










vegetation in the flooded area was dead, and primate biomass at the reserve increased to 255 ( 109) kg km because of the migration of animals from the reservoir to the reserve. By 1991, primate biomass in the reserve returned to levels similar to 1988, 153 ( 81) kg km-2, most likely due to dispersal of animals to adjacent areas. The pattern of biomass fluctuation indicates that primate carrying capacity at the reserve is about 150 kg km2

Terrestrial mammal densities and biomass at the Samuel Ecological Station increased in 1988, immediately after flooding, and remained high until 1990 when they started decreasing.

Understory bird community structure and composition was sampled at the Samuel Ecological Station, and at an area located at the Jamari River's edge, downstream from the dam. The creation of the Samuel Lake destroyed approximately 214 km of pristine riverine habitat, displacing an entire community of birds. Despite the creation of a water-edge habitat by the reservoir, the composition of bird species at the lake's edge is different from that of the downstream site, therefore, different from previously existing river edges at the reservoir site. However, community structure seems to be similar at both sites. Water-related niches were created at the reservoir's edge, yet, the uniqueness of species composition at riverine habitats was not emulated.


ix

















CHAPTER 1
INTRODUCTION


Background


Socioeconomic demands on land in Brazilian Amazonia are very high. Land officially designated for development totals 2,100,000 km2, or 65% of the total Amazonian area; of this area 4.4% is to be flooded by hydroelectric development (Johns 1988). Eletrobras, the Brazilian electric company, identified 80 potential dam sites in the Amazon region in its Plan 2010 (Serra 1989). The objective of the agency is to stimulate development in the region by attracting investors with inexpensive energy sources.

Until 1980, only 2 small hydroelectric dams were operating in the Amazon: Curui-Una, near Santar~m, and Paredio, in Amapi state. Each dam impacted an area less than 100 km2 (Junk and Nunes de Mello 1987). Since then, 3 large dams have been added to the region and are operating in the Amazon; Tucurui, near Bel6m, Balbina, near Manaus, and Samuel, near Porto Velho. Collectively, these 3 dams have flooded an area of 5,350 km2. If Eletronorte, the Brazilian Agency for Hydroelectric Power Development in the Amazon region, succeeds in completing all the dams projected for the Amazon in the 2010 plan, an area of roughly 100,000 bm2 will be flooded (Fearnside 1989).

The flooding of such large areas has a tremendous impact on humans and wildlife. The most significant effect is the loss of land, which


1





2


causes human and animal displacement and/or death, and can also bring about extinction of species (Liao et al. 1988).

The Amazon region is well known for its high plant and animal

diversity. However, the distribution and densities of both plants and animals are basically unknown in Amazonian regions. Despite their negative impact on fauna and flora, hydroelectric dams provide good opportunities for researchers to conduct detailed studies on local distribution and densities of plants and animals if they are contacted when the first feasibility studies begin, long before the creation of the reservoir. Instead, the power companies invite scientist to research the area 1 or 2 years prior the completion of the project, yielding only a short-term evaluation. As a result of this policy, very little was learned of the impact on wildlife resulting from the 3 dams constructed most recently.

This study documents the impacts on avian and mammalian

communities created by the construction of the Samuel Hydroelectric Power Plant on the Jamari River, Rond6nia, Brazil. The Samuel Hydroelectric Power Plant


The Samuel Hydroelectric Power Plant is located on the Jamari

River, a right bank tributary of the Madeira River, Rond6nia, 80 45'S 63* 25'W. The site is 52 km east of the city of Porto Velho, the state's capital, and 96 km from the confluence with the Madeira river (Fig. 11). The Jamari River basin is entirely located in the State of Rond6nia; its watershed is located between 80 28' 110 07'S, and 620 36' 630 57'W. Its head waters are located in the Pacais Novos Mountain chain at






3


N 66 I62! 8









Manaus 4*









BR AZIL
ZVI* Is







I
/
I





-. Porto
Velmo





0 RONDONIA

.R iV 0r s 12*


sale lng
D|


Figure 1-1: Location of the Samuel Dam on the Jamari River.





4


an altitude of approximately 500 m; its total extension is approximately 560 km.

The reservoir, with a volume of 3.2 billion rn3, occupies 560 km2 at normal maximum operating level of 87 m above sea level (Eletronorte 1990), and has the capacity of generating 216 MW of energy. The reservoir, lying in a southeastern orientation, is 40 km in length; its width spans 15 to 20 km in the first 25 km, and between 3 and 1 km for the remaining 15 km. The construction of the hydroelectric power plant started in 1982, and because of the flat terrain characteristic of the region, 57 km of dikes were constructed along the right and left banks in order to contain the river water (Fig. 1-2).

To guarantee conservation of the local fauna and flora,

Eletronorte created the Samuel Ecological Station, 20,865 ha in size, adjacent to the area of influence by the Samuel Dam. However, roughly 20% of the Reserve was lost due to inundation after completion of the reservoir (Mozeto et al. 1990, Eletronorte 1993) (Fig. 1-3). The vegetation cover at the Reserve, before the flooding, consisted of 96.6% terra firme, 2% temporarily flooded forest, and 1.4% secondary forest (Eletronorte 1993). The only road into the Reserve area parallels the dam (Fig. 1-3). Hunting does not occur within the Reserve because of controlled access.


Rescue Operations... a History


To ameliorate the effects of habitat loss on wildlife, Eletronorte employed "rescue operations" to save terrestrial animals once the reservoir began to fill. Drowning animals is bad publicity, which






5


Figure 1-2: Satellite image of the Samuel Dam Reservoir under construction (From Instituto Nacional de Pesquisas Espaciais INPE, July 1987).






6


I


Figure 1-3: Satellite image of the Samuel reservoir, and the Samuel Ecological Station (From Instituto Nacional de Pesquisas Espaciais INPE, August 1992).


. %I --. .


I -,





7


undoubtedly was a relevant factor in the decision by Eletronorte to promote a rescue operation. However, the rescue operations created a moral dilemma among Brazilian and foreign scientists. Was it worth spending millions of dollars on such operations, which did not guarantee the survival of the animals? Or was it more profitable, for environmental conservation, to use the money to create new reserves and to further maintain the already existing ones?

Eletronorte ignored such questions, and initiated the first rescue operation at Tucurui in 1984-85; a total of 284,211 animals (vertebrates and invertebrates) were captured and translocated (Eletronorte 1985). Unfortunately, Eletronorte had not conducted a preliminary study to define appropriate sites for release of the captured animals; thus, they were released on the nearest piece of dry land (Johns 1986).

The Tucurui rescue operation cost US $30 million and employed 300 people (Johns 1986). The cost of rescuing 222,544 vertebrates was $134.80 per individual; considering only the rescue of 107,094 birds and mammals, the cost was $280.13 per individual. Furthermore, based on crude estimates of primate densities, only 4% of the tamarins (Saguinus midas), 6.4% of squirrel monkeys (Saimiri sciureus), 6.9% of capuchin monkeys (Cebus apella), 4.2% of saki monkeys (Chiropotes satanas), and 29.7% of howler monkeys (Alouatta belzebul) were rescued (Johns 1986). According to Johns


these results suggest that rescue operations will remove only a
small proportion of primates. It is likely that there will be
critical overpopulation of lake fringe areas caused by the vast
majority of animals escaping unaided, which suggests that
releasing captured animals on the lake shore is worse than
useless. In fact, the real value of rescue operations is called
into question. (Johns 1986, pg. 20)





8


Despite the unsuccessful results of the Tucurui rescue operation, a similar rescue operation was carried out by Eletronorte at Balbina in 1987. Although the area of Tucurui and Balbina were similar, the number of animals rescued at Balbina was only about 10% of the number rescued at Tucurui (Gribel 1990). The Balbina reservoir is much shallower, which limited boat travel and thus, reduced the number of animals rescued. As at Tucurui, the rescued animals were released with the same carefree attitudes as before, and most likely did not survive the pressures of high densities and hunting.

Only in 1988, after much pressure from the scientific community in Brazil, did Eletronorte change its rescue operation policies. A new rescue plan was developed because it was obvious that releasing animals at random did not benefit the animals rescued or the community into which they were released.

Museums, research institutions, universities, and zoological

gardens were contacted throughout Brazil, before the Samuel Dam rescue operation started, and offered live specimens, study skins, and display skins of species rescued in the Samuel dam reservoir. Priority was given to research institutions that could provide housing facilities for the animals, and to well-known researchers willing to work at the Samuel site during the operation. Most importantly, an ecological station was created by Eletronorte to be the focus of conservation studies, and the only site of animal release.

Several inventories were carried out inside the Reserve prior to the flooding of the reservoir, including vegetation, mammal, reptile and bird surveys, and soil analyses. The projects were conducted in an attempt to understand the community before the translocation of the






9


rescued animals, so that a follow-up would bring some understanding to the changes brought about by the dam construction. The Samuel Rescue Operation


The rescue operation started in November 1988, immediately after the closing of flood gates. The objective of this operation was to recover every animal found stranded on small islands created by the filling of the reservoir. The operation lasted 4.5 months, during which time 16,000 animals were rescued. The rescued animals included 6,590 arthropods, 3,729 mammals, 3,504 reptiles, 2,099 amphibians, and 78 birds (Eletronorte 1989). Of the 16,000 animals rescued, 11,417 were sent alive to research institutions, and 1,729, which were sacrificed or died during the operation, were sent to research institutes or museums. Only 2,854 were released inside the Reserve. The number of animals released at the Samuel Ecological Station was significantly less than the number of animals released at previous dam sites. However, these animals, combined with the animals that moved on their own to the Reserve site, have probably caused some impact on the populations previously inhabiting the site.

The influence of the dam and its reservoir is not a one-time

phenomena. During the dry season, the reservoir shrinks to an estimated 40% of its fullest extent. The approximately 300 km2 area which will be exposed each year has the potential of altering adjacent biological communities in a major way.





10


Research Design


In order to document the changes that occurred after the

translocation of animals, I conducted research at the Samuel Ecological Station during the months of June to October of 1989, 1990 and 1991. My objectives were to sample both large diurnal mammals and understory passerine birds both inside and outside the Reserve area.

My original experimental design included a control site for the Reserve. The control was needed to represent mammal densities, and bird community composition and structure in the Reserve prior to any flooding because I started my work after the filling of the reservoir. However, it was not possible to find an accessible area with the same characteristics as the Reserve that was not affected by the flooding. I then modified my original design and sampled an area downstream from the dam (Jusante), which had similar characteristics to the area that was flooded by the reservoir. Therefore, Jusante was a control site for the area flooded by the reservoir, and not the Reserve.

This work represents the first attempt to understand community changes brought about by hydroelectric dams in the Amazon. There is still a lot of work to be done; however, I hope that the results of my research will stimulate similar projects at future dam sites, and that it will influence Eletronorte officials in their future environmental decisions. Eletronorte should consider environmental impact studies and rescue operations more seriously and make decisions based on scientific facts rather than public indulgence.

















CHAPTER 2
CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION Introduction



Amazonia occupies nearly 6,000,000 km2, with more than half of this in Brazilian territory. Viewed from the air, the Amazon forest appears quite homogeneous; however, when examined in detail, considerable local variations of vegetation and floristic composition are encountered (Pires and Prance 1985). The central plateau of the Amazon Basin is limited to the north and to the south by moderate slopes and to the west by the Andes, opening only to the east where it receives hot and humid winds from the Atlantic. These unique characteristics plus its latitudinal position give the Amazon region singular conditions, such as almost constant daylength throughout the year, unchanged solar energy at the limit of the earth's atmosphere, and little variation in average monthly temperatures (Salati 1985).

Precipitation in the Amazon basin varies from 1,500 to > 3,000 mm annually (IBGE 1977). A well defined dry season (from May through September) is common in the central region.

Due to effects of high temperatures, high rainfall, and geology of the region, the soils of Amazon rainforest have low potential for supplying nutrients to plants. Intense weathering and leaching over millions of years have removed the nutrients from the minerals which form the parent material of the soil (Jordan 1985).


11





12


Forest physiognomy in the Amazon basin is influenced not only by soil but also by the age of the vegetation at specific sites (Lisboa 1990). Water levels in the past and in the present have great influence in the formation of the vegetation. Younger vegetation formations, present today, are located in areas that were submerged in the past.

The objective of this chapter is to characterize the climate,

soil, and vegetation in the area, in order to understand the differences and/or similarities of the study sites and their associated fauna.


Methods




Study Site


The Samuel Hydroelectric Dam is located in the state of Rond6nia, 50 km east of the state capital of Porto Velho. Two sites were selected for the study. The first site was the Samuel Ecological Station with an area of 21,000 ha, located at the southeast border of the Samuel Reservoir approximately 26 km (straight line) from the dam. The Reserve was created by Eletronorte (the Brazilian Agency for Hydroelectric Power Development in the Amazon Region) in an attempt to establish a protected site for the animals rescued from the reservoir area during flooding. The second site was located approximately 3 km below the dam, and will be referred to as Jusante. The Jusante site was comparable, floristically as well as faunistically, to 50% of the area that was flooded by the reservoir (Fig. 2-1).

Within each of these sites, 3, 1 km2 plots were used to sample soil and vegetation (Fig. 2-1). Plots at Jusante were deliberately













9038000 9028000 901000


STUDY SITE

















AV
8
0








0 0









JUSANTE



Figure 2-1: The Samuel Dam Reservoir area showing study sites (Reserve and Jusante), and study plots (1, 2, 3, 4, 5, A, B, C, and D).





14


located 500 m from the River's edge in order to avoid flooded forest, which has a unique but different set of vegetative characteristics. Temperature and Precipitation


Temperature and precipitation data were collected by the Engineering Department at the Samuel Hydroelectric Dam. The climatological station at Samuel is located at 8' 45'S 630 28'W, at an altitude of 80 m, and has been operating since July 1977 (Eletronorte 1988a).

Daily average temperature was recorded with the use of a maximumminimum temperature thermometer. Rainfall was recorded on a daily basis with a rain gauge.


Nutrients in Soil


Soil samples were collected in 6, 1 km2 plots (plots 1, 2, 3, A, B, and C; Fig. 2-1). Samples were taken along a straight line every 100 m for 1 km in each of the plots, totaling 11 stations in each plot. Soil was collected, using a soil auger, at depths of 0-20 cm, 20-40 cm, and 40-60 cm in each of the stations, which yielded 33 samples/plot. Care was taken to collect soil from a non-disturbed area (therefore top soil would be intact), and to avoid contamination of lower depth samples by upper samples.

Samples were kept in marked plastic bags and taken to Embrapa's (Empresa Brasileira de Pesquisa Agropecudria) soil analysis laboratory in Porto Velho for pH and nutrient (phosphorus [P], potassium [K], calcium [Ca], aluminun (Al], and magnesium [Mg]) analysis within 2 days of collection.





15


Vegetation


Vegetation samples were taken for purpose of characterizing the vegetation structure at both sites. A 1-km transect was sampled in each of the 6 plots (soil and vegetation were sampled along the same 1-km strip). The point-centered quarter method was used to sample large trees (Knight 1978), which divides the area around a sample point into 4 imaginary quarters. Measurements of distance were taken from each point to the center of the nearest tree in each quarter. Points in the transect were 20 m apart, totaling 51 stations, and 204 trees measured per transect. Measurements recorded included DBH (diameter at breast height), and height (estimated). Only trees > 10 cm DBH were included.

Understory vegetation was sampled within 5 x 5 m quadrats at

alternate stations in each of the transects, totaling 26 quadrat sample stations per transect. Within each quadrat all trees with DBH ranging from 1 to 10 cm were measured, and all vegetation with DBH < 1 cm (seedlings) was counted.


Results



Temperature and Precipitation


The average annual monthly temperature in the Samuel region is 27*C; however, lower monthly temperatures are recorded during June, July, and August due to a phenomenon called friagem, which is a cold front sweeping over the continent from Antarctica (Table 2-1, Fig. 2-2). This event, which lasts for only a few days, will lower temperatures considerably, especially during nighttime.






16


Table 2-1: Average monthly temperature
(*C) at the Samuel Dam, during 1989, 1990, and 1991 (source Eletronorte).



Month Year
1989 1990 1991

January 27.0 27.2 26.7
February 26.0 26.7 26.8
March 27.0 26.4 26.4
April 27.0 26.6 27.0
May 27.0 26.7 26.6
June 26.5 26.1 26.1
July 25.0 25.7 24.4
August 27.0 27.0 25.5
September 28.0 27.0 26.6 October 27.0 26.4 27.3
November 28.0 28.0 27.8
December 28.0 28.0 27.4

Annual
average 26.9 26.8 26.6


29 ,


1 2 3 4 5


6 7 8 9 10 Months


Figure 2-2: Average monthly temperature ('C), at the Samuel Dam region for 1989,1990, and 1991.


'I


a
0

l
4


28.5 28

27.6 27 4 26.6 26

25.6

26 24.5


24


-0-1989
-E9-1990
--1991


11 12





17


Total annual precipitation in the Samuel region ranges from 2,000 to 2,600 mm year. June, July, and August are characteristically dry, and the dry season lasts from May to September. The rainy season is also well defined by the months of October to April, when monthly precipitation usually exceeds 200 mm (Table 2-2, Fig. 2-3).











Table 2-2: Total monthly precipitation
(mm) in the Samuel Dam, during 1989, 1990, and 1991 (source Eletronorte).


Month Year
1989 1990 1991

January 302.4 400.0 326.9
February 319.4 521.0 223.1 March 248.0 341.0 412.9
April 316.6 147.0 315.1
May 165.8 133.0 100.4
June 107.1 47.0 38.7
July 59.2 24.0 0.0
August 59.9 39.0 0.0
September 83.7 146.3 42.2
October 136.0 274.4 142.6
November 89.9 318.0 233.2
December 275.0 244.0 306.9

Annual
total 2,163.0 2,634.7 2,142.0





18


550

600 -0 1989
--9-1990
450 1991

400

360
0
.H 300
.4J 0

200



100 60 0
1 2 3 4 5 6 7 8 9 10 11 12 Months


Figure 2-3: Total monthly Precipitation (mm),
at the Samuel Dam site for 1989,1990, and 1991.




Nutrients in Soil


Surface soils were extremely acid in all plots. The soils were less acid at progressively lower depths. Amounts of phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), and aluminum (Al) decreased from surface to lower depths. Levels of P, Ca, and Mg were low, whereas levels of K ranged from low to medium. The levels of Al, however, were high in all plots and at all depths (Tables 2-3, and 2-4 ). Among sites, levels of Al were different at all depths (ANOVA, n = 2, d.f. = 1, p < 0.0264, 0.0108, and 0.0100, for 0-20, 20-40, and 40-60 cm depth, respectively), and pH was different at 0-20, and 20-40 cm depth (ANOVA, n = 2, d.f. = 1, p < 0.0450, and 0.0239, respectively). Levels of P, K, Ca, and Mg did not differ among sites.













Table 2-3: Average pH and nutrients per area, per category of depth


at the Samuel Ecological Station ( s.d.). Depth P and K (ppm), Ca, Mg, and Al (mEq/100 ml).


(cm), pH (H20)


AREA DEPTH pH P K Ca Mg Al


01 0-20 3.85 (0.2) 2.55 (0.8) 47.73 (18.1) 0.20 (0.1) 0.16 (0.1) 2.07 (0.4)
20-40 4.30 (0.4) 1.27 (0.5) 27.45 (22.9) 0.07 (0.1) 0.13 (0.1) 1.46 (0.3)
40-60 4.48 (0.2) 1.36 (0.7) 14.00 (22.3) 0.05 (0.1) 0.14 (0.1) 1.15 (0.2)

02 0-20 3.78 (0.1) 1.73 (0.6) 52.82 (26.7) 0.15 (0.1) 0.18 (0.1) 1.83 (0.3)
20-40 4.27 (0.3) 1.00 (0.0) 40.00 (35.6) 0.12 (0.1) 0.22 (0.1) 1.28 (0.4)
40-60 4.40 (0.2) 1.27 (0.5) 18.91 (12.2) 0.11 (0.0) 0.17 (0.1) 1.19 (0.2)

03 0-20 4.08 (0.2) 1.36 (0.5) 43.36 (16.9) 0.75 (0.4) 0.52 (0.2) 0.92 (0.2)
20-40 4.33 (0.3) 1.00 (0.0) 26.91 (20.9) 0.20 (0.1) 0.29 (0.1) 0.79 (0.2)
40-60 4.52 (0.4) 1.09 (0.3) 35.36 (54.0) 0.19 (0.1) 0.30 (0.1) 0.68 (0.3)


Note: Reference values for pH and nutrients (Source: EMBRAPA's Laboratory of Soil Analysis).


pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid


Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high


P: 0 10 ppm = low
11 30 ppm= medium
> 30 ppm = high


Al: 0.0 0.3 mEq = low
> 0.3 mEq =high


K: 0 45 ppm = low
46 150 ppm = medium
> 150 ppm = high


I-, '.0












Table 2-4: Average pH at Jusante ( s.d.). and Al (mEq/100 ml).


and nutrients per area, per category of depth Depth (cm), pH (H20) P and K (ppm), Ca, Mg,


AREA DEPTH pH P K Ca Mg IL


A 0-20 3.45 (0.1) 3.27 (1.4) 49.18 (14.3) 0.13 (0.1) 0.17 (0.1) 3.12 (0.4)
20-40 3.73 (0.2) 1.36 (0.5) 14.73 ( 6.8) 0.04 (0.1) 0.12 (0.1) 2.79 (0.3)
40-60 4.05 (0.2) 1.09 (0.3) 7.73 ( 2.6) 0.00 (0.0) 0.11 (0.1) 2.58 (0.1)

B 0-20 3.51 (0.2) 2.82 (1.3) 37.54 ( 8.2) 0.09 (0.1) 0.15 (0.1) 2.65 (0.8)
20-40 3.86 (0.2) 1.36 (0.7) 11.27 ( 2.3) 0.05 (0.1) 0.12 (0.1) 2.12 (0.3)
40-60 4.15 (0.3) 1.73 (1.3) 10.09 ( 7.3) 0.02 (0.1) 0.14 (0.1) 1.90 (0.2)

C 0-20 3.71 (0.3) 5.18 (2.7) 63.45 (28.4) 0.02 (0.0) 0.18 (0.1) 2.92 (1.0)
20-40 4.11 (0.2) 2.64 (3.5) 29.91 (25.5) 0.01 (0.0) 0.11 (0.1) 2.41 (0.9)
40-60 4.43 (0.2) 1.27 (0.6) 13.09 ( 8.1) 0.01 (0.0) 0.11 (0.1) 2.13 (0.9)


Note: Reference values for pH and nutrients (Source: EMBRAPA's Laboratory of Soil Analysis).


pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid


Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high


P: 0 10 ppm = low
11 30 ppm = medium
> 30 ppm = high


Al: 0.0 0.3 mEq =low
> 0.3 mEq = high


K: 0 45 ppm = low
46 150 ppm = medium
> 150 ppm = high





21


In summary, both Jusante and Reserve sites had low pH values and low levels of nutrients in the soil; however, the soil at Jusante had a lower pH and a higher Al content than at the Reserve indicating more acid soils were characteristic of Jusante. Vegetation


Canopy

Average DBH of trees was 22.3 cm (SD = 14.8, n = 612) and 18.9 cm (SD = 10, n = 612) at the Reserve and Jusante, respectively; 20.6% of trees at the Reserve had diameters than 30 cm, compared to only 11.8% at Jusante (Fig. 2-4). DBH differed between the 2 sites (Chi-Square, p <

0.006, n = 1224). The findings from the Reserve are in agreement with those of Martinelli et al. (1988), who reported a mean DBH of 21.2 cm with 16.6% of the trees with DBH > 30 cm from a survey prior to the filling of the Reservoir.

Total basal area was larger at the Reserve than at Jusante; 24.7 m2, and 18.8 M2, respectively (n = 612 in both areas). The difference was greater if only trees > 30 cm in DBH were considered; 16.7 M2 (n = 126) and 8.6 M2 (n = 72) in the Reserve and at Jusante, respectively.

The height of trees also differed between sites (Chi-Square, p <

0.0001, n = 1224). Despite the fact that average tree height at the Reserve was only slightly greater than at Jusante, 13 m (SD = 4.2, n = 612), and 12 m (SD = 3.1, n = 612), respectively, 5.4% of trees at the Reserve were ; 21 m compared to 1.47% of trees at Jusante (Fig. 2-5). Tree height distribution at the Reserve was also in conformity with findings from Martinelli et al. (1988), with trees predominantly falling











40 35 30 25 20 15 10 5


0







45 40

35 30 25

20 15 10

5

0


22













Reserve















0 10 15 20 25 30 35 40 45 50 55 60 >65

DBH (cm)











Jusante













~AiFIF-iF--


0 10 15 20 25 30 35 40 45 50 55 60 >65

DBH (cm)



Figure 2-4: Percentage of trees in each DBH (cm) category at the Samuel Reserve and Jusante.






23


25 20



16 10


. I I I I I I I I I


<6 6 7 9 11 13 15 17

Height (m)


19 21 23


<5 6 7


Jusante


S I I I


9 11 13 15 17 19 21 23 >26


Height (m)





Figure 2-5: Percentage of trees in each height (m) category at the Samuel Reserve and Jusante.


Reserve


4
44)
0 4.) aD do
14 @4


6 -


0


m-- -


>26


30 26



20


1 10


4)
14
40 4of


62-


. n. r--I. '.


.1-7 .1 1 .





24


in the categories between 10-15 m. However, average tree height found by Martinelli et al. (17.8 m) was greater than in this study. The difference may be due to different techniques of measuring tree height; I estimated height visually, whereas Martinelli et al. used a clinometer and a measuring tape. However, because both the Reserve and Jusante were sampled similarly in this study, comparisons between categories of height between sites should be valid.

Total tree density at the Reserve was 441.3 trees/ha (SD = 17.9, n = 612), and 523.6 trees/ha (SD = 22.4, n = 612) at Jusante. However, considering only trees with DBH 30 cm, the density was higher at the Reserve (90.9/ha; SD = 34.7, n = 126) than at Jusante (61.6/ha; SD = 30.1, n = 72). Martinelli et al. (1988) reported similar findings for the Reserve, total density was 483 trees/ha, and 80/ha for trees

30 cm.

Understory

Average DBH for trees < 10 cm in both areas was very similar, 3.3 cm at the Reserve and 3.4 cm at Jusante, and did not differ in their distribution (Chi-square, d.f. = 8, p = 0.172, n = 573 and 798, respectively) (Fig. 2-6). Similarly, average understory tree height did not differ between sites; being 4.6 m and 5.1 m for the Reserve and Jusante, respectively (Chi-Square, d.f. = 7, p = 0.096, n = 573 and 798) (Fig. 2-7). Seedling density was almost identical at the 2 sites, 15,015 individuals/ha (SD = 7,762, n = 78) at the Reserve, and 15,385 individuals/ha (SD = 7,953, n = 78) at Jusante. However, for trees with DBH between 1 and 10 cm, density was 28% lower at the Reserve (2,933 individuals/ha; SD = 1,416, n = 78) than at Jusante (4,092






25


6


0


30


26 20


16 10


6 -


0


30


26 20


16 10


7


nrlnn-- F


1 2 3 4 6 6 7 8 9 DBH (cm)



Figure 2-6: Percentage of trees in each DBH (cm) category sampled in the understory vegetation at the Reserve and at Jusante.


44

0
4.)


Reserve
















1 2 3 4 5 6 7 8 9

DBH (cm)














-F1 Jusante


4)

'44
0
4.) 4) W)





26


46 -r


Im


Reserve


H1 m--


<2 2 4 6 8 10 12 14

Height (m)


Jusante


<2 2 4 6 8 10 12 14

Height (m)



Figure 2-7: Percentage of trees in each height (m) category sampled in the understory vegetation at the Reserve and at Jusante.


40


36 --


30 -


In @3

0


0 4.)


25 -4-


20 16


10 +


6-


0


36 30 25 20 16 10


0 @3


6


0


lI r I I I I I I I I I






27


individuals/ha; SD = 1,975, n = 78). Basal area was also lower at the Reserve than at Jusante (3.4 m2 and 5.2 M2, respectively).

A striking difference in understory structure between the 2 sites was the number of stemless palms. At the Reserve, the number of individual stemless palms was 234, 65% of which belonged to the species Orbignya barbosiana (babagu) which is of great economic importance in the region as a source of oil. At Jusante, however, only 57 individual stemless palms were counted, and only 8.8% of those were 0. barbosiana. On the other hand, the palm-like species Phanakospermum guianenses (sororoca, Musaceae) was recorded 58 times in the Reserve, but 469 times at Jusante.


Discussion and Conclusions


The Amazon forest is heterogeneous both in the large numbers of species within each community type, and in the large numbers of community types in a given area. Structurally, the Reserve and Jusante are different. DBH and tree height distribution revealed that the Reserve has a higher and denser canopy than Jusante. Basal area and tree density, for trees 2 than 30 cm, were also much greater at the Reserve.

According to Pires and Prance,

similar types of vegetation have approximately the same biomass.
Biomass can be expressed by the basal area of trees per hectare,
using individuals of 30 cm or more in circumference [approximately 10 cm in diameter]. On this basis, the exceptionally large forests
can exceed 40 M2 of basal area. The open forests or vine forests
usually are between 18 and 24 M2. (Pires and Prance 1985, pg. 112) Vine forests in the MarabA region, on the Tocantins river, have basal areas between 18-22 m2 (Pires 1984). In this study, basal area was 18.8 m2 at Jusante and 24.7 M2 at the Reserve, indicating that the forest






28


type at the Samuel Dam region is the open or vine forest, commonly found in Rond6nia (Pires 1984). Furthermore, species most common in open forests, according to Pires (1984), were described by Martinelli et al. (1988) at the Reserve. The absence of epiphytes in the area, another characteristic of open forest, can be related to the occurrence of a marked dry season.

The lower basal area at Jusante may be due to the age of the forest. According to Lisboa (1990), younger vegetational types are located in areas that were submerged in the past. Elevation varies between 60 and 100 m at Jusante, and between 90 and 150 m at the Reserve. This difference in elevation supports the idea that the forest at Jusante is younger, probably due to disturbance effects related to its proximity to the river.

The more open forest at Jusante allows for greater penetration of light, providing an opportunity for shrub and liana species to develop, which creates a forest floor more densely covered by vegetation. The denser vegetation is exemplified by the higher density of understory trees in the 1-9 cm DBH category at Jusante, and the abundance of the low growth musaceous palm-like species, Phanakospermum guianense, which was encountered 88% more on the forest floor of Jusante than at the Reserve.

Even though this study did not record species composition,

differences in the study sites were visually detected. The presence of adult Bertholletia excelsia (Brazil nut), and Orbignya barbosiana (babagu) were very.common at the Reserve, but rare at Jusante. On the other hand, the species, Hevea brasiliensis (seringueira), was very common at Jusante but never seen at the Reserve; the latter observation





29


was expected because this particular species is known to occur along water courses and not in terra firme.

The reason that the soils are so low in nutrients is because they have been subjected to the intense weathering of the tropical climate for many millions of years. However, a very slight difference in soil quality within an area could be reflected in an entirely different forest community through very fine adaptations of each community to subtle differences in the soil (Jordan 1985). The latter could explain the differences in species composition between the study sites.

















CHAPTER 3
INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A CONSEQUENCE OF DAMMING


Introduction


The importance of tropical rain forests to global biodiversity is clearly appreciated when one realizes that they cover only 7% of the earth's land surface, but contain more than half the species of the world's biota (Wilson 1988). Despite the importance of tropical forests, and the fact that very little is known about their fauna and flora, development of tropical areas is occurring at a rapid pace, and will bring about the extinction of species. To avoid mass extinction and to be able to guide developing agencies, a better understanding of the communities and their responses to environmental changes is needed. The

purpose of this chapter is to document the response of mammalian communities to environmental changes resulting from the construction of the Samuel Hydroelectric Dam in the Amazon.

Two sites were monitored after the filling of the Samuel

Hydroelectric Dam reservoir. In the first site (a reserve), animals captured during a rescue operation were released. This site was monitored during 1988 (before the flooding of the reservoir), 1989, 1990, and 1991. The second site (referred to as Jusante) was an undisturbed area located downstream from the dam, on the right bank of the Jamari river, and it was monitored during the 1990 and 1991 field seasons.


30





31


My assumption at the beginning of this study was that the

mammalian community in the Reserve could have been affected in three ways: 1) by the release of rescued animals, 2) by the migration of animals fleeing from the flooded reservoir (I have used the term migration as defined by Baker 1977 throughout the text), and 3) by a combination of both. During the rescue operation, from November 1988 to March 1989, 2,374 mammals were released inside the Reserve (Eletronorte 1989). In addition to the release of rescued animals in the Reserve, I expected migration of animals from the reservoir to occur, because the forest at the reservoir was continuous with the forest at the Reserve. I hypothesized that the Reserve would experience animal overcrowding for an undetermined length of time, possibly surpassing the carrying capacity for the area. My hypothesis could be tested by estimating mammalian densities in the Reserve at different points in time and by examining differences in biomass values for the community.

The Jusante site, on the other hand, could only have been affected by the migration of animals from the reservoir area (because there was no release of animals in the area), or not affected at all.

If my assumptions were correct, the noted responses in density changes, regardless of site, would be immediate in the case of terrestrial mammals (because they would have to flee from the rising water), but possibly delayed for arboreal species (because they could stay on top of trees while the vegetation was still alive). The time frame in which density changes would occur was unknown. To increase the probability of documenting such changes (completely or partially), the sites were sampled repeatedly.





32


Methods



Study Site


The study was conducted in the Samuel Dam region located on the Jamari river, in the state of Rondonia, approximately 50 km east of the state's capital of Porto Velho (Fig. 1-1).

Prior to the filling of the reservoir, Eletronorte (the government agency responsible for hydroelectric dam constructions in northern Brazil) created a 21,000 ha Reserve (Estagio Ecol6gica de Samuel) to compensate for the loss of 56,000 ha of forest due to the creation of the reservoir. The Reserve is located east of the reservoir's embankment, approximately 26 km from the dam. The other study site (referred to as Jusante) was located 3 km below the dam, on the right bank of the Jamari river (Fig. 2-1). Data Collection


Five plots of 1 km2 were established in the Reserve in 1989, while three plots were established at Jusante in 1990 (Fig. 2-1), creating 4 km of transect lines along each plot's perimeter. Transect surveys were conducted by walking slowly (1 km/h), and stopping periodically to watch and listen. Transects were conducted between 0630-0700 hours (depending on the time of sunrise) and 1030-1100 hours in the morning, and between 1300 to 1700 hours in the afternoon. The number of transect samples in each area was divided equally between morning and afternoon surveys. Whenever possible, different transects were walked in the morning and afternoon of the same day. If that was not, possible due to logistics, the two daily surveys of a plot always began in the same direction to





33


give an interval of six hours between the morning and the afternoon survey (i.e. the same point in the trail would be traversed in the afternoon six hours after the morning survey). Transect surveys on different days began at opposite ends of the route to reduce potential biases resulting from direction of travel by the observer. Each transect had equal numbers of surveys originating in both directions. The time, transect identification, location on the trail, species, number of individuals sighted, angle of sighting, and distance from the observer to the animal when first seen was recorded for every non-flying mammal encountered. Surveys were conducted by myself, an assistant, and two field helpers.

The study was conducted from May to August 1989 and from May to October of 1990 and 1991. These months correspond to the dry season in the region (refer to Chapter 2 for detail). During 1989, due to logistical problems and to the lesser amount of time spent at the study site, data were gathered only at the Samuel Reserve.

Two of the five plots sampled in 1989 in the Reserve were abandoned in 1990 and 1991 due to logistic difficulties. However, observations collected in these areas were included in the 1989 analysis to arrive at a density estimate for the entire site. Data from the various plots at a site were pooled within years to give an overall density estimate. The number of kilometers walked per site, plot, and year are listed on Table 3-1.


Data Analysis


Data were analyzed using the computer program TransAn, version

1.00. TransAn is a flexible computer program that uses a non-parametric,






34


shape-restricted, density estimator (Payne 1992). The shape-restricted estimator, first introduced by Johnson and Routledge (1985) and later modified by Fyfe and Routledge (1991), involves modeling the probability of detecting an individual as a function of its perpendicular distance from the transect line.

TransAn requires sightings from at least four independent transect lines to calculate confidence limits. Because I had only three transects per site, the data were divided between morning and afternoon transects to increase the number of transects to six (for the 1989 data the total number of transects was 10, because there were five different plots at the Reserve). Despite inherent biases, the transect censuses are currently the most cost effective method to evaluate large mammal densities in rainforests (Emmons 1984).







Table 3-1: Total number of kilometers walked during surveys, per site, plot, and year.


1M4 WALKED PER YEAR
SITE PLOT 1989 1990 1991 TOTAL
Reserve 1 48 88 84
2 48 96 84
3 48 80 84
4 36 -- -5 36 -- -Total 216 264 252 732
Jusante A -- 80 84
B -- 80 84
C -- 80 84
Total -- 240 252 492
GRAND TOTAL 1,224





35


Biomass estimation

Crude biomass was calculated using average adult body weight (BW) and population density (D) of a species: BW*D (kg km). The average animal weight used for biomass calculations was obtained from data gathered by biologists working at the rescue operation during 14 November 1988 to 29 March 1989 (Eletronorte 1989). Because the rescue operation was conducted as the reservoir was being flooded, and because flooding was accomplished within four months (a relatively short period of time), I assumed that the average animal weight recorded reflected the true weight of the animals in the wild before any stressful situation could cause weight loss.

Metabolic biomass also was calculated. It is an important measurement because species sampled varied greatly in size, and metabolic biomass takes into account that energy expenditure increases allometrically with body weight to the power of 0.75 (Peters 1983). It was then calculated as BW* 5,D (kg0'75 Km'). The density, biomass, and body weight estimates were ln-transformed to linearize the data.


Results



Sightings per Kilometer Walked


The number of individuals sighted per km in the Reserve was very similar for all three years. The mean number of individuals sighted per km walked in the Reserve was 3.34 (SD = 0.49; n = 269 sightings) in 1989, 3.59 (SD = 0.47; n = 304 sightings) in 1990, and 3.34 (SD = 0.39; n = 296 sightings) in 1991 (Fig. 3-1). At Jusante, the mean number of individuals sighted per km walked was 43% lower in 1991 than in 1990:





36


280 270 260 250 240 230 220 210 200 -


1990 1991


Jusante


1990 1991


Year

-)K-kms walked -0--ind/km


Figure 3-1 :Number of kilometers walked and number of individual animals seen per kilometer walked at the Reserve and at Jusante
during the study.




3.90 ( 1.80) ind/km in 1990, and 2.24 ( 0.72) ind/km in 1991 (n = 272 and 224 sightings, respectively; Fig. 3-1).

To evaluate bias in the sampling method, all sightings were

plotted according to the location on the trail where species were seen. The result was an even distribution of sightings and species in each area and in all three years. Therefore, there was no observer bias as to where the animals were observed along the transects. There also was no difference in the number of observations between morning and afternoon censuses, indicating that animal sightings were independent of time of day (Chi-square test = 6.85, 4 d.f., p= 0.05).


UI)
4

0


Reserve


.5

- 4.5

-4

- 3.5

.3

- 2.5

-2

1.5

1


1989





37


Species Observed


The number of species recorded at the Reserve during census was similar for all years. Sixteen species were recorded in 1989, 18 in 1990, and 17 in 1991. At Jusante the number of species seen was 19, in Table 3-2: List of species observed during all transect surveys for both sites. NR = number of sightings for 1989, 1990, and 1991 at the Reserve NJ = number of sightings for 1990 and 1991 at Jusante. ORDER FAMILY SPECIES IM NJ


Primata


Edentata Carnivora Perissodactyla Artiodactyla Rodentia


Cebidae


Callithrichidae Myrmecophagidae Bradypodidae Dasypodidae Procyonidae Mustelidae Felidae Tapiridae Tayassuidae Cervidae Sciuridae Dasyproctidae


Aotus azarae Callicebus bruneus Pithecia irrorata Cebus apella Saimiri ustus Ateles paniscus Callithrix emiliae Saguinus fuscicollis Tamandua tetradactyla Myrmecophaga tridactyla

Choloepus didactylus Cabassous sp.

Nasua nasua

Eira barbara Lutra longicaudis Felis pardalis Panthera onca Tapirus terrestris Tayassu tajacu Tayassu pecani Mazama americana Mazama guazoubira Sciurus sp. Dasyprocta fuliginosa


01 29 62
197 25
165

34 58

02 01 00

00 10 11
00 01 01 03 11 01 67


01 98 37 59
24 05 27 62 05 00 03

02 12 01 01 01 00

02 03 00

49


10 15 180 89






38


1990 and 15 in 1991; a 16% decrease. Table 3-2 lists all species seen during transect surveys in both sites during all three years. The list of species seen is only a fraction of the total number of mammalian species in the area and only represents the medium to large size mammalian community (refer to Appendix A for a complete species list).

Of the 24 species seen during census, only 7 primate species, 2 deer, and the agouti had sample sizes large enough to estimate density. Primates


Primate density estimates prior to damming

Primates censuses were performed at the Reserve by Eletronorte

researchers from September 1987 to February 1988 (Table 3-3, and Figure 3-2). Density estimates were based on 145 km of transect surveys. Techniques used were comparable to the ones used in this study (National Research Council 1981).


Table 3-3: Primate density estimates at the Reserve prior
to the flooding of the reservoir. D = group density,
MGS = mean group size (Eletronorte 1988).Saimiri were
not recorded during the 1988 censuses.


SPECIES D MGS IND/1XM2

Ateles paniscus 2.60 5.2 13.5

Cebus apella 4.02 5.4 21.7

Pithecia irrorata 1.38 3.8 5.2

Callicebus bruneus 0.26 2.6 0.7

Saimiri ustus ---- ---- ---Saguinus fuscicollis 2.00 8.5 17.0

Callithrix emiliae 0.36 15.0 5.4






39


Individuals captured and released

Primates represented 48.4% of all mammals captured, and 47% of all animals released during the rescue/release operation (Eletronorte 1989). The number of individuals per species captured during the operation, and later released at the Reserve is shown in Table 3-4 (refer to Chapter 1 for details of operation). This was a relatively small number and probably had an insignificant effect on most density shifts during the years following flooding. Many changes in density derived form emigration (see below).


Table 3-4: Number of primates captured at the Samuel
Dam reservoir, and number of primates released at the
Samuel Ecological Station (Eletronorte 1989).


Primate density


estimates after damming


The densities of Ateles, Callithrix, and Saimiri were high in 1990 and lower and approximately equal in 1989 and 1991. Cebus and Saguinus densities were also high in 1990 in the Reserve, however, their


SPECIES NUMBERS NUMBERS
CAPTURED RELEASED

Aotus 104 60

Ateles 35 35

Call th.rix 71 42

Saguinus 171 76

Cebus 207 180

Callicebus 348 309

Plthecia 369 324

Saimiri 501 326

TOTAL 1,806 1,352





40


densities in 1991 remained high and similar to the 1990 densities instead of returning to values comparable to 1989. The densities of Callicebus and Pithecia in the Reserve were at their highest in 1989, and decreased steadily through 1990 and 1991(Table 3-5, Fig. 3-2).

With the exception of Pithecia, whose density was similar in 1990 and 1991, all other primate densities decreased substantially from 1990 to 1991 at the Jusante site (Table 3-5, Fig. 3-2).

Ateles paniscus. In 1989, the density of Ateles in the Reserve was

3.15 groups/km2 (13.2 individuals/km2), whereas in 1990 their density increased to 6.06 groups/km2 (23.6 ind/km2). By 1991 their group density was at 3.69 per km2 (higher than in 1989). However, the number of individuals per km2 was 11.1 (less than in 1989) due to the fact that the mean group size went from 4.2 in 1989 to 3.0 in 1991 (Table 3-5, Fig. 3-2). At Jusante the density of Ateles was estimated at 0.60 groups/km2 (5.9 ind/km2) in 1990. In 1991 only one individual was sighted during 252 km of census, making it impossible to estimate density (Table 3-5, Fig. 3-2).

Callithrix emiliae. The density of Callithrix in the Reserve more than doubled from 1989 to 1990, 1.41 groups/km2 (4.1 ind/km2) to 3.06 groups/km2 (10.1 ind/km2), respectively. By 1991, however, the density had decreased to 1.82 groups/km2 (3.6 ind/km2), similar to 1989 (Table 3-5, Fig. 3-2). The density of Callithrix at Jusante decreased by 71% from 1990 to 1991; from 3.13 groups/km2 (9.1 ind/km2) in 1990, to 0.91 groups/km2 (2.8 ind/km2) in 1991 (Table 3-5, Fig. 3-2).

Saimiri ustus. The density of Saimiri in the Reserve increased four fold from 1989 to 1990, then decreased by more than half by 1991. The density in 1989 was 0.91 groups/km2 (6.9 ind/km2), by 1990 it was





41


3.68 groups/km2 (28.7 ind/kM2), and by 1991 it had decreased to 1.41 groups/km2 (8.9 ind/km2) (Table 3-5, Fig. 3-2). The density of Saimiri at Jusante decreased by four fold if we consider the number of individuals per km2. In 1990 the density was 1.69 groups/km2 (23.8 ind/km2), and by 1991 it was 0.75 groups/km2 (6.0 ind/kM2) (Table 3-5, Fig. 3-2).

Cebus apella. Cebus density in the Reserve went from 5.42

groups/km2 (20.1 ind/kM2) in 1989 to 6.63 groups/km2 (28.5 ind/km2) in 1990. The density in 1991 was similar to 1990 with 6.45 groups/km2 (27.1 ind/km2) (Table 3-5, Fig. 3-2). At Jusante Cebus density decreased from 1990 to 1991, going from 2.92 groups/km2 to 2.12 groups/km2. The difference is greater if we consider the number of individuals per km2: 17.5 in 1990, and 8.1 in 1991. This difference is due to the fact that mean group size went from 6.0 in 1990 to 3.8 in 1991 (Table 3-5, Fig. 3-2).

Saguinus fuscicollis. The density of Saguinus in the Reserve more than doubled from 1989 to 1990, going from 1.85 groups/km2 (6.1 ind/km2) in 1989 to 4.08 groups/km2 (14.3 ind/kM2) in 1990. Saguinus maintained a high density in 1991: 3.7 groups/km2 (14.1 ind/km2 ) (Table 3-5, Fig. 32). At Jusante there was a decrease in Saguinus density from 1990 to 1991: 5.83 groups/km2 (20.4 ind/km2) in 1990, to 3.67 groups/km2 (13.6 ind/km2) in 1991 (Table 3-5, Fig. 3-2).

Callicebus bruneus. The density of Callicebus in the Reserve was at its highest in 1989 with 3.61 groups/km2 (7.2 ind/km2). In 1990 the density had decreased to 0.83 groups/km2 (1.3 ind/km2), and by 1991 it was 0.62 groups/km (1.2 ind/km2) (Table 3-5, Fig. 3-2). At Jusante the density of Callicebus went from 8.33 groups/km2 (20.0 ind/km2) in 1990, to 4.85 groups/km2 (11.6 ind/km2) in 1991 (Table 3-5, Fig. 3-2).






42


Pithecia irrorata. Pithecia also had its highest density in the Reserve in 1989, and then decreased thereafter. The density was 3.43 groups/km2 (10.3 ind/km2)in 1989, 2.07 groups/km2 (5.2 ind/kM2) in 1990, and 1.17 groups/km2 (3.5 ind/kM2) in 1991(Table 3-5, Fig. 3-2). At Jusante the density of Pithecia remained similar from 1990 to 1991: 2.60 groups/km2 (7.5 ind/km2), and 2.38 groups/km2 (7.6 ind/km2), respectively (Table 3-5, Fig. 3-2). Density and body weight of primates

When broad geographic regions are examined, differences in density among primate species are related to body size. In general, population densities decline with increasing body mass (Clutton-Brock and Harvey 1977, 1979; Eisenberg 1979; Robinson and Redford 1986; Kinnaird and Eisenberg 1989). In the primate community at the Samuel Dam region, the correlation between density and body weight was positive for primates in the Reserve in 1989. This positive correlation is a contradiction of the rule, because smaller animals should have higher densities than larger ones. There was no significant correlation between density and body weight in 1990 and 1991 (Fig. 3-3).

The correlation between density and body weight for primates at Jusante was not significant in either year, but showed a tendency to be negative in 1990, (Fig. 3-4).

Total primate density estimates

Primates at the Reserve comprised 61% of all mammalian sightings in 1989, 69% in 1990, and 68% in 1991. When data for all primates are pooled and density estimates were calculated for the area as a whole, the result shows a 32% increase in group density from 1989 to 1990, and a 23% decrease from 1990 to 1991 (Table 3-6, Fig. 3-5A).






43


Table 3-5: Primate density estimates at the Reserve and at Jusante. N = number of sightings, D = group density, MGS = mean group size.

SPECIE SITE YEAR N D MGS IND/EM 95* D CI

Ateles Reserve 89 27 3.15 4.2 13.2 1.53-07.43
paniscus Reserve 90 72 6.06 3.9 23.6 3.95-09.91
Reserve 91 66 3.69 3.0 11.1 2.14-07.06

Jusante 90 04 0.60 9.8 5.9 0.17-02.63
Jusante 91 01 ---- 1.0 ---- ---- ---Cebus Reserve 89 54 5.42 3.7 20.1 3.21-09.22
apella Reserve 90 71 6.63 4.3 28.5 4.15-10.31
Reserve 91 72 6.45 4.2 27.1 4.01-10.32

Jusante 90 32 2.92 6.0 17.5 1.59-06.15
Jusante 91 27 2.12 3.8 8.1 0.98-04.57

Pithecia Reserve 89 33 3.43 3.0 10.3 1.76-06.59
irrorata Reserve 90 17 2.07 2.5 5.2 0.75-05.06
Reserve 91 12 1.17 3.0 3.5 0.35-02.49

Jusante 90 21 2.60 2.9 7.5 1.16-05.57
Jusante 91 16 2.38 3.2 7.6 1.08-06.39

Callicebus Reserve 89 15 3.61 2.0 7.2 1.57-08.34
bruneus Reserve 90 05 0.83 1.6 1.3 0.20-03.03
Reserve 91 09 0.62 2.0 1.2 0.24-01.74

Jusante 90 59 8.33 2.4 20.0 4.77-14.72
Jusante 91 39 4.85 2.4 11.6 2.49-09.89

Saimiri Reserve 89 08 0.91 7.6 6.9 0.25-02.43
ustus Reserve 90 05 3.68 7.8 28.7 0.89-05.68
Reserve 91 12 1.41 6.3 8.9 0.49-03.64

Jusante 90 15 1.69 14.1 23.8 0.62-04.55
Jusante 91 09 0.75 8.0 6.0 0.29-02.58

Saguinus Reserve 89 14 1.85 3.3 6.1 0.71-03.88
fuscicollis Reserve 90 24 4.08 3.5 14.3 1.92-09.18
Reserve 91 20 3.70 3.8 14.1 1.51-06.53

Jusante 90 34 5.83 3.5 20.4 2.88-11.23
Jusante 91 28 3.67 3.7 13.6 1.74-07.79

Callithrix Reserve 89 10 1.41 2.9 4.1 0.38-03.82
emiliae Reserve 90 15 3.06 3.3 10.1 1.31-06.69
Reserve 91 09 1.82 2.0 3.6 0.74-06.00

Jusante 90 20 3.13 2.9 9.1 1.34-05.97
Jusante 91 07 0.91 3.1 2.8 0.28-03.75





44


Ateles


7
6
5
4 3
2

- 0


M M 0 MDa 00-


Cebus


---v


I


7 ~ 6
5
4


2
1 0-


Q-) Q/)


0




0


Callithrix


7
6
5
4-


7
6
5
4
3
2

0-


Saimiri D )0 -


Saguinus


r


=IReserve f7 dusante


o 0)0 a) a) a) a


Pithecia








-D )0 a) a3) 0a 00 00 M


Year


Figure 3-2: Primate density estimates for the Samuel Ecological Station during 1988 (Eletronorte 1989), 1989, 1990, 1991 (this study), and the Jusante in 1990, and 1991 (this study).


2


0
CO 3M 0)0
0)0)0)MaM


3D 3 a) a)o


Callicebus 7
6
5
4
3
2


3DE 0)' a) 0 a
00 00 0 M


77 6
5
4
3
2
1
0 -


9
8
7
6
5
4
3
2

0-





45


Primates at Jusante comprised 71 and 58.5% of all mammal sightings for 1990 and 1991, respectively. Primate group density for the area as a whole decreased 44% from 1990 to 1991 (Table 3-6, Fig. 3-5A). Even though a decrease in primate density occurred in both areas from 1990 to 1991, the decrease at Jusante was almost twice that of the Reserve. The changes in densities of ind/km2 show the same pattern as the group density changes, however, Jusante shows a more abrupt reduction in total number of individuals than the Reserve (Fig. 3-5B). Density changes between years at the Reserve

A cluster analysis comparing density results for all 4 years of data for the Reserve shows that the 1988 densities had only a 25.27 degree of similarity with the 1989 densities (all comparisons excluded Saimiri because this species was not recorded during 1988). The 1990 community still only shows a 27.94 degree of similarity with 1988. By 1991, the degree of similarity with the 1988 community, increased to 61.92 (Fig. 3-6).


Table 3-6: Total primate densities for the Reserve and Jusante
for 1989, 1990, and 1991. N = number of sightings, D = group
density, MGS = mean group size (based on total number of
individuals sighted and total number of sightings per year).




SITE YEAR N D MGS flD/Xf2 951 CI

Reserve 1989 161 17.59 3.6 63.3 13.86-22.81

Reserve 1990 209 23.25 3.9 90.7 18.92-29.02

Reserve 1991 200 18.03 3.6 64.9 13.71-23.68




Jusante 1990 185 24.51 4.4 107.8 17.62-33.46

Jusante 1991 127 13.83 3.5 48.4 8.86-23.14





46


0 Ateles A Cebus Pithecia U Saimiri A Callicebus 0 Saguinus V Callithrix


1.' 1.: 1 0.

0.

0.


4


- 1989


A


2 -e
0 8 -o 6 -V

4 '
2.0 2.5 3.0 3.5 4.0


1.6
1.4 1.2 1.0 0.8 0.6
0.4 0.2 0.0


- 1990
-l


A


K>
V


4


I A I


1 o 1.1
1 .
1 .

0.
0.

0.


2.0 2.5 3.0 3.5 4.0


3

2

3


- 1991


A


0


El


-) V V
1.

I A l I i
2.0 2.5 3.0 3.5 4.0


LogiC body weight (g)


Figure 3-3: Relationship between density and body weight for primates at the Samuel Ecological Station. (P = 0.0156 for 1989, 0.6338 for 1990, and 0.5726 for 1991).


1.6

1.4 1.2

1 .0 0.8


I


1990


F-A


A


V


0


0.6 1 1 1 1
2.0 2.5 3.0 3.5 4.0


1.6 1991
1.4
1.2 A
1.0
0.8 L
0.6
0.4 I '
2.0 2.5 3.0 3.5 4.0


LoglO body weight (g)



Figure 3-4: Relationship between density and body weight for primates at Jusante. (P = 0.2135, and
0.6974 for 1990 and 1991 respectively). See Figure
3-3 for legends.


E



(I)

-ul
0
-J


E


C
(1)


0
0






47




30

28 A
Primates
26 24 22

20

N 18

16 14 12

10
1989 1990 1991

reserve --.Jusante







120

110 Primates B

100 90 80 70 60 50

40 I I
1989 1990 1991

Reserve Jusante





Figure 3-5: A) Total primate density (groups/km2) for the Reserve, and Jusante for 1989, 1990, and 1991. B) Total primate density (ind/km2) for the Reserve, and Jusante for 1989, 1990, and 1991.





48


Figure 3-6: Cluster analysis for all four years of density data at the Reserve. Similarity levels = 61.92, 27.94, and
25.27, respectively (Minitab 10, Hierarchical cluster
analysis of observations).






Biomass

In the Reserve in 1989, the three largest primates (Ateles, Cebus, and Pithecia, respectively) were also the three species that contributed most to the crude and metabolic biomass. However, by 1990 Saimiri contributed more than Pithecia, despite weighing only a third of the weight of Pithecia. By 1991, Cebus was the species that contributed most to the metabolic biomass calculations, followed by Ateles and Saimiri. The total primate crude biomass for the three years showed a 55% increase from 1989 to 1990, and then a decrease of 40% by 1991, returning to levels similar to 1989 (Table 3-7). The total primate crude biomass for 1988 was 154 kg/km2 (Eletronorte 1989); a value very similar to the 1989 and 1991 values (Fig. 3-7).


9 im ihr it 25.27 50.19 75.09 100.00
1988 1991 1990 1989
Years





49


There was a positive correlation between body weight and biomass

(kg km-2) at the Reserve in all three years (Fig. 3-8). This correlation is well documented (Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979;

Robinson and Redford 1986).

At Jusante, Cebus contributed most to the metabolic biomass

calculations in both years. Ateles was in second place in 1990, followed

by Saimiri. By 1991, only one group of Ateles was recorded at Jusante

and the contribution by Saimiri had decreased drastically, leaving

Pithecia as the second largest contributor. Total primate crude biomass decreased by 61% from 1990 to 1991 (Table 3-8). The correlation between

biomass and body weight was also positive in both years at Jusante

(Fig. 3-9).



Table 3-7: Body weight, crude biomass, and metabolic biomass of primates
at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size.


SPECIES BODY CRUDE MbTABOLIC
WEIGHT BIOMASS BIMASS
(kg) (kg k=-2) (kg*' 7km-2)
1989 1990 1991 1989 1990 1991

Ateles 6.299(29) 83.15 148.66 69.92 52.48 93.84 44.13

Cebus 2.304(142) 46.31 65.66 62.44 37.59 53.30 50.68

PIthecia 2.102(284) 21.65 10.93 7.36 17.98 9.08 6.11

Callicebux 0.798(279) 5.75 1.04 0.96 6.08 1.10 1.01

Saimirx 0.739(277) 5.10 21.21 6.58 5.50 22.88 7.09

Saguinus 0.329(128) 2.01 4.70 4.64 2.65 6.21 6.13

Callithrix 0.318(61) 1.30 3.21 1.14 1.74 4.28 1."2
TOTAL 165.27 255.41 153.04 124.02 190.69 160.80
( 65.29) ( 109.37) ( 81.30) ( 44.75) ( 73.28) ( 56.10)






50


1988 Reserve


1989 1990

--- )K-


1991 JUsante


Figure 3-7: Total primate crude biomass for the Samuel Ecological Station during 1988 (Eletronorte 1989), 1989, 1990, and 1991 (this study), and for Jusante during 1990 and 1991 (this study).


300 250 -


tM





0


200 150 100 50




0


Primates 55.41








65.27
OW',A53.04 136A







.69


1*


I I





51


0 Ateles A Cebus Pithecia [ Saimiri A Callicebus 0 Saguinus V Callithrix


5.0 1989

4.5

4.0

3.5 3.0
2.0 2.5 3.0 3.5 4.0


6

5

4

3


- 1990

A



- A


5.0

4.5 4.0

3.5 3.0


2 1 1
2.0 2.5 3.0 3.5 4.0


-1991






- A


2.51 I
2.0 2.5 3.0 3.5 4.0


LoglO body weight (g)




Figure 3-8: Relationship between biomass and body weight for primates at the Samuel Ecological Station. (P = 0.0001,
0.0403, and 0.0289 for 1989, 1990, and 1991, respectively).


5.0 1990

4.5 -A0

4.0 /

3.5 V

3.0
2.0 2.5 3.0 3.5 4.0


4.5 1991

4.0

3.5

3.0

2.5
2.0 2.5 3.0 3.5 4.0


LoglO body weight (g)



Figure 3-9: Relationship between biomass and body
weight for primates at Jusante. (P = 0.0131,
and 0.0237 for 1990, and 1991, respectively). See
Figure 3-8 for legends.


C'j



(I)

0 ~0
0
0


E


U)


0

0
-J





52


Table 3-8: Body weight, crude biomass, and metabolic biomass of primates at Jusante during 1990, and 1991. ()= sample size.


SPECIES BODY WEIGHT CRUDE BIOMASS METABOLIC BI0ASS
(kg) (kg kni-2) (kge'"7km-2
1990 1991 1990 1991

Ateles 6.299(29) 37.16 ------ 23.46 -----Cebus 2.304(142) 40.32 18.66 32.73 15.15

Pithecia 2.102(284) 15.77 15.98 13.09 13.27

CaLlicabua 0.798(279) 15.96 9.26 16.89 9.79

Saimiri. 0.739(277) 17.59 4.43 18.97 4.78

Saguinus 0.329(128) 6.71 4.47 8.86 5.91

CaIlithrix 0.318(61) 2.89 0.89 3.85 1.19
TOTAL 136.40 53.69 117.85 50.07
( 36.91) ( 15.8) ( 28.99) (U 13.55)



Terrestrial Diurnal Mammals


Of all the diurnal terrestrial mammals that occur in the region, only the agouti and the two species of deer had sample sizes sufficient for density estimation (Tab. 3-2). Because the deer species are sometimes difficult to identify when seen for a brief moment moving in the understory, and to increase sample size, the two species were pooled (Mazama sp) for the analysis. Density estimates prior to damming

During the 1987-1988 censuses at the Reserve, the density of

Dasyprocta fuliginosa was estimated at 3.33 ind/km2, and Mazama sp. at

0.34 ind/km2 (Fig. 3-10; Eletronorte 1989). Individuals captured and released

Three hundred and three individuals of Dasyprocta fuliginosa were captured during the rescue operation; 214 of those individuals were





53


released at the Reserve. Twenty-three individual Mazama sp. were captured, and 19 were released at the Reserve (Eletronorte 1989). Density estimates after damming

Dasyprocta fuliginosa. Agouti density at the Reserve was at its

highest in 1989 with 21.7 ind/km2. It decreased thereafter, declining to 12.1 ind/km2 in 1990, and to 7.3 ind/km2 in 1991 (Table 3-9, Fig. 3-10). At Jusante, agouti density had a slight increase from 1990 to 1991: 4.3, and 5.0 ind/km2, respectively (Table 3-9, Fig. 3-10).

Mazama sp.. Deer density in the Reserve increased 40% from 1989 to 1990, then decreased 22% from 1990 to 1991, returning to levels similar to 1989. Density went from 3.3 to 4.6 to 3.6 ind/km2, respectively (Table 3-9, Fig. 3-10). At Jusante, however, there was a 74% increase in density from 1990 to 1991, going from 1.9 to 3.3 ind/km2, respectively (Table 3-9, Fig. 3-10).





20 10
18 Reserve
16 Dasyprocta 8 Jusante
E 14
( 12 6 -Mazama
i 10
0 8 4
S6
4 2
2
0 000 0) 0 ~- 00 C) 0
00 00 M) 0) 00 00 Y) 0)
0) 0) 0 a> 0> M 0) 0)


Year

Figure 3-10: Density estimates for Dasyprocta and Mazama at the Samuel Ecological Station for 1988 (Eletronorte 1989), 1989, 1990, and 1991 (this study), and at Jusante in 1990 and 1991 (this study).






54


Table 3-9 Terrestrial diurnal mammal density estimates at the Reserve and at Jusante. N = number of sightings, D = group density, MGS = mean group size.



SPECIE SITE YEAR N D MGS IND/192 95% CI

Dasyprocta Reserve 89 67 19.73 1.1 21.7 12.72-33.68
fuliginosa Reserve 90 63 12.12 1.0 12.1 7.56-21.37
Reserve 91 50 7.29 1.0 7.3 4.60-11.82

Jusante 90 42 4.31 1.0 4.3 2.36-07.27
Jusante 91 47 4.96 1.0 5.0 2.81-08.38

Mazama sp. Reserve 89 24 3.26 1.0 3.3 1.41-05.29
Reserve 90 18 4.58 1.0 4.6 1.98-09.77
Reserve 91 25 3.45 1.0 3.6 1.55-06.95

Jusante 90 16 1.92 1.0 1.9 0.76-05.37
Jusante 91 33 3.04 1.1 3.3 1.52-06.06



Total terrestrial mammal density estimates

Because only the agouti and the two deer species had sample sizes large enough to calculate densities, all terrestrial mammal sightings were pooled to arrive at a terrestrial mammal density for the Reserve as a whole. Sightings for terrestrial mammals comprised 39, 31, and 32% of all mammal sightings at the Reserve for 1989, 1990, and 1991, respectively. From 1989 to 1990, the mean group density decreased only slightly, however, from 1990 to 1991 the total terrestrial mammal density at the Reserve decreased 28%. The number of ind/km2 shows a more obvious decrease in density due to differences in mean group sizes through the years (Table 3-10, Fig. 3-11).

At Jusante the sightings for terrestrial mammals comprised 29 and 41.5% of all sightings for 1990 and 1991, respectively. Group density showed only a slight increase from 1990 to 1991 and the density of terrestrial mammals a slight decline (Table 3-10, Fig. 3-11).






55


Table 3-10: Total terrestrial mammal densities for the Reserve and Jusante for 1989, 1990, and 1991. N.SP = number of species used in calculations, N = number of sightings, D = group density, MGS = mean group size (based on total number of individuals sighted and total number of sightings per year). one sighting of 21 individuals of


T. pecari was eliminated from unrealistic increase in MGS.


40


36 30 25 20


16 10 -T-


6 -


0


the MGS calculations in order to avoid


1989 1990 1991 1989

Reserve


1990 1991

Jusante


40


- 35


- 30


- 25 20 15 10 .5


0


Figure 3-11: Total terrestrial mammal density (groups/km2 and ind/kM2) for the Reserve, and Jusante for 1989, 1990, and 1991.


SITE YEAR N. SP N D MGS =I/D4 95* CI


Reserve 1989 07 102 21.12 1.4 29.6 14.98-32.72 Reserve 1990 09 92 20.68 1.3* 26.9 14.87-31.07 Reserve 1991 09 93 14.93 1.3 19.4 10.96-21.41




Jusante 1990 08 74 8.65 1.4 12.1 5.34-14.13

Jusante 1991 08 90 8.86 1.2 10.6 5.79-12.16


Terrestrial mammals


* i I I I


1


0,-%






56


Biomass

At the Reserve, agouti biomass decreased 44% from 1989 to 1990,

and another 40% from 1990 to 1991. The biomass of the deer species had a 40% increase from 1989 to 1990, and then a 22% decrease in 1991 (Table 3-11). At Jusante, agouti biomass increased slightly from 1990 to 1991 (16%), however, the deer species had a 74% increase in biomass for the same period. (Table 3-12).

There was a 688.5% increase in Mazama and Dasyprocta biomass from 1988 to 1989 (Fig. 3-12). From 1989 to 1990 the biomass remained high, but by 1991 it had decreased 26.5%, although it was still much higher than the 1988 biomass.


Table 3-11: Body weight, crude biomass, and metabolic biomass of agouti and deer at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size, average weight for both species (adapted from Bodmer 1989).

SPECIES BODY CRUDE METABOLIC
WEIGHT (kg) BIOMASS BIOMASS
(kg km) (kg- 7* :-2)
1989 1990 1991 1989 1990 1991

Dasyprocta 2.721(278) 59.05 32.92 19.86 45.97 25.63 15.47

Mazama sp. 20.000* 66.00 92.00 72.00 31.21 43.50 34.05
TOTAL 125.05 124.92 91.86 77.18 69.13 49.52
(U 13.14) ( 5.7) (U 0) ( 10.24) ( 4.45) ( 0)



Table 3-12: Body weight, crude biomass, and metabolic biomass of
agouti and deer at Jusante during 1990, and 1991. ()= sample size,
average weight for both species (adapted from Bodmer 1989).

SPECIES BODY WEIGHT CRUDE BIOMASS MTABOLIC BICMASS
(kg) (kg k=-2) (kg' kkm-2
_ 1990 1991 1990 1991

Danyprocta 2.721(278) 11.70 13.61 9.11 10.59

Mazama op. 20.000* 38.00 66.00 17.97 31.21
TOTAL 49.70 79.61 27.08 41.80

( 1.65) ( 14.08) ( 1.29) ( 7.07)







57


140

Non-primate
120 A


100 80


60


40 20


0
1988 1989 1990 1991


Reserve -- Jusante













140


120 B


100 80


60

Mazama 40 20


0 Das.Procta

1988 1989 1990 1991


Reserve --ii i -)-- Jusante








Figure 3-12: A: Biomass for both Mazama and Dasyprocta at the Reserve and at Jusante (Eletronorte 1988, this study). B: Biomass for Mazama and Dasyprocta at Jusante plotted separately (this study).






58


Discussion and Conclusions


The results suggest that the environmental changes created by the construction of the dam altered the mammal community in the areas adjacent to the reservoir.


Primates


Density estimates

Three different patterns can be seen with primate density changes at the Reserve (Fig. 3-2). The first one, seen with Ateles, Callithrix, and Saimiri, is a large increase in density from 1989 to 1990, and then a sharp decrease in 1991, returning to density levels similar to those found in 1989. The second pattern, which involves Cebus, and Saguinus, also is an increase in density from 1989 to 1990, however, the 1991 densities remaining high. The third pattern is a very different situation, with the densities of Callicebus, and Pithecia being at their highest in 1989, and declining in 1990, and 1991.

The increase in densities from 1989 to 1990 can be explained by the migration of animals from the reservoir into the Reserve. Sixty percent of the reservoir is, on average, only 3.5 m deep, which allowed several tree species to survive for at least 8 months after the flooding began. I observed flowering and fruiting trees inside the reservoir in August 1989 (5 months after the completion of the filling of the reservoir). Child (1968) observed the same phenomenon during the formation of Lake Kariba as a result of the impoundment of the Zambezi river in Zimbabwe. At Lake Kariba, species had different survival times that varied from 4 to 12 months. He also observed that "most species standing in water came into leaf and/or remained in leaf until they






59


died" (Child 1968, pg. 37). Because primates spend most of their time in the mid to upper forest strata, it is reasonable to assume that the vast majority of the primate population was still living inside the reservoir when the 1989 survey was carried out at the Reserve. This assumption can be corroborated by the fact that only 1,806 primates were captured during the rescue operation in a 56,000 ha area (0.03 ind/ha) (Eletronorte 1989). In contrast, at the Tucurui Dam site (which has a much deeper reservoir), a total of 27,039 primates were captured in a 243,000 ha area (0.11 ind/ha) (Eletronorte 1985).

When I arrived at the Samuel Dam in May 1990 all vegetation inside the reservoir was dead. The only exception was in the higher elevation lands, which formed green islands inside the reservoir. Because the reservoir is bordered by dikes on both sides, and by a paved highway on the left river bank, a natural escape route for displaced animals was the Reserve (Fig. 1-3, and 2-1). Thus, the increase in primate densities in the Reserve was likely a result of the natural migration of animals into the Reserve between August 1989 and May 1990, caused by the loss of habitat inside the reservoir.

This scenario does not explain the third pattern of density

changes detected in Callicebus and Pithecia, whose population declined steadily after 1989 However, Callicebus, Pithecia, and Saimiri were the most frequently captured species, with over 300 individuals released (Table 3-3). This suggests that the 1989 census in the Reserve documented the density increase in these species as a result of the release operation. This argument is even more convincing when density estimates from a study done in 1988 (prior to the flooding), is examined (Fig. 3-2).






60


The densities for Ateles and Saguinus were almost identical in 1988 and 1989. Because only a few individuals were released in the Reserve, there was no reason to expect otherwise. The higher 1989 estimates for Saimiri, Callicebus, Pithecia, and to some extent Cebus, reflect the increase in density caused by the released animals. Because my study began only two months after the release was completed, it is reasonable to assume that the animals were still inside the Reserve, and that is why the 1989 densities were higher than the 1988. As for Callithrix, the difference between the 1988 and 1989 densities may be a reflection of the species characteristics. Callithrix are among the most cryptic primates, in both pelage and habitat (Ferrari and Rylands 1994), making them difficult to detect during transect samples. Because the 1988 censuses were not performed by my team, their low density might be a consequence of differences in researcher's detection ability. Despite having similar body weight, the same detectability differences do not apply to Saguinus, "while relatively similar in body size and most aspect of their ecology, S. fuscicollis invariably prefers lower forest strata them its congeners" (Ferrari and Rylands 1994, pg. 82), which makes them more visible during a census.

The decline in density estimates in 1991 for all species is most likely a consequence of the dispersal of animals to adjacent areas or death. The Reserve is located adjacent to an area of continuous forest, without human inhabitation or access roads (Fig. 1-3), and dispersal into those areas would be the expected behavior for animals in a situation of crowding.

There was a sharp decline in density (between 37 and 71%) from 1990 to 1991 at Jusante with all primates, except Pithecia (Fig. 3-2).






61


The decline in density is not very apparent for Cebus and Saimiri if we only consider the number of groups per km2, however, these two species were the only ones to show drastic reduction in mean group sizes over the years (Table 3-5). Hence, if we consider the decline in the number of individuals per kM2, both species also show drastic decreases in densities from 1990 to 1991.

Even though there was no sampling at Jusante during 1988 or 1989, it is logical to assume a similar effect on primate communities in both areas, as a consequence of the creation of the reservoir.

A buffer area for the protection of the dam turbines was created inside the reservoir by clear cutting the forest closest to the dam, which together with several construction projects and the concentration of human activities near the dam could have inhibited animal migration to Jusante. However, migration did occur, probably due to sheer proximity of the area to the reservoir (animals stranded inside the reservoir near the dam could probably see green forest on the other side) (Fig. 2-1).

Despite the lack of data for Jusante in 1989, density estimates for 6 of the 7 primates sampled at Jusante will most likely fit the first pattern of density change described for primates at the Reserve (density increase in 1990, followed by a sharp decrease in 1991). The seventh species, Pithecia, fits the second pattern of density change (density stays at similar levels from 1990 to 1991). Because no data exist for 1989, it is not possible to determined if the third pattern described at the Reserve (a continue decrease in density) was present at Jusante. However, because the explanation given for this pattern in






62


density change was the active release of animals in the Reserve, such pattern would not be expected to appear at Jusante.

Differences in crude densities and in the degree of density

decline among species and sites are most likely due to differences in species behavior and/or habitat requirements.

Ateles. Ateles are frugivore-herbivores (Eisenberg 1981, Robinson and Redford 1986, 1989), with 83 to 90% of their diet consisting of fruits and the remainder of other plant parts (van Roosmalen and Klein 1988). Because the distribution of fruits in a forest is widely scattered, Ateles density is probably restricted by the availability of this food type (Robinson and Ramirez 1982). Home-range size increases as group weight increases (Eisenberg 1979), and Ateles are the largest of all primates in the area requiring, in Surinam, 12.2 ha per individual (van Roosmalen 1980, Robinson and Janson 1987). The new arrivals at the Reserve were most likely displaced to areas outside the Reserve by the resident groups due to the unavailability of fruit crops large enough to maintain the higher population density.

Their almost complete absence from Jusante can be explained by the fact that they are restricted to, or occur in higher densities only in primary forest, using upper levels of canopy and emergent trees (Mittermeier and van Roosmalen 1981, Robinson and Ramirez 1982, van Roosmalen and Klein 1988). Because the forest at Jusante was characterized by having lower, more open canopy with fewer emergent trees (Chapter 2), it is not surprising to find that the Reserve represented a more suitable habitat for the species (Fig. 3-2).

Callithrix. Callithrix are insectivore-omnivores, with more than 50% of their diet consisting of invertebrates (Eisenberg 1981, Robinson






63


and Redford 1986, 1989). They are also adapted to feed on plant exudates at certain times of the year to compensate for seasonal scarcities in the availability of fruits (Ferrari and Lopes Ferrari 1989, Rylands and Faria 1993, Ferrari 1993). They attain highest densities in second growth forest and edge habitat.

Callithrix species tend to have larger group sizes and smaller home-ranges than Saguinus species, and generally occur at higher densities (Ferrari and Lopes Ferrari 1989, Rylands and Faria 1993). Average group size and densities of Callithrix in both of my study sites was lower than that of Saguinus (Table 3-5, Fig. 3-2), in contrast to previous studies. This may be partly a consequence of their crypticity, as described earlier.

Saimiri. Saimiri are classified as frugivore-omnivores, with more than 50% of their diet composed of fruits, and the remainder mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford 1986, 1989). They are habitat specialists, typical of flooded and riverine forests (Eisenberg 1979, Freese et al 1982, Rylands and Keuroghlian 1988). The species is known for its preference for more open, secondary habitats, and they are most often encountered in liana forests (Mittermeier and van Roosmalen 1981, Johns and Skoruppa 1987). Neither of the study sites, the Reserve or Jusante, included flooded forests. Despite the fact that the Jusante site is closer to the Jamari river, and has a more open forest structure, transect censuses started at a distance of 500 meters away from the river's edge. Therefore, high densities of Saimiri were not expected at either site. The high densities in the Reserve in 1990, as well as the very high number of individuals per km2 at Jusante (due to a larger mean group size; Table





64


3-5), probably occurred when animals living along the Jamari river inside the reservoir moved to these areas in search of new suitable habitat. Because neither azea is suitable habitat, the animals most likely dispersed along the Jamari river, causing the density decrease seen in 1991 (Fig. 3-2).

Cebus. Cebus are also classified as frugivore-omnivore, with more than 50% of their diet composed of fruits, and the remainder mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford 1986, 1989). The species has a broad habitat tolerance (Eisenberg 1979). They are opportunistic, and usually well able to persist in disturbed forest (Johns and Skoruppa 1987), which makes them the most adaptable primate species in the neotropics (Mittermeier and van Roosmalen 1981). It is not surprising then that they were able to maintain high population density in the Reserve. The reduction of 54% in the number of individuals estimated per km2 at Jusante from 1990 to 1991 (Table 3-5) seems inconsistent with their ecology. However, because Palmae species were more abundant in the Reserve than at Jusante (Chapter 2), and because Cebus apella rely heavily on palms in a number of different ways (insect foraging, fruits, seeds, flowers, and many other plant parts (Mittermeier and van Roosmalen 1981, Terborgh 1983)), it is reasonable to suggest that the Reserve holds a higher carrying capacity for the species, and hence an increased ability to maintain the increased densities.

Saguinus. The species is classified as insectivore-omnivore, with more than 50% of their diet consisting of invertebrates (Eisenberg 1981, Robinson and Redford 1986, 1989). According to Rylands and Keuroghlian (1988), optimal habitat for this species includes secondary forest and





65


forest edge mixed with tall primary forest. Several studies have demonstrated that the species occurs in greater densities in secondary forest near natural clearings than in mature forest(Eisenberg and Thorington 1973, Mittermeier and van Roosmalen 1981, Robinson and Ramirez 1982, Johns and Skoruppa 1987). Emmons (1984) concluded that Saguinus density appeared to have increased in some areas where large monkeys had been exterminated. They also overlap with Cebus in most habitat and diet categories (Mittermeier and van Roosmalen 1981). The lower densities of Ateles and Cebus at Jusante, and the increased edged habitat at the Reserve created by the reservoir, were most likely favorable factors influencing the maintenance of higher Saguinus densities at both sites during 1990 and 1991.

Callicebus. This species is also classified as frugivore-omnivore, with more than 50% of its diet composed of fruits, and the remainder mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford 1986, 1989). The species occurs in greatest densities in areas characterized by forest openings with early successional vegetation, and spends more time in the lower canopy levels and understory vegetation (Kinsey 1981, Terborgh 1983, Robinson and Redford 1986, Robinson et all. 1987). The more open vegetation at the Jusante site (Chapter 2) most likely created a more suitable habitat for Callicebus than in the Reserve, which may explain their much higher densities at Jusante (Fig. 3-2). The possibility of interference competition with Cebus might also affect Callicebus densities. Both species are catholic in their diet, however, Cebus generally have larger group size, and are more aggressive during interspecific encounters, possibly displacing Callicebus groups from feeding trees. According to Emmons (1984), troops of larger






66


monkeys, such as Cebus, physically prevent access to fruit sources by small ones, such as Saguinus and Callicebus. The lower densities of Cebus at Jusante might benefit Callicebus, affecting their population positively.

Pithecia. Pithecia are also frugivore-omnivores, with more than 50% of their diet composed of fruits, and the remainder mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford 1986, 1989). They are usually found in the understory and lower to middle parts of the canopy (Mittermeier and van Roosmalen 1981), and they occur in gallery and both primary and secondary forest (Robinson and Ramirez 1982). Pithecia are always rare (Mittermier and van Roosmalen 1981, Robinson et all. 1987, Rylands and Keuroghlian 1988), despite the fact that they have no distinct habitat preference. However, their rarity may indicate that they are specialists within the forest they occupy, or at least dependent on certain floristic communities.(Rylands and Keuroghlian 1988). According to Johns and Skoruppa (1987) Pithecia are able to feed on fruits from some of the early colonizing trees, which might explain their higher densities at Jusante.

Density changes between years at the Reserve

The low degree of similarity between the 1988 and 1989 densities was most likely due to the increased densities of Pithecia, and Callicebus, as a result of both the release of captured animals and the movement of free ranging individuals into the area. The 1990 community still show a low degree of similarity with 1988, possibly due to the increase in densities as a consequence of the heavy migration of animals to the Reserve. By 1991, the degree of similarity with the 1988






67


community, increased to 61.92. The community seems to be in the process of returning to its original community structure, apparently restoring its stability (Fig. 3-6).

Density comparisons with other western Amazonian sites

There is great variation in primate density among Amazonian sites (Table 3-13). I limited the comparison to data from neighboring states and/or countries in an attempt to avoid comparisons among areas with distinctly different climate and vegetation. Only unhunted or slightly hunted sites were used (n= 9, and n = 3, respectively). Mean primate density for the neotropics, calculated by Robinson and Redford (1986), is presented for reference.

It is clear that even light hunting strongly affects Ateles

density, and I cannot reject the possibility that Ateles has been hunted at the Jusante site. There are rubber tappers in the area, and fisherman sometimes hunt along the river. During the time I worked in the area I observed a few hunting incidents, however, the main targets were peccaries, deer, and the agouti. I never saw a captured or dead monkey, and when questioned, the rubber tappers confirmed that they mainly killed peccaries and agouties. I believe that the low density of Ateles at Jusante was related to habitat and not to hunting.

Saimiri shows an interesting situation where, in contrast to other species, the number of groups per km2 is somewhat fixed around 2 despite variation in the number of individuals per group. Total density

Estimates of total primate density show a general trend of density increase from 1989 to 1990 and then a decrease in 1991 (Table 3-6, Fig. 3-5). These changes are consistent with the assumption that the










Table 3-13: Primate densities in western Amazonia. Numbers per km2 (groups per kin2) All sites are unhunted, with the exception of Ponta da Castanha, Yavari Miri, and Mamore, which were lightly hunted.

SITE SPECIES REFERENCE

Ateles Cebus Pithecia Callicebus Saimiri Saguinus Callithrix
paniscus apella irrorata bruneus ustus fuscicollis emiliae

Samuel1988 13.5 (2.6) 21.7 (4.0) 5.2 (1.4) 0.7 (0.3) 17.0 (2.0) 5.4 (0.4) Eletronorte,1989
Samuel 1989 13.2 (3.2) 20.1 (5.4) 10.3 (3.4) 7.2 (3.6) 6.9 (0.9) 6.1 (1.9) 4.1 (1.4) Thisstudy
Samuel1990 23.6 (6.1) 28.5 (6.6) 5.2 (2.1) 1.3 (0.8) 28.7 (3.7) 14.3 (4.1) 10.1 (3.1)
Samuel 1991 11.1 (3.7) 27.1 (6.5) 3.5 (1.2) 1.2 (0.6) 8.9 (1.4) 14.1 (3.7) 3.6 (1.8)
Jusante 1990 5.9 (0.6) 17.5 (2.9) 7.5 (2.6) 20.0 (8.3) 23.8 (1.7) 20.4 (5.8) 9.1 (3.1)
Jusante1991 8.1 (2.1) 7.6 (2.4) 11.6 (4.9) 6.0 (0.8) 13.6 (3.7) 2.8 (0.9)
Acaituba 8.8 (1.5) 7.8 (1.0) Johns, 1986
Ponta da Cast. 1.3 (0.1) 11.5 (1.0) 32.0 (0.5) Johns, 1986
Igarape Acu 8.9 Peres, 1990
Tefe 11.9 9.7 Peres, 1990
Urucu 32.2 10.2 15.0 Peres, 1990
CochaCashu 25.0 40.0 60.0 16.0 Terborgh,1983
CochaCashu 22.4 (3.2) 36.0 (3.6) 84.0 (2.1) 10.8 (1.8) Freeseetal. 1982
Peru/Iquitos 25.0 (2.5) 72.0 (1.8) 15.0 (2.5) Freese et al. 1982
Yavari Miri 2.5 11.3 37.8 Puertas & Bodmer

1993
Mamore 2.0 (0.4) 55.0 (5.5) 100.0 (2.5) Freeseetal. 1982
MEAN 11.3 12.4 62.3 26.9 Robinson&

Redford 1986





69


animals inside the reservoir moved to both study sites between August 1989 and May 1990, and then dispersed to adjacent forest between November 1990 and May 1991. Even though 1989 data for Jusante do not exist, I suspect that the total 1989 primate density for Jusante was, like the Reserve, similar to its 1991 estimate.

The sharper density decrease at Jusante is most likely related to lower capacity of the forest to support high primate densities. This is perhaps most obvious for Cebus, whose density remained at high levels in the Reserve, but decreased drastically at Jusante (Table 3-5).

Because mean group size for Cebus and Saimiri at Jusante was much higher in 1990 than in 1991, the number of individuals per km2 shows a more abrupt decline in density than the group density (Fig. 3-5). Biomass

Metabolic biomass takes into consideration the energy consumption of the animal in relation to its body weight, therefore it makes a better assessment of the ecological importance of the animal. The three most important species accounted for 87, 89, and 63% of the metabolic biomass in the Reserve during 1989, 1990, and 1991, respectively (Table 3-7). The three dominant species in the 1988 metabolic biomass calculations were: Ateles = 53.68, Cebus = 40.58, and Pithecia = 9.08 kg0*75 km2, which are also the three largest species, and they accounted for 91% of the total 1988 metabolic biomass (it is important to remember that Saimiri was not sighted in the Reserve during the 1988 census, and hence that these calculations were based on six species instead of seven) (Eletronorte 1989). This pattern repeated itself in the 1989 biomass calculations (Table 3-7).






70


There is a significant relationship between biomass and body weight for primates, the largest the species the higher its biomass (Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979; Robinson and Redford 1986). This principle was observed in the 1989 biomass estimates (Table 3-7). However, the same was not true for the following years. The community went from having Ateles, Cebus, and Pithecia as the most important species in 1988 and 1989 (also the three largest), to having Ateles, Cebus, and Saimiri in 1990, and Cebus, Ateles, and Saimiri as the three most important species in 1991.

Ateles and Cebus were the first and second most important species during 1988, 1989 and 1990; however, in 1991 Cebus had greater metabolic biomass than Ateles. This is probably due to the migration of Ateles to adjacent forest between 1990 and 1991, and to the maintenance of the increased density of Cebus in the Reserve. Pithecia came in third place in 1988 and 1989, but it was replaced by Saimiri in 1990 and 1991. The high density increase of Saimiri probably resulted from their migration from the reservoir area to the Reserve.

Ateles and Cebus had similar biomass values in 1988 and 1989, while Pithecia's biomass in 1989 was much higher than in 1988, most likely a reflection of the increased density due to the number of released animals. Despite the increase in Pithecia's biomass from 1988 to 1989, it is clear that the release operation did not change the primate community significantly. The natural migration of the animals to the Reserve, on the other hand, changed the community considerably.

The 7% increase in total primate biomass from 1988 to 1989 was

probably a consequence of the release operation; the 55% increase from 1990 to 1991 was most likely due to the natural migration of the





71


animals; and the 40% decrease in 1991, which returned the total biomass to the same level found in 1988 (Fig. 3-7), was presumably due to dispersal of the animals to adjacent forests, and to mortality. This pattern of biomass changes strongly suggests that the primate carrying capacity for the Reserve is around 150 kg km2.

Despite the lack of data for 1988 and 1989 at Jusante, it is clear that the primate community at this site was also disrupted by the migration of animals from the reservoir (Table 3-8).

The similarity of the 1990-1991 pattern between sites suggests

that the Jusante site was affected in a similar fashion to the Reserve, and if so, the carrying capacity for the area would be around 50 kg km2. The much lower biomass values for Jusante is expected because the habitat favors smaller primate species, with Ateles occurring at very low densities.

The most likely explanation for the great increase in biomass

during 1990 is the migration of animals from the flooded reservoir to the Reserve and Jusante areas. The total lake area is 502 km2, of which 22 km2 remained green in the form of islands (measured from landsat images 1:250,000 by Adolfo de La Pria Pereira, SEDAM-RO). Before flooding, the area was undisturbed primary forest, with little or no hunting pressure. The rescue operation only removed a small fraction of the animals, the waters did not cover the tree tops, and there was no case of high primate mortality inside the reservoir. Therefore, I can safely state that the animals moved into adjacent areas.

In order to determine how many animals made it out of the reservoir area alive, I used the biomass values estimated for the Reserve and Jusante to estimate biomass inside the reservoir prior to





72


the disturbance. My sample areas at Jusante extended up to two kilometers away from the river bank. However, the northeast corner of the Reserve is located approximately two kilometers from the river bank (Fig. 2-1), and the vegetation in the Reserve had a different structure, and somewhat different composition than that of Jusante (Chapter 2). Based on these observations I measured the area along two kilometers on the right and left bank of the Jamari river, and considered it as "Jusante habitat type" flooded by the reservoir. The remaining area was then considered to be similar to the Reserve habitat type. The measurements indicated that 50% of the reservoir consisted of Jusante habitat, and the other 50% by the Reserve habitat. Then, I used the 1988 biomass estimates to calculate the biomass for 50% of the reservoir area, and assuming that the 1988 biomass values for Jusante were similar to the 1991 values (as is the case for the Reserve), I used a biomass value of 50 kg km2 to calculate the remaining 50% of the reservoir's primate biomass. The difference in biomass increase from 1989 and 1990 at the Reserve and Jusante, divided by the 1988 estimated biomass for the reservoir should then represent the proportion of the total primate biomass that successfully fled the reservoir area. The Reserve is officially 210 km2 in size, however, 30 km2 were also flooded by the reservoir, so I used the value of 180 km2 to calculate biomass for the Reserve. To be able to make a comparison, the same area was used to calculate total biomass for Jusante, even though both areas are located in a continuous forest. The result is presented in Table 3-14. If all the assumptions above are correct, the results show that at least 65% of the biomass of primates in the reservoir reached "safe" grounds within the Reserve and Jusante sites, even though they subsequently





73


Table 3-14: Estimation of primate biomass in the reservoir prior to flooding, and the percentage of which reached the Reserve and Jusante sites safely. (biomass for G and H calculations based on 50% of the reservoir's total area).



Site Year Biomass Area Total Biomass
(kg km2) (kM2) (kg kMf2)

a)Reserve 1989 165.27 180 29,748.60

b)Reserve 1990 255.41 180 45,973.80

c) (b minus a) = Amount increased 16,225.20

d)Jusante 1989 50.00 180 9,000.00

e)Jusante 1990 136.40 180 24,552.00

f)(e minus d) = Amount increased 15,552.00

g)Reservoir 1988 154.00 240 36,960.00

h)Reservoir 1988 50.00 240 12,000.00

i) (g plus h) = Total reservoir biomass 48,960.00


PERCENT OF RESERVOIR PRIMATE BIOMASS ACCOUNTED FOR


[(c + f) + iJ 100 = 48,960.00 = 65% of reservoir






dispersed to other areas or died (these calculations do not include the night monkey, Aotus, which also inhabits the area). This high percentage suggests that the migration of animals was not at random, but directed to the Reserve and Jusante areas.

If the estimation of the reservoir's biomass is accurate, only

4.5% of the primate biomass was rescued during the operation. I observed Cebus preying on snails and living inside the reservoir area during 1991, and I am certain that Cebus and other species still remain on some of the islands. The 22 km2 area of islands could harbor another 7% of






74


the reservoir's original biomass. The total biomass accounted for would then represent 76.5% of the total original biomass. That would leave 23.5% of the biomass unaccounted for. However, I have no doubt that some animals died of starvation, and others were preyed upon inside the reservoir (air born predators can easily kill an animal which cannot run to the ground for cover, or hide among entangled vegetation), others have certainly moved to other areas besides the Reserve and Jusante, and that would account for a significant portion of the remaining biomass. Terrestrial Diurnal Mammal


Density

The density of both terrestrial mammals censused at the Reserve were much higher in 1989 than in 1988 (Fig. 3-10), presumably as a result of the migration of animals to the area as a consequence of the rising water levels in the reservoir, plus the active release of animals captured inside the reservoir. Dasyprocta showed a much higher increase in density from 1988 to 1989 than Mazama. This is not only due to the fact that they naturally occur at higher densities than Mazama (therefore more animals migrated into the Reserve), but also to the fact that 214 individuals of Dasyprocta were released in the Reserve (of a total of 303 captured), as opposed to only 19 individuals of Mazama released (out of 23 captured) (Eletronorte 1989). The constant decrease in Dasyprocta density may be due to predation rather than dispersal. The Reserve has several species of carnivores, which certainly also increased in numbers due to the rising of the water level in the reservoir area. The apparent increase in Mazama density from 1989 to 1990 might be artificial due to restrictions of the TransAn program.





75


Even though the density estimate for 1990 only increased by about one individual per km2, the 95% confidence interval for 1990 is 100% larger than that for the 1989 estimate (Table 3-9).

Dasyprocta at Jusante showed a slight increase in density from 1990 to 1991, however, Mazama showed a 74% increase in density for the same period (Table 3-9, Fig. 3-10). There is no other obvious explanation for this increase, except perhaps that the species had a very good reproductive year. The densities of Dasyprocta are much higher at all times in the Reserve. The greater abundance of mature Bertholletia excelsia (Brazil nut) and Orbignya barbosiana (babaqu) at the Reserve (Chapter 2) most likely makes this area a better habitat for Dasyprocta. Mazama species had similar densities in both areas in 1991, and if the 1990 density at the Reserve was overestimated (due to TransAn restrictions, as mentioned above) then the 1990 densities for both areas might also have been similar.

Total density

The density estimate for all terrestrial mammals sampled at the Reserve was at its highest in 1989. Considering that only 313 individuals (belonging to nine species that were sampled during census) were released in the Reserve (Eletronorte 1989), the migration of animals to the Reserve is the most likely explanation for the higher 1989 density. Total group densities were virtually the same from 1989 to 1990, however, from 1990 to 1991 there was a 28% decrease in total group density. The sharper decrease in the number of ind/km2 is due to a reduction in mean group size from 1989 to 1990 (Table 3-10, Fig. 3-11).

The drastic density decrease from 1990 to 1991 was possibly a result of such factors such as:






76


a) mortality: high population densities in 1989 and 1990 may have lead to higher mortality due to inter- and intraspecific fighting and interference competition. Assuming that the 1988 densities were much lower for all terrestrial mammals (as they were for Dasyprocta and Mazama sp.; Fig. 3-10), total mammal density in 1989 and 1990 were most likely above the carrying capacity for the area, and a decline in density would be expected through migration to other areas and/or higher mortality rates. Small and mid-size mammals have relatively small home ranges (home range increases with body size (Eisenberg 1979, CluttonBrook 1979)), and they are less likely to migrate long distances then larger ones. Furthermore, migration might be expected to occur shortly after the increase in density, and result in density changes similar to that of the primate population (Fig. 3-5). Hence, high mortality rate is more likely to have occurred than migration. b)increased predation: even though I did not census the carnivore community, it is natural to assume that their numbers also increased due to migration into the Reserve. The increase in the number of animals in the area, plus a probable increase in litter size, due to increased food availability in 1989 and 1990, may have contributed to the decrease in total terrestrial mammal densities as a consequence of higher carnivore populations, and thus higher predation rates.

Even though group densities were virtually the same in 1990 and 1991 at Jusante, the number of ind/km2 had a 12% reduction during this time as a result of smaller mean group size in 1991 (Table 3-10, Fig. 311). The reasons for the decrease in density at Jusante are probably similar to those for the Reserve.





77


Biomass

As expected, the increase in terrestrial mammal biomass in the

Reserve occurred between 1988 and 1989 with the rush of animals fleeing the reservoir area (Fig. 3-12). Data for both deer species and the agouti should exemplify the response of the entire community to the flooding. The possible reasons for such changes have already been discussed in the terrestrial mammal density section of this discussion.

At Jusante the density for all three species increased from 1990 to 1991, however, this is a reflection of the increase in Mazama densities, and not an increase for the community as a whole (Fig. 3-12).

















CHAPTER 4
UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL DAM



Introduction


The Amazonian region is recognized by its highly diverse bird

communities (Terborgh et al. 1984, Bierregaard 1990), with the greatest concentration of species in its western portion (Haffer 1990). Bird species richness is greater in the Amazon Basin where riverine habitats occupy extensive areas (Remsen and Parker 1983).

The closing of the Samuel Hydroelectric Dam flood gates in the

Jamari River in November 1988 caused the submersion of approximately 107 km of pristine river bank, which represented a loss of approximately 214 km of riverine habitat. This figure considers only that portion of the river that was completely submerged. However, an even larger stretch of river was affected by the widening of the river caused by increased depth as a result of the dam.

Riverine habitats differ in plant and animal composition from adjacent habitats (Remsen and Parker 1983). Despite the fact that the filling of the reservoir created a water-edge habitat this edge was a terra firme forest before flooding, consequently its plant structure and composition are different from that of the previously existing riveredge. "The bird species composition of Amazonian river-created habitats is generally distinct from that of adjacent terra firme forest" (Remsen and Parker 1983, pg 226), therefore, the composition of bird species at the reservoir's edge will be somehow different from that of river's


78






79


edge. Will the species displaced from river edge habitats be found at the lakes's edge? Will the creation of edge habitat at the lake's perimeter increase bird diversity? A better understanding of the structure and composition of bird communities in such areas is essential for the preservation of bird diversity in the tropics.

The purpose of this chapter is to investigate the differences and/or similarities between understory bird communities in the newly created habitat and in its previously existing riverine habitat. In order to make comparisons, the understory bird community was sampled with the use of mist nets in a pristine riverine habitat (located downstream from the dam), as well as in an area located at the edge of the newly formed lake.


Methods



Study Site


Birds were sampled in the same sites as mammals, with the

exception of a new transect line created at the Jusante site (line D; Fig. 2-1). For a complete description of the study site see Chapter 3. Study Design


Three 2-km transects at the Reserve (located in areas 1, 2, and 3), and 3, 2-km transects at Jusante (located in areas A, B, and C in 1990, and B, C, and D in 1991), were sampled for understory bird community structure and composition (transect A was substituted for transect D in 1991 due to the proximity of its distant end to the river) (Fig. 2-1). Each transect ran from the edge of the water (lake in the






80


Reserve, and river at Jusante) to the forest interior at a 90 degree angle to the water course. Transects were between 2 and 4 km apart. Mist nets were placed along transects within three zones: (1) at the water's edge, (2) at 1 km, and (3) at 2 km distance from the water. At each zone, 15, 12-meter-long nets (2.6 m high, with 36 mm black nylon mesh), were placed along a trail which ran parallel to the main transect. Mist nets were placed with the lowest shelf at the ground, and in precisely the same location during replicate surveys. A period of at least three weeks separated samples of the same net-line.

This research was designed to sample small understory species; no attempt was made to sample birds by visual or auditory methods. Despite biases associated with the use of mist nets, such as the fact that resource availability may affect the probability of a bird's capture (Greenberg and Gradwohl 1986, Visscher 1981), and the fact that they may only capture 40% of the species present (Terborgh et al. 1990, Thiollay 1994), nets are a widely used technique in the study of understory bird communities (Karr 1981), and remain the most productive technique for this type of study where comparisons are to be made. Furthermore, mist nets provide an opportunity to collect reliable information in a relatively short period of time for an area of tropical forest with high species diversity (Karr and Freemark 1983). Data Collection


Sites were sampled for 9 days each month (three days in each transect: one day per zone/per transect) during the months of June through October of 1990 and 1991, which represents the dry season in the






81


region (Chapter 2) Number of days sampled per zone/per site are listed on Table 4-1.


Table 4-1: Number of mist-net days per site, year, and zone at the Reserve and Jusante during 1990 and 1991.


RESERVE JUSA=T
ZONE 1990 1991 1990 1991

EDGE 09 09 07 09

1 KM 09 09 07 09

2 KM 09 09 07 09

TOTAL 27 27 21 27

GRANDTOTAL 54 48




Nets were operated for 6 hours, from dawn to 1200/1300 hours. All identified birds were sexed and aged (whenever possible), banded with government-issued aluminum bands (except for hummingbirds), and released at the site of capture. Measurements on body mass (g), lengths of wing, tail, tarsus, bill (mm), and total body length (cm) were taken on all birds, along with records of date, time, and place of capture and/or recapture. Some specimens were collected for identification at the Goeldi Museum in Bel~m, others were photographed alive for later confirmation of field identification. Eight individual birds were not identified at species level during the course of this study. They were eliminated from the data set.


Data Analysis


The data set was initially described by species accumulation

curves, histograms, and tables of individuals and species captured per






82


site, year, and zone. Chi-square tests were used to compare number of captures and species between years within zones and sites. This was done in the interest of combining data from both years within each site.

Similarity among the species detected at the Reserve and Jusante, and among species detected in each zone (within and across areas) was quantified using Jaccard's Coefficient of Similarity, which equals zero if no species are in common and one if the lists are identical. Cluster analyses (group average strategy) were performed using the results of the similarity index in order to identify any resemblance between sites (Ludwig and Reynolds 1988).

To examine differences and/or similarities between understory bird community structure at the Reserve and Jusante sites, Correspondance Analysis (COA) were performed using the program CANOCO Version 3.12 (Ter Braak 1987).

For the COA analysis, species were first grouped into 6 foraging strata, 7 diet, and 8 substrata categories according to Karr et al. (1990). The initial arrangement resulted in 40 different categories, which were then condensed to 10 guild categories; insectivorous guilds were based on primary foraging substrate. The guild categories included six insectivore guilds (flycatchers=sallying/hawking/snatching [IA], dead-leaf gleaners [ID], live-foliage gleaners [IL], terrestrial [IG], bark-gleaners [IB], and army ant-followers [AF]), one nectarivore guild

(NI), one frugivore guild (FR), one omnivore guild (OM), and one guild of birds which were directly associated with water (WA). The latter guild includes piscivorous species, as well as insectivorous species which catch insects over the water. The species Xenops minutus was the only species that uses twigs as a substrate, so it was grouped with






83


those that use bark. The variables included in the COA consisted of the number of species detected in each guild, and the sampling units consisted of 18 survey areas. Information on the location of sampling sites along transects is not used in the COA; therefore, if samples become ordinated in a way that parallels this gradient, the COA will lend credence to the ecological significance of this gradient (Canaday 1991).


Results



Inter-Year Comparison


There were no significant differences between number of captures or number of species between 1990 and 1991 samples in either area (ChiSquare, p= 0.756, n= 546, p= 0.773, n= 603, for number of captures at the Reserve, and Jusante, respectively, and p= 0.980, n= 207, p= 0.918, n= 215, for number of species at the Reserve and Jusante, respectively) (Tables 4-2, and 4-3). Thus, data from 1990 and 1991 were pooled. Table 4-2: Number of captures per site, zone, and year. Chi-square test performed for each area on number of captures per zone between years. () percent of total capture.



RESERVE JUSANTE
ZORE 1990 1991 1990 1991

EDGE 127 (50) 139 (47) 78 (28) 93 (29)

1 KM 68 (27) 86 (29) 103 (36) 107 (34)

2 KM 57 (23) 69 (24) 103 (36) 119 (37)

TOTAL 252 (100) 294 (100) 284 (100) 319 (100)

CHI-SQUARE P=0.756 P=0.773






84


Table 4-3: Number of species per site, zone, and year. Chi-square test performed for each area on number of species per zone between years. () percent of total species.



RESERVE JUSANTE
ZONE 1990 1991 1990 1991

EDGE 43 (69) 42 (71) 37 (56) 42 (55)

1 KM 30 (48) 31 (53) 30 (45) 39 (51)

2 KM 31 (50) 30 (51) 30 (45) 37 (49))

TOTAL 62 (84) 59 (80) 66 (68) 76 (78)

CHI-SQUARE P=0.980 P=0.918





Overall Site Comparison


I found no significant difference in the number of species per zone (Chi-Square, p = 0.979, n = 293). However, the sites were highly differentiated when tested against the number of captures per zone (ChiSquare, p = 0.000, n = 1,149 (Table 4-4, Fig. 4-1).





Table 4-4: Number of species and number of captures per site, and zone (years combined) Chi-square test performed on number of species and capture per zone between sites. () percent of species present and total capture.



SPECIES CAPTURES
ZONE RESERVE JUSAZT RESERVE JUSANTE

EDGE 56 (76) 61 (63) 266 (49) 171 (28)

1 KM 43 (58) 47 (48) 154 (28) 210 (35)

2 KM 40 (54) 46 (47) 126 (23) 222 (37)

TOTAL 74 (100) 97 (100) 546 (100) 603 (100)

CHI-SQUARE P=0.979 P=0.000






85



EDGE


500 ,


7-


7

-z


I I
1KM 2KM


--] Reserve

Jusante


I I
EDGE 1 KM

Zones


2 KM


Figure 4-1: Number of species and number of captures at each zone at the Reserve and Jusante sites during 1990 and 1991.


140 120 100 80 60 40 -


0.
4)
0
E z3


20

0


400


300
0
200
E
Z 100


0 -





86


Despite the slightly higher sampling effort at the Reserve (n = 54 days), than at Jusante (n = 48 days), both the number of total captures and number of species were higher at Jusante than at the Reserve (n = 603 captures and 97 species, and n = 546 captures and n = 74 species, respectively) (Fig. 4-2). The species accumulation curve at the Reserve seems to be reaching an asymptote, while at Jusante the curve continues to extend upwards (Fig. 4-3 and 4-4); however, the cumulative number of captures and species do not differ significantly among sites (Kolmogorov-Smirnov, p = 0.6325 and 0.1462, n = 102, for number of species and number of captures, respectively). Within sites, the species accumulation curves are almost identical for the 1 and 2 km zones, while at the edge the number of species captured is higher than in the interior of the forest (Fig. 4-3 and 4-4). The Jusante site differs from the reserve in that there were fewer individuals captured at the edge than at the 1 and 2 km zones (Fig. 4-1, Table 4-4).




Species Richness


During 1990 and 1991, a total of 1,149 individuals of 118 species (Appendix B) were captured during 9,180 net-hours in both the Reserve and Jusante sites (see Appendix C see for list of species with occurrence confirmed). The two most speciose families for both areas combined were the Formicaridae and Tyrannidae, which were represented by 25 and 20 species, respectively. At Jusante those two families were also the most speciose, represented by 24 and 16 species, respectively. At the Reserve the Formicaridae had the highest number of species (14), while the Tyrannidae and the Furnaridae came in second represented by 10





87


100 90 80 U) 70 Q)

S60
0
z 50


40
E
E 30 -_Reserve
Jusante 20

10



0 100 200 300 400 500 600 700

Cumulative N of captures


Figure 4-2: Cumulative number of captures and species at the Reserve and Jusante sites for 1990 and 1991.





88


80 0 Total 0 Edge 0 1 km A 2 km



70



60

~O

*5 50U) / DC
q)



Z040 U)/



Reserve E 30



20



10

0



0 50 100 150 200 250 300 350 400 450 500 550 Cumulative N of captures






Figure 4-3: Cumulative number of captures and species per zone
for the Reserve during 1990 and 1991.





89


1 Total 0 Edge E 1 km A 2 km
1 00-90


80


70
V)

60 z60


50 -/uat
300
0




30
20






10


0 I I I I I I I T i i i
0 50 100 150 200 250 300 350 400 450 500 550 600 650 Cumulative N of captures






Figure 4-4: Cumulative number of captures and species per zone
for Jusante during 1990 and 1991.





90


species each. Tropical forest communities have long been recognized for their high diversity as well as their peculiar structure involving many rare species (Richards 1952). This study shows that the 20 most common species represent 70% of all captures at the Reserve, 67.8% at Jusante, and 63% for both areas combined. No one species dominates the rank abundance distribution, and a long tail of rare species is detectable in both areas separately, as well as in the combined samples (Fig. 4-5). Rarity is defined by Karr (1971) as those species making up less than 2% of individuals captured. At the Reserve 57 species (77% of the sample) occurred at rates lower than 2%; at Jusante, this proportion was higher (83 species or 86% of the sample; Kolmogorov-Smirnov p = 0.0003, n 171). The combined data set shows that of the 118 species captured, 105 occurred at rates lower than 2% (89% of the sample), and 27 (23%) were represented by only one individual.


Guild Structure


Of the total 118 species, and 1,149 individuals captured, 77 species (65%) and 789 captures (69%) were insectivores. Guild distribution for the Reserve and Jusante, and for both areas combined are shown in Figure 4-6.

Insectivores represented 68 and 69% of all captures at the Jusante and Reserve sites, respectively. The 2 guilds that had the greatest number of species at Jusante were the live-foliage gleaner insectivores (n = 30), and the omnivores (n = 11). At the Reserve the 2 guilds that had the greatest number of species were the live-foliage gleaners (n = 20), and the bark-gleaner insectivores (n = 9). When number of captures were considered, the 2 most common guilds at Jusante were the live-





91




Rank-ordered abundances


Reserve n=546


11 10 9
8
7
6

4
3
2
1
0-


11 10 9
8
7
6
5
4
3
2
1
0-


Number of species









Both areas n=1,1149




Ulii MMh b nionnfiu~inmomn


10 20 30


40


I I 1 8 9 50 60 70 80 90


Jusante n=603


. I02I3040I060I09, I
0 10 20 30 40 50 60 70 80 90100


100 110 120


Number of species







Figure 4-5: Rank-ordered abundance for the Reserve, Jusante, and both areas combined for 1990 and 1991.


(0)
q1)



0
C
n)


0 10 20 30 40 50 60 70 80 90100


1
1


CO Q)

0




C
Q)


Q)
a_


1 0
9
8
7
6
5
4

2

0-


0




Full Text
This dissertation was submitted to the Graduate Faculty of the
School of Forest Resources and Conservation in the College of
Agriculture and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of Doctor of JPhilosophy.
December, 1995
Director/ Forest Resources and
Conservation
Dean, Graduate School


73
Table 3-14: Estimation of primate biomass in the reservoir prior to
flooding, and the percentage of which reached the Reserve and Jusante
sites safely, (biomass for G and H calculations based on 50% of the
reservoir's total area).
Site
Year
Biomass
(kg km2)
Area
(km2)
Total Biomass
(kg km2)
a)Reserve
1989
165.27
180
29,748.60
b)Reserve
1990
255.41
180
45,973.80
c)(b minus a)
= Amount
increased
16,225.20
d)Jusante
1989
50.00
180
9,000.00
e)Jusante
1990
136.40
180
24,552.00
f)(e minus d)
= Amount
increased
15,552.00
g)Reservoir
1988
154.00
240
36,960.00
h)Reservoir
1988
50.00
240
12,000.00
i) (g plus h)
= Total
reservoir
biomass
48,960.00
PERCENT OF
RESERVOIR
PRIMATE
BIOMASS
ACCOUNTED FOR
[(c + f) -r i]
* 100 =
48,960.00
= 65%
of reservoir
dispersed to other areas or died (these calculations do not include the
night monkey, Aotus, which also inhabits the area). This high percentage
suggests that the migration of animals was not at random, but directed
to the Reserve and Jusante areas.
If the estimation of the reservoir's biomass is accurate, only
4.5% of the primate biomass was rescued during the operation. I observed
Cebus preying on snails and living inside the reservoir area during
1991, and I am certain that Cebus and other species still remain on some
of the islands. The 22 km2 area of islands could harbor another 7% of


122
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Leptopogon amaurocephalus
2
0
S-SI-F
Mionectes oleaginea
3
14
S-SO-F
Myiarchus ferox
1
0
C-SI-A
Myiarchus tuberculifer
1
0
C-SI-F
Myiobius barbatus
2
2
S-SI-A
Ochthoeca littoralis
1
0
W-SI-A
Onychorhynchus coronatus
1
20
U-SI-A
Platyrinchus coronatus
2
0
S-SI-F
Platyrinchus saturatus
2
0
S-SI-F
Ramphotrigon ruficauda
0
2
U-SI-F
Rhynchocyclus olivaceus
2
0
U-SI-F
Rhytipterna simplex
2
3
C-LI-F
Terenotriccus erythrurus
4
3
S-SI-A
Tolmomyias poliocephalus
2
0
C-SI-F
Hirundinidae
Stelgidopteryx ruficollis
2
0
W-SI-A
Troglodytidae
Cyphorhinus arada
4
6
G-SI-G
Microcerculus marginatus
4
10
G-SI-G
Thryothorus genibarbis
6
2
S-SI-D
Thryothorus leucotis
1
0
S-SI-F
Turdidae
Turdus albicollis
2
13
S-SO-F
Turdus amaurochalinus
0
1
S-SO-F
Parulidae
Basileuterus fulvicauda
3
1
S-SI-F
Thraupidae
Euphonia laniirostris
2
0
C-FR-F
Haba rubica
25
0
S-SO-F
Tachyphonus cristatus
1
0
C-SO-F


65
forest edge mixed with tall primary forest. Several studies have
demonstrated that the species occurs in greater densities in secondary
forest near natural clearings than in mature forest(Eisenberg and
Thorington 1973, Mittermeier and van Roosmalen 1981, Robinson and
Ramirez 1982, Johns and Skoruppa 1987). Emmons (1984) concluded that
Saguinus density appeared to have increased in some areas where large
monkeys had been exterminated. They also overlap with Cebus in most
habitat and diet categories (Mittermeier and van Roosmalen 1981). The
lower densities of Ateles and Cebus at Jusante, and the increased edged
habitat at the Reserve created by the reservoir, were most likely
favorable factors influencing the maintenance of higher Saguinus
densities at both sites during 1990 and 1991.
Callicebus. This species is also classified as frugivore-omnivore,
with more than 50% of its diet composed of fruits, and the remainder
mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and
Redford 1986, 1989). The species occurs in greatest densities in areas
characterized by forest openings with early successional vegetation, and
spends more time in the lower canopy levels and understory vegetation
(Kinsey 1981, Terborgh 1983, Robinson and Redford 1986, Robinson et all.
1987). The more open vegetation at the Jusante site (Chapter 2) most
likely created a more suitable habitat for Callicebus than in the
Reserve, which may explain their much higher densities at Jusante (Fig.
3-2). The possibility of interference competition with Cebus might also
affect Callicebus densities. Both species are catholic in their diet,
however, Cebus generally have larger group size, and are more aggressive
during interspecific encounters, possibly displacing Callicebus groups
from feeding trees. According to Emmons (1984), troops of larger


Percent of trees (%) Percent of trees (%)
26
Height (m)
Height (m)
Figure 2-7: Percentage of trees in each height (m)
category sampled in the understory vegetation at the
Reserve and at Jusante.


58
Discussion and Conclusions
The results suggest that the environmental changes created by the
construction of the dam altered the mammal community in the areas
adjacent to the reservoir.
Primates
Density estimates
Three different patterns can be seen with primate density changes
at the Reserve (Fig. 3-2). The first one, seen with Ateles, Callithrix,
and Saimir, is a large increase in density from 1989 to 1990, and then
a sharp decrease in 1991, returning to density levels similar to those
found in 1989. The second pattern, which involves Cebus, and Saguinus,
also is an increase in density from 1989 to 1990, however, the 1991
densities remaining high. The third pattern is a very different
situation, with the densities of Callicebus, and Pithecia being at their
highest in 1989, and declining in 1990, and 1991.
The increase in densities from 1989 to 1990 can be explained by
the migration of animals from the reservoir into the Reserve. Sixty
percent of the reservoir is, on average, only 3.5 m deep, which allowed
several tree species to survive for at least 8 months after the flooding
began. I observed flowering and fruiting trees inside the reservoir in
August 1989 (5 months after the completion of the filling of the
reservoir). Child (1968) observed the same phenomenon during the
formation of Lake Kariba as a result of the impoundment of the Zambezi
river in Zimbabwe. At Lake Kariba, species had different survival times
that varied from 4 to 12 months. He also observed that "most species
standing in water came into leaf and/or remained in leaf until they


42
Pithecia irrorata. Pithecia also had its highest density in the
Reserve in 1989, and then decreased thereafter. The density was 3.43
groups/km2 (10.3 ind/km2)in 1989, 2.07 groups/km2 (5.2 ind/km2) in 1990,
and 1.17 groups/km2 (3.5 ind/km2) in 1991(Table 3-5, Fig. 3-2). At
Jusante the density of Pithecia remained similar from 1990 to 1991: 2.60
groups/km2 (7.5 ind/km2), and 2.38 groups/km2 (7.6 ind/km2),
respectively (Table 3-5, Fig. 3-2).
Density and body weight of primates
When broad geographic regions are examined, differences in density
among primate species are related to body size. In general, population
densities decline with increasing body mass (Clutton-Brock and Harvey
1977, 1979; Eisenberg 1979; Robinson and Redford 1986; Kinnaird and
Eisenberg 1989). In the primate community at the Samuel Dam region, the
correlation between density and body weight was positive for primates in
the Reserve in 1989. This positive correlation is a contradiction of the
rule, because smaller animals should have higher densities than larger
ones. There was no significant correlation between density and body
weight in 1990 and 1991 (Fig. 3-3).
The correlation between density and body weight for primates at
Jusante was not significant in either year, but showed a tendency to be
negative in 1990, (Fig. 3-4).
Total primate density estimates
Primates at the Reserve comprised 61% of all mammalian sightings
in 1989, 69% in 1990, and 68% in 1991. When data for all primates are
pooled and density estimates were calculated for the area as a whole,
the result shows a 32% increase in group density from 1989 to 1990, and
a 23% decrease from 1990 to 1991 (Table 3-6, Fig. 3-5A) .


92
N of species
N of captures
Guilds
Figure 4-6: Guild structure of the community at the Samuel Ecological
Station, at Jusante, and both areas combined for 1990 and 1991.
Guild denominations are: IL = live-foliage gleaners, IB = bark-
gleaners, IA = flycatchers, IG = terrestrial insectivores, ID = dead-
leaf gleaners, AF = army ant-followers, OM = omnivores, FR =
frugivores, WA = birds which were directly associated with water, and
NI = nectarivores.


69
animals inside the reservoir moved to both study sites between August
1989 and May 1990, and then dispersed to adjacent forest between
November 1990 and May 1991. Even though 1989 data for Jusante do not
exist, I suspect that the total 1989 primate density for Jusante was,
like the Reserve, similar to its 1991 estimate.
The sharper density decrease at Jusante is most likely related to
lower capacity of the forest to support high primate densities. This is
perhaps most obvious for Cebus, whose density remained at high levels in
the Reserve, but decreased drastically at Jusante (Table 3-5).
Because mean group size for Cebus and Saimir at Jusante was much
higher in 1990 than in 1991, the number of individuals per km2 shows a
more abrupt decline in density than the group density (Fig. 3-5).
Biomass
Metabolic biomass takes into consideration the energy consumption
of the animal in relation to its body weight, therefore it makes a
better assessment of the ecological importance of the animal. The three
most important species accounted for 87, 89, and 63% of the metabolic
biomass in the Reserve during 1989, 1990, and 1991, respectively (Table
3-7). The three dominant species in the 1988 metabolic biomass
calculations were: Ateles = 53.68, Cebus = 40.58, and Pithecia = 9.08
kg0-75 km2, which are also the three largest species, and they accounted
for 91% of the total 1988 metabolic biomass (it is important to remember
that Saimir was not sighted in the Reserve during the 1988 census, and
hence that these calculations were based on six species instead of
seven) (Eletronorte 1989). This pattern repeated itself in the 1989
biomass calculations (Table 3-7).


Table 2-3: Average pH and nutrients per area, per category of depth
at the Samuel Ecological Station ( s.d.). Depth (cm), pH (H20)
P and K (ppm), Ca, Mg, and A1 (mEq/100 ml).
AREA
DEPTH
pH
P
1C
Ca
Mg
Al
01
0-20
3.85 (0.2)
2.55 (0.8)
47.73 (18.1)
0.20 (0.1)
0.16(0.1)
2.07 (0.4)
20-40
4.30 (0.4)
1.27 (0.5)
27.45 (22.9)
0.07 (0.1)
0.13 (0.1)
1.46 (0.3)
40-60
4.48 (0.2)
1.36 (0.7)
14.00 (22.3)
0.05 (0.1)
0.14(0.1)
1.15(0.2)
02
0-20
3.78 (0.1)
1.73 (0.6)
52.82 (26.7)
0.15(0.1)
0.18(0.1)
1.83 (0.3)
20-40
4.27 (0.3)
1.00 (0.0)
40.00 (35.6)
0.12(0.1)
0.22(0.1)
1.28 (0.4)
40-60
4.40 (0.2)
1.27 (0.5)
18.91 (12.2)
0.11 (0.0)
0.17(0.1)
1.19(0.2)
03
0-20
4.08 (0.2)
1.36(0.5)
43.36 (16.9)
0.75 (0.4)
0.52 (0.2)
0.92 (0.2)
20-40
4.33 (0.3)
1.00(0.0)
26.91 (20.9)
0.20(0.1)
0.29(0.1)
0.79 (0.2)
40-60
4.52 (0.4)
1.09(0.3)
35.36 (54.0)
0.19(0.1)
0.30 (0.1)
0.68 (0.3)
Note: Reference values for pH and nutrients (Source: EMBRAPAs Laboratory of Soil Analysis).
pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid
P: 0-10 ppm = low
11-30 ppm = medium
> 30 ppm = high
K: 0 45 ppm = low
46- 150 ppm = medium
>150 ppm = high
Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high
Al: 0.0 0.3 mEq = low
> 0.3 mEq = high


83
those that use bark. The variables included in the COA consisted of the
number of species detected in each guild, and the sampling units
consisted of 18 survey areas. Information on the location of sampling
sites along transects is not used in the COA; therefore, if samples
become ordinated in a way that parallels this gradient, the COA will
lend credence to the ecological significance of this gradient
(Canaday 1991).
Results
Inter-Year Comparison
There were no significant differences between number of captures
or number of species between 1990 and 1991 samples in either area (Chi-
Square, p= 0.756, n= 546, p= 0.773, n= 603, for number of captures at
the Reserve, and Jusante, respectively, and p= 0.980, n= 207, p= 0.918,
n= 215, for number of species at the Reserve and Jusante, respectively)
(Tables 4-2, and 4-3) Thus, data from 1990 and 1991 were pooled.
Table 4-2: Number of captures per site, zone, and year. Chi-square
test performed for each area on number of captures per zone between
years. () percent of total capture.
ZONE
RESERVE
1990
1991
JUSANTE
1990
1991
EDGE
127 (50)
139
(47)
78
(28)
93
(29)
1 KM
68 (27)
86
(29)
103
(36)
107
(34)
2 KM
57 (23)
69
(24)
103
(36)
119
(37)
TOTAL
252 (100)
294
(100)
284
(100)
319
(100)
CHI-SQUARE
P=0.756
P
=0.773


136
. 1988. Economic development and wildlife conservation in Brazilian
Amazonia. Ambio 17(5):302-306.
Johns, A. D., and J. P. Skoruppa. 1987. Responses of rain-forest
primates to habitat disturbances: a review. International Journal
of Primatology 8 (2): 157-191.
Johnson, E. G., and R. D. Routledge. 1985. The line transect method: a
nonparametric estimator based on shape restrictions. Biometrics
41:669-679.
Jordan, C. F. 1985. Soils of the Amazon rainforest. In: Amazonia. G. T.
Prance and T. E. Lovejoy (eds.). Pergamon Press, New York, N.Y.
Junk, W. J., and J. A. S. Nunes de Mello. 1987. Impactos ecolgicos das
represas hidreltricas na bacia Amaznica Brasileira. Tubinger
Geographische Studien 95:367-385.
Karr, J. R., 1971. Structure of avian communities in selected Panama and
Illinois habitats. Ecological Monographs 41:207-231.
. 1981. Surveying birds with mist nets. Studies in Avian Biology
6:62-67.
Karr, J. R., and K. E. Freemark. 1983. Habitat selection and
environmental gradients: Dynamics in the stable" tropics.
Ecology 64(6):1481-1494.
Karr, J. R., S. K. Robinson, J. G. Blake, and R. O. Bierregaard, Jr.
1990. Birds of four netropical forests. In: Four Neotropical
Forests. A. H. Gentry (ed.). Yale University Press, London.
Kinnaird, M., and J. F. Eisenberg. 1989. A consideration of body size,
diet, and population biomass for neotropical mammals. In: Advances
in Neotropical Mammalogy. K. H. Redford and J. F. Eisenberg
(eds.). The Sandhill Crane Press Inc, Gainesville, FL.
Kinsey, W. G. 1977. Diet and feeding behavior of C. torquatus. In:
Primate Ecology; Studies of Feeding and Ranging Behaviour in
Lemurs, Monkeys and Apes. T. H. Clutton-Brock (ed.). Academic
Press, New York, N.Y.
. 1981. The titi monkeys, genus Callicebus. In: Ecology and
Behavior of Neotropical Primates, Vol. 1. A. F. Coimbra-Filho, and
R. A. Mittermeier (eds.). Academia Brasileira de Ciencias. Rio de
Janeiro, R.J.
Knight, D. H. 1978. Methods for sampling vegetation: An instructional
manual. Department of Botany, University of Wyoming, Laramie.
Liao, W., D. S. Bhargava, and J. Das. 1988. Some effects of dams on
wildlife. Environmental Conservation 15(l):68-70.
Lisboa, P. L. B. 1990. Rondnia: Colonizago e Floresta. Programa
Polonoroeste. Relatrio de Pesquisa n 9, SCT/PR-CNPq, Brasilia,
D.F.


110
Rescue operations have become a public relations strategy used by
power companies to appease public opinion. The only way to make power
companies change their policies is to inform the public, and to give
power companies better options for rescue/conservation programs. Some
suggestions for future rescue programs are listed below:
1. Rescue operations should be confined to:
(a) endangered species. Several species classified as "endangered" by
IBAMA (Instituto Brasileiro do Meio Ambiente e Recursos Naturais
Renovveis), and "vulnerable" by IUCN (International Union for the
Conservation of Nature) were present at Samuel. These included giant
anteaters (Mirmecophaga tridactyla) giant armadillos (Priodontes
maximus), spider monkeys (Ateles paniscus), and many others. Species
known to be endangered and/or vulnerable should be rescued when
stranded.
(b) species that are unable to escape their flooded environment.
Sometimes females with infants are trapped on small islands unable to
swim with their young. An effort should be made to rescue those animals.
(c) species that could be used for research. At Samuel, poisonous
snakes, scorpions, and spiders were sent to several centers for
development and/or production of vaccines.
(d) species that could be used for re-establishing depleted populations
elsewhere. Some areas in the Amazon have been heavily hunted, and large
species are sometimes rare. Rescued animals such as deer, peccaries, and
large monkeys could be released at nearby sites where populations of
those species are low.


BIOGRAPHICAL SKETCH
Rosa Maria Lemos de S was born in the city of Belo Horizonte,
Minas Gerais, Brazil. Being from a large city, and from a cosmopolitan
family, she remained unaware of nature until the images of "The Wild
Kingdom" reached her through a colored screen. It was love at first
sight! She packed her bags and journed to the University of Wisconsin-
Stevens Point, where she earned a B.S. degree in Wildlife Management.
From there, she returned to Brazil to get hands-on experience with the
"real" wildlife. She earned her Master's Degree from the University of
Brasilia working with the highly endangered species Brachyteles
arachnoides (woolly spider monkey) in one of the few remaining small
patches of the Atlantic Forest. She was then lured to the University of
Florida's Program for Studies in Tropical Conservation, where she was,
finally, introduced to the Amazon forest the ultimate place for
wildlife! The Amazon's magnificence has forever changed her image of
wilderness, and will keep her "busy" for the rest of her life with the
unending task of trying to preserve one of the few places on earth still
relatively unknown to mankind.
140


Log10 biomass (g/km2)
51
O Ateles A Cebus Pithecia Saimir Callicebus O Saguinus V Callithrix
Log10 body weight (g)
Figure 3-8: Relationship between biomass and body weight for
primates at the Samuel Ecological Station. (P = 0.0001,
0.0403, and 0.0289 for 1989, 1990, and 1991, respectively).
Log10 body weight (g)
Figure 3-9: Relationship between biomass and body
weight for primates at Jusante. (P = 0.0131,
and 0.0237 for 1990, and 1991, respectively). See
Figure 3-8 for legends.


vegetation in the flooded area was dead, and primate biomass at the
reserve increased to 255 ( 109) kg km2, because of the migration of
animals from the reservoir to the reserve. By 1991, primate biomass in
the reserve returned to levels similar to 1988, 153 ( 81) kg km2, most
likely due to dispersal of animals to adjacent areas. The pattern of
biomass fluctuation indicates that primate carrying capacity at the
reserve is about 150 kg km"2.
Terrestrial mammal densities and biomass at the Samuel Ecological
Station increased in 1988, immediately after flooding, and remained high
until 1990 when they started decreasing.
Understory bird community structure and composition was sampled at
the Samuel Ecological Station, and at an area located at the Jamari
River's edge, downstream from the dam. The creation of the Samuel Lake
destroyed approximately 214 km of pristine riverine habitat, displacing
an entire community of birds. Despite the creation of a water-edge
habitat by the reservoir, the composition of bird species at the lake's
edge is different from that of the downstream site, therefore, different
from previously existing river edges at the reservoir site. However,
community structure seems to be similar at both sites. Water-related
niches were created at the reservoir's edge, yet, the uniqueness of
species composition at riverine habitats was not emulated.
ix


60
The densities for Ateles and Saguinus were almost identical in
1988 and 1989. Because only a few individuals were released in the
Reserve, there was no reason to expect otherwise. The higher 1989
estimates for Saimir, Callicebus, Pithecia, and to some extent Cebus,
reflect the increase in density caused by the released animals. Because
my study began only two months after the release was completed, it is
reasonable to assume that the animals were still inside the Reserve, and
that is why the 1989 densities were higher than the 1988. As for
Callithrix, the difference between the 1988 and 1989 densities may be a
reflection of the species characteristics. Callithrix are among the most
cryptic primates, in both pelage and habitat (Ferrari and Rylands 1994),
making them difficult to detect during transect samples. Because the
1988 censuses were not performed by my team, their low density might be
a consequence of differences in researcher's detection ability. Despite
having similar body weight, the same detectability differences do not
apply to Saguinus, "while relatively similar in body size and most
aspect of their ecology, S. fuscicollis invariably prefers lower forest
strata them its congeners" (Ferrari and Rylands 1994, pg. 82), which
makes them more visible during a census.
The decline in density estimates in 1991 for all species is most
likely a consequence of the dispersal of animals to adjacent areas or
death. The Reserve is located adjacent to an area of continuous forest,
without human inhabitation or access roads (Fig. 1-3), and dispersal
into those areas would be the expected behavior for animals in a
situation of crowding.
There was a sharp decline in density (between 37 and 71%) from
1990 to 1991 at Jusante with all primates, except Pithecia (Fig. 3-2).


131
ORDER
FAMILY SPECIES
Myiarchus spp
Contopus virens
Tolmomyias sulphurescens
Todirostrum ciereum
Pipromorpha oleaginea
Terenotriccus erythrurus
Elaenia sp
Myiobius barbatus
Hemitriccus zosterops
Laniocerca hypopyrrha
Mionectes oleagines
Onychorhynchus coronatus
Platyrhinchus saturinus
Ramphotrigon ruficauda
Rhytiopterna simplex
Hirundinidae Atticora fasciata
Stelgidopterix ruficollis
Tachycineta albiventer
Progne chalybea
Troglodytidae Troglodytes aedon
Tryothorus genibarbis
Tryothorus coraya
Campylorhynchus turdinus
Cyphorhinus arada
Microcerculus marginatus


112
size as the forested area lost by the creation of the reservoir, and
should contain habitats similar to those lost.
The maintenance (including management, research, and protection)
of the conservation unit has to be the responsibility of the power
company in charge of the hydroelectric dam construction. Their profit
does not end with the construction of the power plant; therefore, their
responsibility to conservation should not end either, but rather
continue, at least, throughout the life span of the hydroelectric dam.
Preliminary Studies
Preliminary studies of plant and animal communities at future
reservoir sites, as well as at the conservation unit, are key
instruments for conservation strategies, and should be carried out well
in advance of dam constructions. Most researchers are called to dam
sites only a few months before the flooding occurs. It is imperative
that environmental studies begin at the same time as do feasibility
studies for the construction of a particular dam. Eletronorte started
the feasibility study at the Samuel Dam site some 10 years before
construction (Francisco Silveira Pereira, pers. communication); however,
ecological studies at the site started only 1 year before the filling of
the reservoir (Eletronorte 1989). Data on mammal density, and bird
community composition from the reservoir area, prior to its filling,
would have greatly helped in the understanding of the impact of the dam
in those communities.


Groups/kilometer square
44
7
6
5
4
3
2
1
0
Cebus
7
6
5
4
3
2
1
0
Callithrix
1
7 n
6
5 -
4 -
3 -
2 -
1 -
0
Saimir
n
i
oo
cr>
o

00
CD
O
r
OO
CD
O
00
oo
CD
CD
OO
OO
CD
CD
OO
OO
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
7
6
5
4
3
2
1
0
Saguinus
\
\
_N
\
\
\
\
\
\
00
CD
O
t
00
CD
O
OO
OO
CD
CD
OO
00
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
9
8
7
6
5
4
3
2
1
0
Callicebus
J=t-
7
6
5
4
3
2
1
0
Pithecia
00
CD
O
T
00
CD
O
,
00
00
CD
CD
00
00
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
'

*
^
^
<
T
s
Year
Figure 3 2: Primate density estimates for the Samuel Ecological Station
during 1988 (Eletronorte 1989), 1989, 1990, 1991 (this study), and the
Jusante in 1990, and 1991 (this study).


90
species each. Tropical forest communities have long been recognized for
their high diversity as well as their peculiar structure involving many
rare species (Richards 1952). This study shows that the 20 most common
species represent 70% of all captures at the Reserve, 67.8% at Jusante,
and 63% for both areas combined. No one species dominates the rank
abundance distribution, and a long tail of rare species is detectable in
both areas separately, as well as in the combined samples (Fig. 4-5).
Rarity is defined by Karr (1971) as those species making up less than 2%
of individuals captured. At the Reserve 57 species (77% of the sample)
occurred at rates lower than 2%; at Jusante, this proportion was higher
(83 species or 86% of the sample; Kolmogorov-Smirnov p = 0.0003, n =
171). The combined data set shows that of the 118 species captured, 105
occurred at rates lower than 2% (89% of the sample), and 27 (23%) were
represented by only one individual.
Guild Structure
Of the total 118 species, and 1,149 individuals captured, 77
species (65%) and 789 captures (69%) were insectivores. Guild
distribution for the Reserve and Jusante, and for both areas combined
are shown in Figure 4-6.
Insectivores represented 68 and 69% of all captures at the Jusante
and Reserve sites, respectively. The 2 guilds that had the greatest
number of species at Jusante were the live-foliage gleaner insectivores
(n = 30), and the omnivores (n = 11). At the Reserve the 2 guilds that
had the greatest number of species were the live-foliage gleaners (n =
20), and the bark-gleaner insectivores (n = 9). When number of captures
were considered, the 2 most common guilds at Jusante were the live-


Individuals captured and released 52
Density estimates after damming 53
Total terrestrial mammal density estimates 54
Biomass 56
Discussion and Conclusions 58
Primates 58
Density estimates 58
Density changes between years at the Reserve 66
Density comparisons with other western Amazonian sites 67
Total density 67
Biomass 69
Terrestrial Diurnal Mammal 74
Density 74
Total density 75
Biomass 77
4 UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL
DAM 78
Introduction 78
Methods 79
Study Site 79
Study Design 79
Data Collection 80
Data Analysis 81
Results 83
Inter-Year Comparison 83
Overall Site Comparison 84
Species Richness 86
Guild Structure 90
Faunal Similarity Between and Within Sites 93
Ecological Similarities Among Sites 97
Discussion and Conclusions 97
Overall Site Comparison 97
Species richness 97
Species abundance 101
Guild Structure 101
Faunal Similarities 102
Ecological Similarities 103
5 RECOMMENDATIONS FOR FUTURE HYDROELECTRIC DAM CONSERVATION PROGRAMS.105
Overview 105
Considerations 107
Rescue Operations 107
Creation of Conservation Units Ill
Preliminary Studies 112
Follow-up Studies 113
Conclusions 114
APPENDIX A MAMMALIAN SPECIES RECORDED AT THE SAMUEL DAM SITE 115
APPENDIX B BIRD SPECIES CAPTURED DURING THIS STUDY 118
APPENDIX C BIRD SPECIES RECORDED AT THE SAMUEL DAM SITE 124
LIST OF REFERENCES 134
BIOGRAPHICAL SKETCH 140
vii


Percent of trees (%) Percent of trees (%)
22
DBH (cm)
DBH (cm)
Figure 2-4: Percentage of trees in each DBH (cm) category
at the Samuel Reserve and Jusante.


75
Even though the density estimate for 1990 only increased by about one
individual per km2, the 95% confidence interval for 1990 is 100% larger
than that for the 1989 estimate (Table 3-9).
Dasyprocta at Jusante showed a slight increase in density from
1990 to 1991, however, Mazama showed a 74% increase in density for the
same period (Table 3-9, Fig. 3-10). There is no other obvious
explanation for this increase, except perhaps that the species had a
very good reproductive year. The densities of Dasyprocta are much higher
at all times in the Reserve. The greater abundance of mature
Bertholletia excelsia (Brazil nut) and Orbignya barbosiana (babagu) at
the Reserve (Chapter 2) most likely makes this area a better habitat for
Dasyprocta. Mazama species had similar densities in both areas in 1991,
and if the 1990 density at the Reserve was overestimated (due to TransAn
restrictions, as mentioned above) then the 1990 densities for both areas
might also have been similar.
Total density
The density estimate for all terrestrial mammals sampled at the
Reserve was at its highest in 1989. Considering that only 313
individuals (belonging to nine species that were sampled during census)
were released in the Reserve (Eletronorte 1989), the migration of
animals to the Reserve is the most likely explanation for the higher
1989 density. Total group densities were virtually the same from 1989 to
1990, however, from 1990 to 1991 there was a 28% decrease in total group
density. The sharper decrease in the number of ind/km2 is due to a
reduction in mean group size from 1989 to 1990 (Table 3-10, Fig. 3-11).
The drastic density decrease from 1990 to 1991 was possibly a
result of such factors such as:


17
Total annual precipitation in the Samuel region ranges from 2,000 .
to 2,600 mm year. June, July, and August are characteristically dry, and
the dry season lasts from May to September. The rainy season is also
well defined by the months of October to April, when monthly
precipitation usually exceeds 200 mm (Table 2-2, Fig. 2-3).
Table 2-2: Total monthly precipitation
(mm) in the Samuel Dam, during 1989,
1990, and 1991 (source Eletronorte).
Month
Year
1989
1990
1991
January
302.4
400.0
326.9
February
319.4
521.0
223.1
March
248.0
341.0
412.9
April
316.6
147.0
315.1
May
165.8
133.0
100.4
June
107.1
47.0
38.7
July
59.2
24.0
0.0
August
59.9
39.0
0.0
September
83.7
146.3
42.2
October
136.0
274.4
142.6
November
89.9
318.0
233.2
December
275.0
244.0
306.9
Annual
total
2,163.0
2,634.7
2,142.0


130
ORDER
FAMILY SPECIES
Myrmeciza hemimelaena
Myrmoborus myotherinus
Myrmotherula hauxwelli
Myrmotherula longipennis
Myrmotherula ornata
Percnostola leucostigma
Sclateria naevia
Cotingidae Lipaugus vociferans
Iodopleura isabellae
Tityra cayana
Attila spadiceus
Pipridae Pipra rubrocapilla
Pipra nattereri
Pipra fasciicauda
Pipra sp
Machaeropterus pyrocephalus
Schiffornis turdinus
Chiroxiphia parela
Heterocercus flavivertex
Tyrannidae Xolmis irupero
Pitangus sulphuratus
Muscvora tyrannus
Tyrannus melancholicus
Megarhynchus pitanga
Myiozetetes similis


139
van Roosmallen, M. G. M. 1980. Habitat Preferences, Diet, Feeding
Strategy and Social Organization of the Black Spider Monkey
(Ateles p. paniscus Linnaeus 1758) in Surinam. Rijksinstituut voor
Natuurbeheer. Arnhem, The Netherlands.
van Roosmalen, M. G. M., and L. L. Klein. 1988. The spider monkeys,
genus Ateles. In: Ecology and Behavior of Neotropical Primates,
Vol. 2. R. A. Mittermeier, A. B. Rylands, A. Coimbra-Filho, and G.
A. B. Fonseca (eds.). World Wildlife Fund. Washington, D.C.
Visscher, M. N. 1981. Consideraciones sobre el uso de redes de neblina
en el anlisis de comunidades de aves en habitat tropical. Acta
Biolgica Venezuelica 11(2):89-107.
Walsh, J., and R. Gannon. 1967. Time Is Short and the Water Rises. E. P.
Dutton & Co., Inc. New York, N.Y.
Wilson, E. 0. 1988. Biodiversity. National Academy Press, Washington,
D.C.


Ind/km2 Groups/knr
47
Reserve
Jusante
Figure 3-5: A) Total primate density (groups/km2) for the Reserve,
and Jusante for 1989, 1990, and 1991. B) Total primate density
(ind/km2) for the Reserve, and Jusante for 1989, 1990, and 1991.


71
animals; and the 40% decrease in 1991, which returned the total biomass
to the same level found in 1988 (Fig. 3-7), was presumably due to
dispersal of the animals to adjacent forests, and to mortality. This
pattern of biomass changes strongly suggests that the primate carrying
capacity for the Reserve is around 150 kg km2.
Despite the lack of data for 1988 and 1989 at Jusante, it is clear
that the primate community at this site was also disrupted by the
migration of animals from the reservoir (Table 3-8).
The similarity of the 1990-1991 pattern between sites suggests
that the Jusante site was affected in a similar fashion to the Reserve,
and if so, the carrying capacity for the area would be around 50 kg km2.
The much lower biomass values for Jusante is expected because the
habitat favors smaller primate species, with Ateles occurring at very
low densities.
The most likely explanation for the great increase in biomass
during 1990 is the migration of animals from the flooded reservoir to
the Reserve and Jusante areas. The total lake area is 502 km2, of which
22 km2 remained green in the form of islands (measured from landsat
images 1:250,000 by Adolfo de La Pria Pereira, SEDAM-RO). Before
flooding, the area was undisturbed primary forest, with little or no
hunting pressure. The rescue operation only removed a small fraction of
the animals, the waters did not cover the tree tops, and there was no
case of high primate mortality inside the reservoir. Therefore, I can
safely state that the animals moved into adjacent areas.
In order to determine how many animals made it out of the
reservoir area alive, I used the biomass values estimated for the
Reserve and Jusante to estimate biomass inside the reservoir prior to


Temperature (C)
16
Table 2-1: Average monthly temperature
(C) at the Samuel Dam, during 1989,
1990, and 1991 (source Eletronorte).
Month
rear
1989
1990
1991
January
27.0
27.2
26.7
February
26.0
26.7
26.8
March
27.0
26.4
26.4
April
27.0
26.6
27.0
May
27.0
26.7
26.6
June
26.5
26.1
26.1
July
25.0
25.7
24.4
August
27.0
27.0
25.5
September
28.0
27.0
26.6
October
27.0
26.4
27.3
November
28.0
28.0
27.8
December
28.0
28.0
27.4
Annual
average
26.9
26.8
26.6
Figure 2-2: Average monthly temperature (C),
at the Samuel Dam region for 1989,1990, and 1991


5
Figure 1-2: Satellite image of the Samuel Dam Reservoir under
construction (From Instituto Nacional de Pesquisas Espaciis -
INPE, July 1987).


93
foliage gleaner insectivores with 24% of all captures, and the
frugivores with 20% of all captures. At the Reserve the 2 guilds with
the greatest numbers of captures were the live-foliage gleaner, and the
bark-gleaner insectivores with 22 and 21% of all captures, respectively.
Despite differences, community structure did not differ significantly
between sites (Kolmogorov-Smirnov, p = 0.7591 and 0.9883, n = 20, for
percentage of captures per guild, and percentage of species per guild,
respectively).
Faunal Similarity Between and Within Sites
The 12 most abundant species in each area are shown in Table 4-5.
The lists show great differences in abundance of species between sites.
Jaccard's Similarity Index for the Reserve and Jusante as a whole was
0.449, which indicates that the sites shared approximately half of the
species detected. Of the 118 species detected, the Reserve had 21
species that were unique, while Jusante had 44 unique species. The
number of species in common between sites was 52. The index was also
used to determine similarity of the species lists between zones (across
and within sites). The similarity indexes for all zones are shown in
Table 4-6.
Note that the edge zone at Jusante had very little in common with
any other zone (the least similar zones were the edge and the 2 km sites
at the Jusante). The edge at the Reserve had an index of approximately
0.35 across zones, while the 1 and 2 km zones for both sites had indices
ranging from 0.45 to 0.63 (with the highest similarity index between the
1 and 2 km zones at Jusante). A cluster analysis was performed using the


48
25.27 -
3U.I0
75.09 -
19
88 19
91 IS
Years
90 19
89
Figure 3-6: Cluster analysis for all four years of density
data at the Reserve. Similarity levels = 61.92, 27.94, and
25.27, respectively (Minitab 10, Hierarchical cluster
analysis of observations).
Biomass
In the Reserve in 1989, the three largest primates (Ateles, Cebus,
and Pithecia, respectively) were also the three species that contributed
most to the crude and metabolic biomass. However, by 1990 Saimir
contributed more than Pithecia, despite weighing only a third of the
weight of Pithecia. By 1991, Cebus was the species that contributed most
to the metabolic biomass calculations, followed by Ateles and Saimir.
The total primate crude biomass for the three years showed a 55%
increase from 1989 to 1990, and then a decrease of 40% by 1991,
returning to levels similar to 1989 (Table 3-7). The total primate crude
biomass for 1988 was 154 kg/km2 (Eletronorte 1989); a value very similar
to the 1989 and 1991 values (Fig. 3-7).


77
Biomass
As expected, the increase in terrestrial mammal biomass in the
Reserve occurred between 1988 and 1989 with the rush of animals fleeing
the reservoir area (Fig. 3-12). Data for both deer species and the
agouti should exemplify the response of the entire community to the
flooding. The possible reasons for such changes have already been
discussed in the terrestrial mammal density section of this discussion.
At Jusante the density for all three species increased from 1990
to 1991, however, this is a reflection of the increase in Mazama
densities, and not an increase for the community as a whole (Fig. 3-12).


125
ORDER FAMILY
SPECIES
Cathartes burrovianus
Coragyps atratus
Accipitridae
Heterosnizias meridionalis
Buteogallus urubitinga
Leptodon cayanensis
Ictinia plmbea
Buteo nitidus
Leocopternis shistacea
Elanoides forficatus
Accipiter superciliosus
Falconidae
Daptrius americanus
Daptrius ater
Milvago chima chima
Polyborus plancus
Falco rufigularis
Herpetotheres cachinnans
Galliformes Phasianidae
Odontophorus gujanensis
Cracidae
Mi tu mitu
Notocrax urumutum
Opisthocomiformes Opisthocomidae
Notocrax sp
Opistochomus hoazin
Gruiformes Eurypygidae
Eurypyga helias
Psophiidae
Psophia viridis
Rallidae
Pomphinula martinika
Heliornithidae
Heliornis flica


121
ORDER/FAMILY
Pipridae
Cotingidae
Tyrannidae
SPECIES
J(N)
R(N)
GUILD
Percnostola leucostigma
0
4
S-LI-F
Phlegopsis nigromaculata
10
18
S-LI-R
Rhegmatorhina hoffmannsi
14
0
S-LI-R
Sclateria naevia
1
3
S-SI-G
Thamnomanes caesius
5
0
S-LI-F
Thamnomanes saturinus
13
2
S-LI-F
Thamnophilus aethiops
5
2
S-LI-F
Thamnophilus amazonicus
4
0
U-LI-F
Thamnophilus schistaceus
3
0
U-LI-F
Chiroxiphia parela
6
5
S-FR-F
Heterocercus flavivertex
0
1
S-FR-F
Heterocercus linteatus
1
0
S-FR-F
Manacus manacus
1
0
S-FR-F
Pipra fasciicauda
21
1
S-FR-F
Pipra nattereri
61
39
S-FR-F
Pipra rubrocapilla
11
34
S-FR-F
Schiffornis turdinus
21
16
S-SI-F
Tyranneutes stolzmanni
1
0
U-SO-F
Lipaugus vociferaos
0
3
C-LO-F
Attila cinnamomeus
0
1
U-LI-F
Attila spadiceus
2
5
U-LI-F
Cnemotriccus fuscatus
1
0
S-SI-A
Elaenea parvirostris
1
0
C-SO-F
Hemitriccus zosterops
0
4
U-SI-F
Laniocera hypopyrrha
0
3
C-LO-F


43
Table 3-5: Primate density estimates at the Reserve and at Jusante.
N = number of
sightings
, D =
group
density
, MGS
= mean group size.
SPECIE
SITE
YEAR
N
D
MGS
IND/XM1
95% D Cl
Ateles
Reserve
89
21
3.15
4.2
13.2
1.53-07.43
paniscus
Reserve
90
72
6.06
3.9
23.6
3.95-09.91
Reserve
91
66
3.69
3.0
11.1
2.14-07.06
Jusante
90
04
0.60
9.8
5.9
0.17-02.63
Jusante
91
01

1.0

Cebus
Reserve
89
54
5.42
3.7
20.1
3.21-09.22
apella
Reserve
90
71
6.63
4.3
28.5
4.15-10.31
Reserve
91
72
6.45
4.2
27.1
4.01-10.32
Jusante
90
32
2.92
6.0
17.5
1.59-06.15
Jusante
91
27
2.12
3.8
8.1
0.98-04.57
Pithecia
Reserve
89
33
3.43
3.0
10.3
1.76-06.59
irrorata
Reserve
90
17
2.07
2.5
5.2
0.75-05.06
Reserve
91
12
1.17
3.0
3.5
0.35-02.49
Jusante
90
21
2.60
2.9
7.5
1.16-05.57
Jusante
91
16
2.38
3.2
7.6
1.08-06.39
Callicebus
Reserve
89
15
3.61
2.0
7.2
1.57-08.34
bruneus
Reserve
90
05
0.83
1.6
1.3
0.20-03.03
Reserve
91
09
0.62
2.0
1.2
0.24-01.74
Jusante
90
59
8.33
2.4
20.0
4.77-14.72
Jusante
91
39
4.85
2.4
11.6
2.49-09.89
Saimir
Reserve
89
08
0.91
7.6
6.9
0.25-02.43
ustus
Reserve
90
05
3.68
7.8
28.7
0.89-05.68
Reserve
91
12
1.41
6.3
8.9
0.49-03.64
Jusante
90
15
1.69
14.1
23.8
0.62-04.55
Jusante
91
09
0.75
8.0
6.0
0.29-02.58
Saguinus
Reserve
89
14
1.85
3.3
6.1
0.71-03.88
fuscicollis
Reserve
90
24
4.08
3.5
14.3
1.92-09.18
Reserve
91
20
3.70
3.8
14.1
1.51-06.53
Jusante
90
34
5.83
3.5
20.4
2.88-11.23
Jusante
91
28
3.67
3.7
13.6
1.74-07.79
Callithrix
Reserve
89
10
1.41
2.9
4.1
0.38-03.82
emiliae
Reserve
90
15
3.06
3.3
10.1
1.31-06.69
Reserve
91
09
1.82
2.0
3.6
0.74-06.00
Jusante
90
20
3.13
2.9
9.1
1.34-05.97
Jusante
91
07
0.91
3.1
2.8
0.28-03.75


107
released (Fig. 5-1) (Eletronorte 1989). It also was the first time that
a land conservation unit, or preserve, was created to compensate for the
loss of habitat caused by flooding. Despite the improvements made in the
Samuel rescue operation in contrast to previous efforts, many aspects of
rescue operations need to be reconsidered.
Considerations
Rescue Operations
The cost-benefit ratio of rescue operations need to be reevaluated
before future rescue operations are initiated. The rationale behind
rescue operations has always been to save animals from unquestionable
death. However, animal mortality at hydroelectric dam sites is related
more to loss of habitat than to drowning. Several large mammals have
relatively good swimming capabilities (Child 1968). Kennerly (1963) has
even indicated the importance of swimming in ensuring gene flow between
populations separated by rivers. On the other hand, habitat loss is
irreversible. Kariba reservoir, the largest of all dams discussed above,
flooded an area of 5,462 km2 (ironically enough, its rescue operation
captured the least total number of animals, but most were ungulates
(Fig. 5-1)), Brokopondo flooded 1,683 km2, Tucurui 2,430 km2, Balbina
2,600 km2, and Samuel, the smallest of all, flooded 560 km2. Mammals,
amphibians, and some reptiles may survive by swimming to dry land, and
birds may fly away; however, the habitat lost can never be recovered.
Furthermore, animals captured during rescue operations are usually
released on the nearest piece of dry land, without any concern for the
animal community inhabiting the area of release. My results on mammal


102
frugivores have the lowest capture probability, and flycatchers the
highest (flycatchers in this study had low capture rates) (Brawn et al.
1995). Mist net data, therefore, most likely reflected the actual guild
structure of the forests.
The 3 guilds that most differed structurally among sites were the
bark-gleaner, and ant-follower insectivores, and the guild of birds
associated with water. The latter had a greater number of species at the
Jusante site than at the Reserve, however, the percentage of captures
were similar. According to Remsen and Parker (1983), species restricted
to river-created habitats comprise 15% of the total land bird fauna of
the Amazon Basin, therefore, the higher number of species at Jusante
might be a reflection of the greater diversity at the river edge. The
bark-gleaner insectivores had greater number of captures at the Reserve,
while the ant-followers were more abundant at Jusante, which may be due
to differences in forest structure. As discussed earlier, the Jusante
site has probably a richer understory, resulting in a greater abundance
of ants, which, in turn, would account for the higher abundance of ant
following birds. The Reserve, on the other hand, has larger DBH trees
(Chapter 2) which means older trees, and probably a greater number of
dead trees, which could account for the greater densities of bark-
gleaner insectivores.
Faunal Similarities
Members of the Pipridae family are often the most frequently
captured individuals in tropical areas (Brawn et al. 1995, Terborgh et
al. 1990, Remsen and Parker 1983, Robinson et al. 1989) and, despite
differences in species abundance, Pipra nattereri was the most


129
ORDER FAMILY SPECIES
Sittasomus griseicapillus
Furnariidae Phylidor erythrocercus
Phylidor pyrrhodes
Phylidor ruficaudatus
Xenops minutus
Sclerulus mexicanus
Sclerulus rufigularis
Hyloctistes subulatus
Automulus infuscatus
Automulus ochrolaemus
Hypocnemoides maculicauda
Synallaxis rutilans
Formicaridae Thamnophilus schistaceus
Thamnophilus doliatus
Thamnophilus aethiops
Thamnomanes caesius
Thamnomanes saturinus
Cercomacra nigrescens
Phlegopsis nigromaculata
Cymbilaimus lineatus
Hylophylax poecilonota
Hylophylax naevia
Hylophylax punctulata
Hypocnemis cantator
Formicarius colma


116
ORDER
FAMILY SPECIES
Rodentia
Cetacea
Carnivora
Myrmecophagidae
Sciuridae
Hydrochaeridae
Dasyproctidae
Echimydae
Muridae
Erethizontidae
Delphinidae
Platanistidae
Felidae
Cyclopes didactylus
Tamanda tetradactyla
Myrmecophaga tridactyla
Sciurus langsdorffi
Sciurus sp
Hydrochaeris hydrochaeris
Agouti paca
Dasyprocta fuliginosa
Echymys macrurus
Echymys grandis
Isothrix bistriata
Lonchothrix emiliae
Proechimys longicaudatus
Proechimys sp
Mesomys hispidus
Rhipidomys leucodactylus
Nectomys squamipes
Oryzomys concolor
Oryzomys megacephalus
Coendou prehensilis
Sphiggurus insidiosus
Sotalia fluviatilis
Inia geoffrensis
Felis concolor
Felis pardalis


To Joo Paulo and Daniela
My sources of strength and inspiration


EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD
COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST
By
ROSA MARIA LEMOS DE S
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
1995

Copyright 1995
by
Rosa Maria Lemos de S

To Joo Paulo and Daniela
My sources of strength and inspiration

ACKNOWLEDGMENTS
Most students consider themselves lucky if they find the right
academic advisor; I consider myself blessed for having found not one but
two! John Robinson was my first advisor, who helped define my work and
gave support during the initial field work. John Eisenberg took over
when Robinson left the University, and he has provided valuable advice
and moral support throughout this lengthy portion of my work. I thank
both of them for their help and guidance! Kent Redford was greatly
responsible for my interest in the University of Florida, and together
with his family made life in Gainesville even more enjoyable. His
knowledge and keen interest in the Neotropics has often motivated me,
and for that I am thankful. I am also extremely thankful to Ron Labisky,
Jay Malcolm, Richard Bodmer, Doug Levey, and Mel Sunquist for their
comments, advice, and editorial help. Special thanks go to Richard
Bodmer for agreeing to substitute for John Robinson at such short
notice.
Field work in Rondnia would not have been possible without the
technical support from ELETRONORTE. Several people have greatly helped
with the bureaucracy, allowing me to do field work: Edgar Menezes
Cardoso, Bruno Payolla, Rubens Guilhardi, and Carlos Fabbris. I also had
iv

excellent field assistance from Carry Ann Cadmam, Rodrigo Mariano,
Barroso, Antonio, and Chico. David Oren, Jos Maria Cardoso, and Chris
Canaday identified bird specimens and provided exceptional friendship.
This study was generously supported by the World Wildlife Fund -
WWF/US, Conservation International Cl, Lincoln Park Zoo "Scott
Neotropical Fund", Tropical Conservation and Development Program TCD,
Tinker Foundation, and the Program for Study in Tropical Conservation -
PSTC. Various individuals from these institutions simplified my life by
providing easy flow between finances and field work: Gustavo Fonseca,
and Sonia Rigueira from Cl, Cleber Alho, and LouAnn Dietz from WWF-US,
Steven Thompson from the Lincoln Zoo, and Kent Redford, Steven
Sanderson, and Peter Polshek from TCD. The Conselho Nacional de
Desenvolvimento Cientfico e Tecnolgico CNPq granted me a scholarship
for my academic work. The Empresa Brasileira de Pesquisa Agro-Pecuria -
EMBRAPA analyzed my soil samples. The Centro de Pesquisas para
Conservagao das Aves Silvestres CEMAVE/IBAMA supplied bird banding
permits and aluminum bands.
Life in Gainesville has been enriched by the ephemeral presence of
friends such as Jay Malcolm, Justina Ray, Joe Fragoso, Chris Canaday,
Wendy Townsend, John Payne, Ann Edwards, Miriam Marmontel, Andres
Navarro, Susan Walker, Denise Imbroise, Dener Martins, Peter Crawshaw,
Damian Rumiz, Cludio and Suzana Pdua, Rajanathan Rajaratnam, and many
others.
I am much grateful to Joao Paulo Viana for his endless help
throughout all phases of this work. His love, patience, and assistance
during the final weeks were specially cherished.
v

TABLE OF CONTENTS
ACKNOWLEDGMENTS iv
ABSTRACT viii
CHAPTERS
1 INTRODUCTION 1
Background 1
The Samuel Hydroelectric Power Plant 2
Rescue Operations... a History 4
The Samuel Rescue Operation 9
Research Design 10
2 CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION 11
Introduction 11
Methods 12
Study Site 12
Temperature and Precipitation 14
Nutrients in Soil 14
Vegetation 15
Results 15
Temperature and Precipitation 15
Nutrients in Soil 18
Vegetation 21
Canopy 21
Understory 24
Discussion and Conclusions 27
3 INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A
CONSEQUENCE OF DAMMING 30
Introduction 30
Methods 32
Study Site 32
Data Collection 32
Data Analysis 33
Biomass estimation 35
Results 35
Sightings per Kilometer Walked 35
Species Observed 37
Primates 38
Primate density estimates prior to damming 38
Individuals captured and released 39
Primate density estimates after damming 39
Density and body weight of primates 42
Total primate density estimates 42
Density changes between years at the Reserve 45
Biomass 48
Terrestrial Diurnal Mammals 52
Density estimates prior to damming 52
vi

Individuals captured and released 52
Density estimates after damming 53
Total terrestrial mammal density estimates 54
Biomass 56
Discussion and Conclusions 58
Primates 58
Density estimates 58
Density changes between years at the Reserve 66
Density comparisons with other western Amazonian sites 67
Total density 67
Biomass 69
Terrestrial Diurnal Mammal 74
Density 74
Total density 75
Biomass 77
4 UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL
DAM 78
Introduction 78
Methods 79
Study Site 79
Study Design 79
Data Collection 80
Data Analysis 81
Results 83
Inter-Year Comparison 83
Overall Site Comparison 84
Species Richness 86
Guild Structure 90
Faunal Similarity Between and Within Sites 93
Ecological Similarities Among Sites 97
Discussion and Conclusions 97
Overall Site Comparison 97
Species richness 97
Species abundance 101
Guild Structure 101
Faunal Similarities 102
Ecological Similarities 103
5 RECOMMENDATIONS FOR FUTURE HYDROELECTRIC DAM CONSERVATION PROGRAMS.105
Overview 105
Considerations 107
Rescue Operations 107
Creation of Conservation Units Ill
Preliminary Studies 112
Follow-up Studies 113
Conclusions 114
APPENDIX A MAMMALIAN SPECIES RECORDED AT THE SAMUEL DAM SITE 115
APPENDIX B BIRD SPECIES CAPTURED DURING THIS STUDY 118
APPENDIX C BIRD SPECIES RECORDED AT THE SAMUEL DAM SITE 124
LIST OF REFERENCES 134
BIOGRAPHICAL SKETCH 140
vii

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
EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD
COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST
By
ROSA MARIA LEMOS DE S
December 1995
Chairman: Dr. John F. Eisenberg
Major Department: Forest Resources and Conservation (Wildlife Ecology
and Conservation)
Two sites located near a hydroelectric dam reservoir, in the
northwestern region of Brazil, were monitored in order to document
changes in mammal and bird communities brought about by: a) the creation
of the reservoir, and b) by the release of rescued animals into one of
the study areas. Mammals were sampled with transect surveys, and
understory bird communities with mist nets.
Primate densities and biomass at the Samuel Ecological Station
(located at the southeast border of the Samuel Dam reservoir and
hereafter referred to as reserve) increased after the Samuel Dam flood
gates closed. Primate biomass in the reserve was 154 kg km'2 in 1988,
prior to damming, and increased to 165 ( 65) kg km'2 in 1989,
immediately after flooding, due to the release of animals captured in
the area flooded by the reservoir. However, because of shallow water,
most primates remained in the flooded forest for the first year since
the vegetation in the area was still alive. By 1990, all woody
viii

vegetation in the flooded area was dead, and primate biomass at the
reserve increased to 255 ( 109) kg km2, because of the migration of
animals from the reservoir to the reserve. By 1991, primate biomass in
the reserve returned to levels similar to 1988, 153 ( 81) kg km2, most
likely due to dispersal of animals to adjacent areas. The pattern of
biomass fluctuation indicates that primate carrying capacity at the
reserve is about 150 kg km"2.
Terrestrial mammal densities and biomass at the Samuel Ecological
Station increased in 1988, immediately after flooding, and remained high
until 1990 when they started decreasing.
Understory bird community structure and composition was sampled at
the Samuel Ecological Station, and at an area located at the Jamari
River's edge, downstream from the dam. The creation of the Samuel Lake
destroyed approximately 214 km of pristine riverine habitat, displacing
an entire community of birds. Despite the creation of a water-edge
habitat by the reservoir, the composition of bird species at the lake's
edge is different from that of the downstream site, therefore, different
from previously existing river edges at the reservoir site. However,
community structure seems to be similar at both sites. Water-related
niches were created at the reservoir's edge, yet, the uniqueness of
species composition at riverine habitats was not emulated.
ix

CHAPTER 1
INTRODUCTION
Background
Socioeconomic demands on land in Brazilian Amazonia are very high.
Land officially designated for development totals 2,100,000 km2, or 65%
of the total Amazonian area; of this area 4.4% is to be flooded by
hydroelectric development (Johns 1988). Eletrobras, the Brazilian
electric company, identified 80 potential dam sites in the Amazon region
in its Plan 2010 (Serra 1989). The objective of the agency is to
stimulate development in the region by attracting investors with
inexpensive energy sources.
Until 1980, only 2 small hydroelectric dams were operating in the
Amazon: Curu-Una, near Santarm, and Paredao, in Amap state. Each dam
impacted an area less than 100 km2 (Junk and Nunes de Mello 1987). Since
then, 3 large dams have been added to the region and are operating in
the Amazon; Tucurui, near Belm, Balbina, near Manaus, and Samuel, near
Porto Velho. Collectively, these 3 dams have flooded an area of 5,350
km2. If Eletronorte, the Brazilian Agency for Hydroelectric Power
Development in the Amazon region, succeeds in completing all the dams
projected for the Amazon in the 2010 plan, an area of roughly 100,000
km2 will be flooded (Fearnside 1989).
The flooding of such large areas has a tremendous impact on humans
and wildlife. The most significant effect is the loss of land, which
1

2
causes human and animal displacement and/or death, and can also bring
about extinction of species (Liao et al. 1988).
The Amazon region is well known for its high plant and animal
diversity. However, the distribution and densities of both plants and
animals are basically unknown in Amazonian regions. Despite their
negative impact on fauna and flora, hydroelectric dams provide good
opportunities for researchers to conduct detailed studies on local
distribution and densities of plants and animals if they are contacted
when the first feasibility studies begin, long before the creation of
the reservoir. Instead, the power companies invite scientist to research
the area 1 or 2 years prior the completion of the project, yielding only
a short-term evaluation. As a result of this policy, very little was
learned of the impact on wildlife resulting from the 3 dams constructed
most recently.
This study documents the impacts on avian and mammalian
communities created by the construction of the Samuel Hydroelectric
Power Plant on the Jamari River, Rondnia, Brazil.
The Samuel Hydroelectric Power Plant
The Samuel Hydroelectric Power Plant is located on the Jamari
River, a right bank tributary of the Madeira River, Rondnia, 8 45'S -
63 25'W. The site is 52 km east of the city of Porto Velho, the state's
capital, and 96 km from the confluence with the Madeira river (Fig. 1-
1). The Jamari River basin is entirely located in the State of Rondnia;
its watershed is located between 8 28' 11 07'S, and 62 36' 63
57'W. Its head waters are located in the Pacas Novos Mountain chain at

3
Figure 1-1: Location of the Samuel Oam on the Jamari River

4
an altitude of approximately 500 m; its total extension is approximately
560 km.
The reservoir, with a volume of 3.2 billion m3, occupies 560 km2
at normal maximum operating level of 87 m above sea level (Eletronorte
1990), and has the capacity of generating 216 MW of energy. The
reservoir, lying in a southeastern orientation, is 40 km in length; its
width spans 15 to 20 km in the first 25 km, and between 3 and 1 km for
the remaining 15 km. The construction of the hydroelectric power plant
started in 1982, and because of the flat terrain characteristic of the
region, 57 km of dikes were constructed along the right and left banks
in order to contain the river water (Fig. 1-2).
To guarantee conservation of the local fauna and flora,
Eletronorte created the Samuel Ecological Station, 20,865 ha in size,
adjacent to the area of influence by the Samuel Dam. However, roughly
20% of the Reserve was lost due to inundation after completion of the
reservoir (Mozeto et al. 1990, Eletronorte 1993) (Fig. 1-3). The
vegetation cover at the Reserve, before the flooding, consisted of 96.6%
terra firme, 2% temporarily flooded forest, and 1.4% secondary forest
(Eletronorte 1993). The only road into the Reserve area parallels the
dam (Fig. 1-3). Hunting does not occur within the Reserve because of
controlled access.
Rescue Operations... a History
To ameliorate the effects of habitat loss on wildlife, Eletronorte
employed "rescue operations" to save terrestrial animals once the
reservoir began to fill. Drowning animals is bad publicity, which

5
Figure 1-2: Satellite image of the Samuel Dam Reservoir under
construction (From Instituto Nacional de Pesquisas Espaciis -
INPE, July 1987).

6
Figure 1-3: Satellite image of the Samuel reservoir, and
the Samuel Ecological Station (From Instituto Nacional
de Pesquisas Espaciis INPE, August 1992).

7
undoubtedly was a relevant factor in the decision by Eletronorte to
promote a rescue operation. However, the rescue operations created a
moral dilemma among Brazilian and foreign scientists. Was it worth
spending millions of dollars on such operations, which did not guarantee
the survival of the animals? Or was it more profitable, for
environmental conservation, to use the money to create new reserves and
to further maintain the already existing ones?
Eletronorte ignored such questions, and initiated the first rescue
operation at Tucurui in 1984-85; a total of 284,211 animals (vertebrates
and invertebrates) were captured and translocated (Eletronorte 1985).
Unfortunately, Eletronorte had not conducted a preliminary study to
define appropriate sites for release of the captured animals; thus, they
were released on the nearest piece of dry land (Johns 1986).
The Tucurui rescue operation cost US $30 million and employed 300
people (Johns 1986). The cost of rescuing 222,544 vertebrates was
$134.80 per individual; considering only the rescue of 107,094 birds and
mammals, the cost was $280.13 per individual. Furthermore, based on
crude estimates of primate densities, only 4% of the tamarins (Saguinus
midas), 6.4% of squirrel monkeys (Saimir sciureus), 6.9% of capuchin
monkeys (Cebus apella), 4.2% of saki monkeys (Chiropotes satanas), and
29.7% of howler monkeys (Alouatta belzebul) were rescued (Johns 1986).
According to Johns
these results suggest that rescue operations will remove only a
small proportion of primates. It is likely that there will be
critical overpopulation of lake fringe areas caused by the vast
majority of animals escaping unaided, which suggests that
releasing captured animals on the lake shore is worse than
useless. In fact, the real value of rescue operations is called
into question. (Johns 1986, pg. 20)

8
Despite the unsuccessful results of the Tucurui rescue operation,
a similar rescue operation was carried out by Eletronorte at Balbina in
1987. Although the area of Tucurui and Balbina were similar, the number
of animals rescued at Balbina was only about 10% of the number rescued
at Tucurui (Gribel 1990). The Balbina reservoir is much shallower, which
limited boat travel and thus, reduced the number of animals rescued. As
at Tucurui, the rescued animals were released with the same carefree
attitudes as before, and most likely did not survive the pressures of
high densities and hunting.
Only in 1988, after much pressure from the scientific community in
Brazil, did Eletronorte change its rescue operation policies. A new
rescue plan was developed because it was obvious that releasing animals
at random did not benefit the animals rescued or the community into
which they were released.
Museums, research institutions, universities, and zoological
gardens were contacted throughout Brazil, before the Samuel Dam rescue
operation started, and offered live specimens, study skins, and display
skins of species rescued in the Samuel dam reservoir. Priority was given
to research institutions that could provide housing facilities for the
animals, and to well-known researchers willing to work at the Samuel
site during the operation. Most importantly, an ecological station was
created by Eletronorte to be the focus of conservation studies, and the
only site of animal release.
Several inventories were carried out inside the Reserve prior to
the flooding of the reservoir, including vegetation, mammal, reptile and
bird surveys, and soil analyses. The projects were conducted in an
attempt to understand the community before the translocation of the

9
rescued animals, so that a follow-up would bring some understanding to
the changes brought about by the dam construction.
The Samuel Rescue Operation
The rescue operation started in November 1988, immediately after
the closing of flood gates. The objective of this operation was to
recover every animal found stranded on small islands created by the
filling of the reservoir. The operation lasted 4.5 months, during which
time 16,000 animals were rescued. The rescued animals included 6,590
arthropods, 3,729 mammals, 3,504 reptiles, 2,099 amphibians, and 78
birds (Eletronorte 1989). Of the 16,000 animals rescued, 11,417 were
sent alive to research institutions, and 1,729, which were sacrificed or
died during the operation, were sent to research institutes or museums.
Only 2,854 were released inside the Reserve. The number of animals
released at the Samuel Ecological Station was significantly less than
the number of animals released at previous dam sites. However, these
animals, combined with the animals that moved on their own to the
Reserve site, have probably caused some impact on the populations
previously inhabiting the site.
The influence of the dam and its reservoir is not a one-time
phenomena. During the dry season, the reservoir shrinks to an estimated
40% of its fullest extent. The approximately 300 km2 area which will be
exposed each year has the potential of altering adjacent biological
communities in a major way.

10
Research Design
In order to document the changes that occurred after the
translocation of animals, I conducted research at the Samuel Ecological
Station during the months of June to October of 1989, 1990 and 1991. My
objectives were to sample both large diurnal mammals and understory
passerine birds both inside and outside the Reserve area.
My original experimental design included a control site for the
Reserve. The control was needed to represent mammal densities, and bird
community composition and structure in the Reserve prior to any flooding
because I started my work after the filling of the reservoir. However,
it was not possible to find an accessible area with the same
characteristics as the Reserve that was not affected by the flooding. I
then modified my original design and sampled an area downstream from the
dam (Jusante), which had similar characteristics to the area that was
flooded by the reservoir. Therefore, Jusante was a control site for the
area flooded by the reservoir, and not the Reserve.
This work represents the first attempt to understand community
changes brought about by hydroelectric dams in the Amazon. There is
still a lot of work to be done; however, I hope that the results of my
research will stimulate similar projects at future dam sites, and that
it will influence Eletronorte officials in their future environmental
decisions. Eletronorte should consider environmental impact studies and
rescue operations more seriously and make decisions based on scientific
facts rather than public indulgence.

CHAPTER 2
CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION
Introduction
Amazonia occupies nearly 6,000,000 km2, with more than half of
this in Brazilian territory. Viewed from the air, the Amazon forest
appears quite homogeneous; however, when examined in detail,
considerable local variations of vegetation and floristic composition
are encountered (Pires and Prance 1985). The central plateau of the
Amazon Basin is limited to the north and to the south by moderate slopes
and to the west by the Andes, opening only to the east where it receives
hot and humid winds from the Atlantic. These unique characteristics plus
its latitudinal position give the Amazon region singular conditions,
such as almost constant daylength throughout the year, unchanged solar
energy at the limit of the earth's atmosphere, and little variation in
average monthly temperatures (Salati 1985).
Precipitation in the Amazon basin varies from 1,500 to > 3,000 mm
annually (IBGE 1977). A well defined dry season (from May through
September) is common in the central region.
Due to effects of high temperatures, high rainfall, and geology of
the region, the soils of Amazon rainforest have low potential for
supplying nutrients to plants. Intense weathering and leaching over
millions of years have removed the nutrients from the minerals which
form the parent material of the soil (Jordan 1985).
11

12
Forest physiognomy in the Amazon basin is influenced not only by
soil but also by the age of the vegetation at specific sites (Lisboa
1990) Water levels in the past and in the present have great influence
in the formation of the vegetation. Younger vegetation formations,
present today, are located in areas that were submerged in the past.
The objective of this chapter is to characterize the climate,
soil, and vegetation in the area, in order to understand the differences
and/or similarities of the study sites and their associated fauna.
Methods
Study Site
The Samuel Hydroelectric Dam is located in the state of Rondnia,
50 km east of the state capital of Porto Velho. Two sites were selected
for the study. The first site was the Samuel Ecological Station with an
area of 21,000 ha, located at the southeast border of the Samuel
Reservoir approximately 26 km (straight line) from the dam. The Reserve
was created by Eletronorte (the Brazilian Agency for Hydroelectric Power
Development in the Amazon Region) in an attempt to establish a protected
site for the animals rescued from the reservoir area during flooding.
The second site was located approximately 3 km below the dam, and will
be referred to as Jusante. The Jusante site was comparable,
floristically as well as faunistically, to 50% of the area that was
flooded by the reservoir (Fig. 2-1).
Within each of these sites, 3, 1 km2 plots were used to sample
soil and vegetation (Fig. 2-1). Plots at Jusante were deliberately

Figure 2-1: The Samuel Dam Reservoir area showing study sites (Reserve and Jusante), and study
plots (1, 2, 3, 4, 5, A, B, C, and D) .

14
located 500 m from the River's edge in order to avoid flooded forest,
which has a unique but different set of vegetative characteristics.
Temperature and Precipitation
Temperature and precipitation data were collected by the
Engineering Department at the Samuel Hydroelectric Dam. The
climatological station at Samuel is located at 8 45'S 63 28'W, at an
altitude of 80 m, and has been operating since July 1977 (Eletronorte
1988a) .
Daily average temperature was recorded with the use of a maximum-
minimum temperature thermometer. Rainfall was recorded on a daily basis
with a rain gauge.
Nutrients in Soil
Soil samples were collected in 6, 1 km2 plots (plots 1, 2, 3, A,
B, and C; Fig. 2-1) Samples were taken along a straight line every 100
m for 1 km in each of the plots, totaling 11 stations in each plot. Soil
was collected, using a soil auger, at depths of 0-20 cm, 20-40 cm, and
40-60 cm in each of the stations, which yielded 33 samples/plot. Care
was taken to collect soil from a non-disturbed area (therefore top soil
would be intact), and to avoid contamination of lower depth samples by
upper samples.
Samples were kept in marked plastic bags and taken to Embrapa's
(Empresa Brasileira de Pesquisa Agropecuria) soil analysis laboratory
in Porto Velho for pH and nutrient (phosphorus [P], potassium (K],
calcium [Ca], aluminun [Al], and magnesium [Mg]) analysis within 2 days
of collection.

15
Vegetation
Vegetation samples were taken for purpose of characterizing the
vegetation structure at both sites. A 1-km transect was sampled in each
of the 6 plots (soil and vegetation were sampled along the same 1-km
strip). The point-centered quarter method was used to sample large trees
(Knight 1978), which divides the area around a sample point into 4
imaginary quarters. Measurements of distance were taken from each point
to the center of the nearest tree in each quarter. Points in the
transect were 20 m apart, totaling 51 stations, and 204 trees measured
per transect. Measurements recorded included DBH (diameter at breast
height), and height (estimated). Only trees > 10 cm DBH were included.
Understory vegetation was sampled within 5 x 5 m quadrats at
alternate stations in each of the transects, totaling 26 quadrat sample
stations per transect. Within each quadrat all trees with DBH ranging
from 1 to 10 cm were measured, and all vegetation with DBH < 1 cm
(seedlings) was counted.
Results
Temperature and Precipitation
The average annual monthly temperature in the Samuel region is
27C; however, lower monthly temperatures are recorded during June,
July, and August due to a phenomenon called friagem, which is a cold
front sweeping over the continent from Antarctica (Table 2-1, Fig. 2-2).
This event, which lasts for only a few days, will lower temperatures
considerably, especially during nighttime.

Temperature (C)
16
Table 2-1: Average monthly temperature
(C) at the Samuel Dam, during 1989,
1990, and 1991 (source Eletronorte).
Month
rear
1989
1990
1991
January
27.0
27.2
26.7
February
26.0
26.7
26.8
March
27.0
26.4
26.4
April
27.0
26.6
27.0
May
27.0
26.7
26.6
June
26.5
26.1
26.1
July
25.0
25.7
24.4
August
27.0
27.0
25.5
September
28.0
27.0
26.6
October
27.0
26.4
27.3
November
28.0
28.0
27.8
December
28.0
28.0
27.4
Annual
average
26.9
26.8
26.6
Figure 2-2: Average monthly temperature (C),
at the Samuel Dam region for 1989,1990, and 1991

17
Total annual precipitation in the Samuel region ranges from 2,000 .
to 2,600 mm year. June, July, and August are characteristically dry, and
the dry season lasts from May to September. The rainy season is also
well defined by the months of October to April, when monthly
precipitation usually exceeds 200 mm (Table 2-2, Fig. 2-3).
Table 2-2: Total monthly precipitation
(mm) in the Samuel Dam, during 1989,
1990, and 1991 (source Eletronorte).
Month
Year
1989
1990
1991
January
302.4
400.0
326.9
February
319.4
521.0
223.1
March
248.0
341.0
412.9
April
316.6
147.0
315.1
May
165.8
133.0
100.4
June
107.1
47.0
38.7
July
59.2
24.0
0.0
August
59.9
39.0
0.0
September
83.7
146.3
42.2
October
136.0
274.4
142.6
November
89.9
318.0
233.2
December
275.0
244.0
306.9
Annual
total
2,163.0
2,634.7
2,142.0

18
Figure 2-3: Total monthly Precipitation (mm),
at the Samuel Dam site for 1989,1990, and 1991.
Nutrients in Soil
Surface soils were extremely acid in all plots. The soils were
less acid at progressively lower depths. Amounts of phosphorus (P),
potassium (K), calcium (Ca), magnesium (Mg), and aluminum (Al) decreased
from surface to lower depths. Levels of P, Ca, and Mg were low, whereas
levels of K ranged from low to medium. The levels of Al, however, were
high in all plots and at all depths (Tables 2-3, and 2-4 ). Among sites,
levels of Al were different at all depths (ANOVA, n = 2, d.f. = 1,
p < 0.0264, 0.0108, and 0.0100, for 0-20, 20-40, and 40-60 cm depth,
respectively), and pH was different at 0-20, and 20-40 cm depth (ANOVA,
n = 2, d.f. = 1, p < 0.0450, and 0.0239, respectively). Levels of P, K,
Ca, and Mg did not differ among sites.

Table 2-3: Average pH and nutrients per area, per category of depth
at the Samuel Ecological Station ( s.d.). Depth (cm), pH (H20)
P and K (ppm), Ca, Mg, and A1 (mEq/100 ml).
AREA
DEPTH
pH
P
1C
Ca
Mg
Al
01
0-20
3.85 (0.2)
2.55 (0.8)
47.73 (18.1)
0.20 (0.1)
0.16(0.1)
2.07 (0.4)
20-40
4.30 (0.4)
1.27 (0.5)
27.45 (22.9)
0.07 (0.1)
0.13 (0.1)
1.46 (0.3)
40-60
4.48 (0.2)
1.36 (0.7)
14.00 (22.3)
0.05 (0.1)
0.14(0.1)
1.15(0.2)
02
0-20
3.78 (0.1)
1.73 (0.6)
52.82 (26.7)
0.15(0.1)
0.18(0.1)
1.83 (0.3)
20-40
4.27 (0.3)
1.00 (0.0)
40.00 (35.6)
0.12(0.1)
0.22(0.1)
1.28 (0.4)
40-60
4.40 (0.2)
1.27 (0.5)
18.91 (12.2)
0.11 (0.0)
0.17(0.1)
1.19(0.2)
03
0-20
4.08 (0.2)
1.36(0.5)
43.36 (16.9)
0.75 (0.4)
0.52 (0.2)
0.92 (0.2)
20-40
4.33 (0.3)
1.00(0.0)
26.91 (20.9)
0.20(0.1)
0.29(0.1)
0.79 (0.2)
40-60
4.52 (0.4)
1.09(0.3)
35.36 (54.0)
0.19(0.1)
0.30 (0.1)
0.68 (0.3)
Note: Reference values for pH and nutrients (Source: EMBRAPAs Laboratory of Soil Analysis).
pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid
P: 0-10 ppm = low
11-30 ppm = medium
> 30 ppm = high
K: 0 45 ppm = low
46- 150 ppm = medium
>150 ppm = high
Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high
Al: 0.0 0.3 mEq = low
> 0.3 mEq = high

Table 2-4: Average pH and nutrients per area, per category of depth
at Jusante ( s.d.)* Depth (cm), pH (H20) P and K (ppm), Ca, Mg,
and A1 (mEq/100 ml).
APEA
DEPTH
pH
P
K
Ca
Mg
Al
A
0-20
3.45(0.1)
3.27(1.4)
49.18(14.3)
0.13 (0.1)
0.17(0.1)
3.12(0.4)
20-40
3.73 (0.2)
1.36 (0.5)
14.73 ( 6.8)
0.04 (0.1)
0.12(0.1)
2.79 (0.3)
40-60
4.05 (0.2)
1.09 (0.3)
7.73 ( 2.6)
0.00 (0.0)
0.11 (0.1)
2.58 (0.1)
B
0-20
3.51 (0.2)
2.82 (1.3)
37.54 ( 8.2)
0.09(0.1)
0.15(0.1)
2.65 (0.8)
20-40
3.86 (0.2)
1.36 (0.7)
11.27 ( 2.3)
0.05(0.1)
0.12 (0.1)
2.12 (0.3)
40-60
4.15 (0.3)
1.73 (1.3)
10.09 ( 7.3)
0.02(0.1)
0.14(0.1)
1.90 (0.2)
C
0-20
3.71 (0.3)
5.18(2.7)
63.45 (28.4)
0.02 (0.0)
0.18(0.1)
2.92 (1.0)
20-40
4.11 (0.2)
2.64 (3.5)
29.91 (25.5)
0.01 (0.0)
0.11 (0.1)
2.41 (0.9)
40-60
4.43 (0.2)
1.27 (0.6)
13.09 ( 8.1)
0.01 (0.0)
0.11 (0.1)
2.13(0.9)
Note: Reference values for pH and nutrients (Source: EMBRAPAs Laboratory of Soil Analysis).
pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid
P: 0-10 ppm = low
11-30 ppm = medium
> 30 ppm = high
K: 0 45 ppm = low
46- 150 ppm = medium
> 150 ppm = high
Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high
Al: 0.0 0.3 mEq = low
> 0.3 mEq = high

21
In summary, both Jusante and Reserve sites had low pH values and
low levels of nutrients in the soil; however, the soil at Jusante had a
lower pH and a higher A1 content than at the Reserve indicating more
acid soils were characteristic of Jusante.
Vegetation
Canopy
Average DBH of trees was 22.3 cm (SD =14.8, n = 612) and 18.9 cm
(SD = 10, n = 612) at the Reserve and Jusante, respectively; 20.6% of
trees at the Reserve had diameters £ than 30 cm, compared to only 11.8%
at Jusante (Fig. 2-4). DBH differed between the 2 sites (Chi-Square, p <
0.006, n = 1224). The findings from the Reserve are in agreement with
those of Martinelli et al. (1988), who reported a mean DBH of 21.2 cm
with 16.6% of the trees with DBH > 30 cm from a survey prior to the
filling of the Reservoir.
Total basal area was larger at the Reserve than at Jusante; 24.7
m2, and 18.8 m2, respectively (n = 612 in both areas). The difference
was greater if only trees > 30 cm in DBH were considered; 16.7 m2 (n =
126) and 8.6 m2 (n = 72) in the Reserve and at Jusante, respectively.
The height of trees also differed between sites (Chi-Square, p <
0.0001, n = 1224). Despite the fact that average tree height at the
Reserve was only slightly greater than at Jusante, 13 m (SD =4.2, n =
612), and 12 m (SD =3.1, n = 612), respectively, 5.4% of trees at the
Reserve were i 21 m compared to 1.47% of trees at Jusante (Fig. 2-5).
Tree height distribution at the Reserve was also in conformity with
findings from Martinelli et al. (1988), with trees predominantly falling

Percent of trees (%) Percent of trees (%)
22
DBH (cm)
DBH (cm)
Figure 2-4: Percentage of trees in each DBH (cm) category
at the Samuel Reserve and Jusante.

Percent of trees (%) Percent of trees (%)
23
Height (m)
Height (m)
Figure 2-5: Percentage of trees in each height (m) category
at the Samuel Reserve and Jusante.

24
in the categories between 10-15 m. However, average tree height found by
Martinelli et al. (17.8 m) was greater than in this study. The
difference may be due to different techniques of measuring tree height;
I estimated height visually, whereas Martinelli et al. used a clinometer
and a measuring tape. However, because both the Reserve and Jusante were
sampled similarly in this study, comparisons between categories of
height between sites should be valid.
Total tree density at the Reserve was 441.3 trees/ha (SD = 17.9,
n = 612), and 523.6 trees/ha (SD = 22.4, n = 612) at Jusante. However,
considering only trees with DBH ^ 30 cm, the density was higher at the
Reserve (90.9/ha; SD = 34.7, n = 126) than at Jusante (61.6/ha;
SD = 30.1, n = 72). Martinelli et al. (1988) reported similar findings
for the Reserve, total density was 483 trees/ha, and 80/ha for trees
l> 30 cm.
Understory
Average DBH for trees < 10 cm in both areas was very similar, 3.3
cm at the Reserve and 3.4 cm at Jusante, and did not differ in their
distribution (Chi-square, d.f. =8, p = 0.172, n = 573 and 798,
respectively) (Fig. 2-6). Similarly, average understory tree height did
not differ between sites; being 4.6 m and 5.1 m for the Reserve and
Jusante, respectively (Chi-Square, d.f. =7, p = 0.096, n = 573 and 798)
(Fig. 2-7). Seedling density was almost identical at the 2 sites, 15,015
individuals/ha (SD = 7,762, n = 78) at the Reserve, and 15,385
individuals/ha (SD = 7,953, n = 78) at Jusante. However, for trees with
DBH between 1 and 10 cm, density was 28% lower at the Reserve (2,933
individuals/ha; SD = 1,416, n = 78) than at Jusante (4,092

Percent of trees (%) Percent of trees (%)
25
DBH (cm)
DBH (cm)
Figure 2-6: Percentage of trees in each DBH (cm)
category sampled in the understory vegetation
at the Reserve and at Jusante.

Percent of trees (%) Percent of trees (%)
26
Height (m)
Height (m)
Figure 2-7: Percentage of trees in each height (m)
category sampled in the understory vegetation at the
Reserve and at Jusante.

27
individuals/ha; SD = 1,975, n = 78). Basal area was also lower at the
Reserve than at Jusante (3.4 m2 and 5.2 m2, respectively).
A striking difference in understory structure between the 2 sites
was the number of stemless palms. At the Reserve, the number of
individual stemless palms was 234, 65% of which belonged to the species
Orbignya barbosiana (babagu) which is of great economic importance in
the region as a source of oil. At Jusante, however, only 57 individual
stemless palms were counted, and only 8.8% of those were 0. barbosiana.
On the other hand, the palm-like species Phanakospermum guianenses
(sororoca, Musaceae) was recorded 58 times in the Reserve, but 469 times
at Jusante.
Discussion and Conclusions
The Amazon forest is heterogeneous both in the large numbers of
species within each community type, and in the large numbers of
community types in a given area. Structurally, the Reserve and Jusante
are different. DBH and tree height distribution revealed that the
Reserve has a higher and denser canopy than Jusante. Basal area and tree
density, for trees £ than 30 cm, were also much greater at the Reserve.
According to Pires and Prance,
similar types of vegetation have approximately the same biomass.
Biomass can be expressed by the basal area of trees per hectare,
using individuals of 30 cm or more in circumference [approximately
10 cm in diameter]. On this basis, the exceptionally large forests
can exceed 40 m2 of basal area. The open forests or vine forests
usually are between 18 and 24 m2. (Pires and Prance 1985, pg. 112)
Vine forests in the Marab region, on the Tocantins river, have basal
areas between 18-22 m2 (Pires 1984). In this study, basal area was 18.8
m2 at Jusante and 24.7 m2 at the Reserve, indicating that the forest

28
type at the Samuel Dam region is the open or vine forest, commonly found
in Rondnia (Pires 1984). Furthermore, species most common in open
forests, according to Pires (1984), were described by Martinelli et al.
(1988) at the Reserve. The absence of epiphytes in the area, another
characteristic of open forest, can be related to the occurrence of a
marked dry season.
The lower basal area at Jusante may be due to the age of the
forest. According to Lisboa (1990), younger vegetational types are
located in areas that were submerged in the past. Elevation varies
between 60 and 100 m at Jusante, and between 90 and 150 m at the
Reserve. This difference in elevation supports the idea that the forest
at Jusante is younger, probably due to disturbance effects related to
its proximity to the river.
The more open forest at Jusante allows for greater penetration of
light, providing an opportunity for shrub and liana species to develop,
which creates a forest floor more densely covered by vegetation. The
denser vegetation is exemplified by the higher density of understory
trees in the 1-9 cm DBH category at Jusante, and the abundance of the
low growth musaceous palm-like species, Phanakospermum guianense, which
was encountered 88% more on the forest floor of Jusante than at the
Reserve.
Even though this study did not record species composition,
differences in the study sites were visually detected. The presence of
adult Bertholletia excelsia (Brazil nut), and Orbignya barbosiana
(babagu) were very.common at the Reserve, but rare at Jusante. On the
other hand, the species, Hevea brasiliensis (seringueira), was very
common at Jusante but never seen at the Reserve; the latter observation

29
was expected because this particular species is known to occur along
water courses and not in terra firme.
The reason that the soils are so low in nutrients is because they
have been subjected to the intense weathering of the tropical climate
for many millions of years. However, a very slight difference in soil
quality within an area could be reflected in an entirely different
forest community through very fine adaptations of each community to
subtle differences in the soil (Jordan 1985). The latter could explain
the differences in species composition between the study sites.

CHAPTER 3
INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A
CONSEQUENCE OF DAMMING
Introduction
The importance of tropical rain forests to global biodiversity is
clearly appreciated when one realizes that they cover only 7% of the
earth's land surface, but contain more than half the species of the
world's biota (Wilson 1988). Despite the importance of tropical forests,
and the fact that very little is known about their fauna and flora,
development of tropical areas is occurring at a rapid pace, and will
bring about the extinction of species. To avoid mass extinction and to
be able to guide developing agencies, a better understanding of the
communities and their responses to environmental changes is needed. The
purpose of this chapter is to document the response of mammalian
communities to environmental changes resulting from the construction of
the Samuel Hydroelectric Dam in the Amazon.
Two sites were monitored after the filling of the Samuel
Hydroelectric Dam reservoir. In the first site (a reserve), animals
captured during a rescue operation were released. This site was
monitored during 1988 (before the flooding of the reservoir), 1989,
1990, and 1991. The second site (referred to as Jusante) was an
undisturbed area located downstream from the dam, on the right bank of
the Jamari river, and it was monitored during the 1990 and 1991 field
seasons.
30

31
My assumption at the beginning of this study was that the
mammalian community in the Reserve could have been affected in three
ways: 1) by the release of rescued animals, 2) by the migration of
animals fleeing from the flooded reservoir (I have used the term
migration as defined by Baker 1977 throughout the text), and 3) by a
combination of both. During the rescue operation, from November 1988 to
March 1989, 2,374 mammals were released inside the Reserve (Eletronorte
1989). In addition to the release of rescued animals in the Reserve, I
expected migration of animals from the reservoir to occur, because the
forest at the reservoir was continuous with the forest at the Reserve. I
hypothesized that the Reserve would experience animal overcrowding for
an undetermined length of time, possibly surpassing the carrying
capacity for the area. My hypothesis could be tested by estimating
mammalian densities in the Reserve at different points in time and by
examining differences in biomass values for the community.
The Jusante site, on the other hand, could only have been affected
by the migration of animals from the reservoir area (because there was
no release of animals in the area), or not affected at all.
If my assumptions were correct, the noted responses in density
changes, regardless of site, would be immediate in the case of
terrestrial mammals (because they would have to flee from the rising
water), but possibly delayed for arboreal species (because they could
stay on top of trees while the vegetation was still alive). The time
frame in which density changes would occur was unknown. To increase the
probability of documenting such changes (completely or partially), the
sites were sampled repeatedly.

32
Methods
Study Site
The study was conducted in the Samuel Dam region located on the
Jamari river, in the state of Rondnia, approximately 50 km east of the
state's capital of Porto Velho (Fig. 1-1).
Prior to the filling of the reservoir, Eletronorte (the government
agency responsible for hydroelectric dam constructions in northern
Brazil) created a 21,000 ha Reserve (Estago Ecolgica de Samuel) to
compensate for the loss of 56,000 ha of forest due to the creation of
the reservoir. The Reserve is located east of the reservoir's
embankment, approximately 26 km from the dam. The other study site
(referred to as Jusante) was located 3 km below the dam, on the right
bank of the Jamari river (Fig. 2-1).
Data Collection
Five plots of 1 km2 were established in the Reserve in 1989, while
three plots were established at Jusante in 1990 (Fig. 2-1), creating 4
km of transect lines along each plot's perimeter. Transect surveys were
conducted by walking slowly (1 km/h), and stopping periodically to watch
and listen. Transects were conducted between 0630-0700 hours (depending
on the time of sunrise) and 1030-1100 hours in the morning, and between
1300 to 1700 hours in the afternoon. The number of transect samples in
each area was divided equally between morning and afternoon surveys.
Whenever possible, different transects were walked in the morning and
afternoon of the same day. If that was not, possible due to logistics,
the two daily surveys of a plot always began in the same direction to

33
give an interval of six hours between the morning and the afternoon
survey (i.e. the same point in the trail would be traversed in the
afternoon six hours after the morning survey). Transect surveys on
different days began at opposite ends of the route to reduce potential
biases resulting from direction of travel by the observer. Each transect
had equal numbers of surveys originating in both directions. The time,
transect identification, location on the trail, species, number of
individuals sighted, angle of sighting, and distance from the observer
to the animal when first seen was recorded for every non-flying mammal
encountered. Surveys were conducted by myself, an assistant, and two
field helpers.
The study was conducted from May to August 1989 and from May to
October of 1990 and 1991. These months correspond to the dry season in
the region (refer to Chapter 2 for detail). During 1989, due to
logistical problems and to the lesser amount of time spent at the study
site, data were gathered only at the Samuel Reserve.
Two of the five plots sampled in 1989 in the Reserve were
abandoned in 1990 and 1991 due to logistic difficulties. However,
observations collected in these areas were included in the 1989 analysis
to arrive at a density estimate for the entire site. Data from the
various plots at a site were pooled within years to give an overall
density estimate. The number of kilometers walked per site, plot, and
year are listed on Table 3-1.
Data Analysis
Data were analyzed using the computer program TransAn, version
1.00. TransAn is a flexible computer program that uses a non-parametric,

34
shape-restricted, density estimator (Payne 1992). The shape-restricted
estimator, first introduced by Johnson and Routledge (1985) and later
modified by Fyfe and Routledge (1991), involves modeling the probability
of detecting an individual as a function of its perpendicular distance
from the transect line.
TransAn requires sightings from at least four independent transect
lines to calculate confidence limits. Because I had only three transects
per site, the data were divided between morning and afternoon transects
to increase the number of transects to six (for the 1989 data the total
number of transects was 10, because there were five different plots at
the Reserve). Despite inherent biases, the transect censuses are
currently the most cost effective method to evaluate large mammal
densities in rainforests (Emmons 1984).
Table 3-1: Total number of kilometers walked during surveys, per
site, plot, and year.
KM WALKED PER YEAR
SITE
PLOT
1989
1990
1991
TOTAL
Reserve
1
48
88
84
2
48
96
84
3
48
80
84
4
36


5
36


Total
216
264
252
732
Jusante
A

80
84
B

80
84
C

80
84
Total

240
252
492
GRAND TOTAL
1,224

35
Biomass estimation
Crude biomass was calculated using average adult body weight (BW)
and population density (D) of a species: BW*D (kg km2). The average
animal weight used for biomass calculations was obtained from data
gathered by biologists working at the rescue operation during 14
November 1988 to 29 March 1989 (Eletronorte 1989) Because the rescue
operation was conducted as the reservoir was being flooded, and because
flooding was accomplished within four months (a relatively short period
of time), I assumed that the average animal weight recorded reflected
the true weight of the animals in the wild before any stressful
situation could cause weight loss.
Metabolic biomass also was calculated. It is an important
measurement because species sampled varied greatly in size, and
metabolic biomass takes into account that energy expenditure increases
allometrically with body weight to the power of 0.75 (Peters 1983). It
was then calculated as BW'75*D (kg0'75 Km2) The density, biomass, and
body weight estimates were ln-transformed to linearize the data.
Results
Sightings per Kilometer Walked
The number of individuals sighted per km in the Reserve was very
similar for all three years. The mean number of individuals sighted per
km walked in the Reserve was 3.34 (SD = 0.49; n = 269 sightings) in
1989, 3.59 (SD = 0.47; n = 304 sightings) in 1990, and 3.34 (SD = 0.39;
n = 296 sightings) in 1991 (Fig. 3-1). At Jusante, the mean number of
individuals sighted per km walked was 43% lower in 1991 than in 1990:

36
Year
5
- 4.5
H-
3
a
kms walked
ind/km
J
Figure 3-1 :Number of kilometers walked and number of individual
animals seen per kilometer walked at the Reserve and at Tusante
during the study.
3.90 ( 1.80) ind/km in 1990, and 2.24 ( 0.72) ind/km in 1991 (n = 272
and 224 sightings, respectively; Fig. 3-1).
To evaluate bias in the sampling method, all sightings were
plotted according to the location on the trail where species were seen.
The result was an even distribution of sightings and species in each
area and in all three years. Therefore, there was no observer bias as to
where the animals were observed along the transects. There also was no
difference in the number of observations between morning and afternoon
censuses, indicating that animal sightings were independent of time of
day (Chi-square test = 6.85, 4 d.f., p= 0.05).

37
Species Observed
The number of species recorded at the
similar for all years. Sixteen species were
1990, and 17 in 1991. At Jusante the number
Reserve during census was
recorded in 1989, 18 in
of species seen was 19, in
Table 3-2: List of species observed during all transect surveys for both
sites. NR = number of sightings for 1989, 1990, and 1991 at the Reserve
NJ = number of
sightings for 1990
and 1991 at Jusante.
ORDER
FAMILY
SPECIES
NR
NJ
Primata
Cebidae
Aotus azarae
01
01
Callicebus bruneus
29
98
Pithecia irrorata
62
37
Cebus apella
197
59
Saimiri ustus
25
24
Ateles paniscus
165
05
Callithrichidae
Callithrix emiliae
34
27
Saguinus fuscicollis
58
62
Edentata
Myrmecophagidae
Tamanda tetradactyla
02
05
Myrmecophaga tridactyla
01
00
Bradypodidae
Choloepus didactylus
00
03
Dasypodidae
Cabassous sp.
00
02
Carnivora
Procyonidae
Nasua nasua
10
12
Mustelidae
Eira barbara
11
01
Lutra longicaudis
00
01
Felidae
Felis pardalis
01
01
Panthera onca
01
00
Perissodactyla
Tapiridae
Tapirus terrestris
03
02
Artiodactyla
Tayassuidae
Tayassu tajacu
11
03
Tayassu pcari
01
00
Cervidae
Mazama americana
Mazama guazoubira "
67
49
Rodentia
Sciuridae
Sciurus sp.
10
15
Dasyproctidae
Dasyprocta fuliginosa
180
89

38
1990 and 15 in 1991; a 16% decrease. Table 3-2 lists all species seen
during transect surveys in both sites during all three years. The list
of species seen is only a fraction of the total number of mammalian
species in the area and only represents the medium to large size
mammalian community (refer to Appendix A for a complete species list).
Of the 24 species seen during census, only 7 primate species, 2
deer, and the agouti had sample sizes large enough to estimate density.
Primates
Primate density estimates prior to damming
Primates censuses were performed at the Reserve by Eletronorte
researchers from September 1987 to February 1988 (Table 3-3, and Figure
3-2). Density estimates were based on 145 km of transect surveys.
Techniques used were comparable to the ones used in this study (National
Research Council 1981) .
Table 3-3: Primate density estimates at the Reserve prior
to the flooding of the reservoir. D = group density,
MGS = mean group size (Eletronorte 1988).Saimir were
not recorded during the 1988 censuses.
SPECIES
D
MGS
IND/KM2
Ateles paniscus
2.60
5.2
13.5
Cebus apella
4.02
5.4
21.7
Pithecia irrorata
1.38
3.8
5.2
Callicebus bruneus
0.26
2.6
0.7
Saimir ustus



Saguinus fuscicollis
2.00
8.5
17.0
Callithrix emiliae
0.36
15.0
5.4

39
Individuals captured and released
Primates represented 48.4% of all mammals captured, and 47% of all
animals released during the rescue/release operation (Eletronorte 1989).
The number of individuals per species captured during the operation, and
later released at the Reserve is shown in Table 3-4 (refer to Chapter 1
for details of operation). This was a relatively small number and
probably had an insignificant effect on most density shifts during the
years following flooding. Many changes in density derived form
emigration (see below).
Table 3-4: Number of primates captured at the Samuel
Dam reservoir, and number of primates released at the
Samuel Ecological Station (Eletronorte 1989).
SPECIES
NUMBERS
CAPTURED
NUMBERS
RELEASED
Aotus
104
60
Ateles
35
35
Callithrix
71
42
Saguinus
171
76
Cebus
207
180
Callicebus
348
309
Pithecia
369
324
Saimir
501
326
TOTAL
1, 806
1,352
Primate density estimates after damming
The densities of Ateles, Callithrix, and Saimir were high in 1990
and lower and approximately equal in 1989 and 1991. Cebus and Saguinus
densities were also high in 1990 in the Reserve, however, their

40
densities in 1991 remained high and similar to the 1990 densities
instead of returning to values comparable to 1989. The densities of
Callicebus and Pithecia in the Reserve were at their highest in 1989,
and decreased steadily through 1990 and 1991(Table 3-5, Fig. 3-2).
With the exception of Pithecia, whose density was similar in 1990
and 1991, all other primate densities decreased substantially from 1990
to 1991 at the Jusante site (Table 3-5, Fig. 3-2).
Ateles paniscus. In 1989, the density of Ateles in the Reserve was
3.15 groups/km2 (13.2 individuals/km2), whereas in 1990 their density
increased to 6.06 groups/km2 (23.6 ind/km2). By 1991 their group density
was at 3.69 per km2 (higher than in 1989). However, the number of
individuals per km2 was 11.1 (less than in 1989) due to the fact that
the mean group size went from 4.2 in 1989 to 3.0 in 1991 (Table 3-5,
Fig. 3-2). At Jusante the density of Ateles was estimated at 0.60
groups/km2 (5.9 ind/km2) in 1990. In 1991 only one individual was
sighted during 252 km of census, making it impossible to estimate
density (Table 3-5, Fig. 3-2).
Callithrix emiliae. The density of Callithrix in the Reserve more
than doubled from 1989 to 1990, 1.41 groups/km2 (4.1 ind/km2) to 3.06
groups/km2 (10.1 ind/km2), respectively. By 1991, however, the density
had decreased to 1.82 groups/km2 (3.6 ind/km2), similar to 1989 (Table
3-5, Fig. 3-2). The density of Callithrix at Jusante decreased by 71%
from 1990 to 1991; from 3.13 groups/km2 (9.1 ind/km2) in 1990, to 0.91
groups/km2 (2.8 ind/km2) in 1991 (Table 3-5, Fig. 3-2).
Saimir ustus. The density of Saimiri in the Reserve increased
four fold from 1989 to 1990, then decreased by more than half by 1991.
The density in 1989 was 0.91 groups/km2 (6.9 ind/km2), by 1990 it was

41
3.68 groups/km2 (28.7 ind/km2), and by 1991 it had decreased to 1.41
groups/km2 (8.9 ind/km2) (Table 3-5, Fig. 3-2). The density of Saimir at
Jusante decreased by four fold if we consider the number of individuals
per km2. In 1990 the density was 1.69 groups/km2 (23.8 ind/km2), and by
1991 it was 0.75 groups/km2 (6.0 ind/km2) (Table 3-5, Fig. 3-2).
Cebus apella. Cebus density in the Reserve went from 5.42
groups/km2 (20.1 ind/km2) in 1989 to 6.63 groups/km2 (28.5 ind/km2) in
1990. The density in 1991 was similar to 1990 with 6.45 groups/km2 (27.1
ind/km2) (Table 3-5, Fig. 3-2). At Jusante Cebus density decreased from
1990 to 1991, going from 2.92 groups/km2 to 2.12 groups/km2. The
difference is greater if we consider the number of individuals per km2:
17.5 in 1990, and 8.1 in 1991. This difference is due to the fact that
mean group size went from 6.0 in 1990 to 3.8 in 1991 (Table 3-5,
Fig. 3-2).
Saguinus fuscicollis. The density of Saguinus in the Reserve more
than doubled from 1989 to 1990, going from 1.85 groups/km2 (6.1 ind/km2)
in 1989 to 4.08 groups/km2 (14.3 ind/km2) in 1990. Saguinus maintained a
high density in 1991: 3.7 groups/km2 (14.1 ind/km2 ) (Table 3-5, Fig. 3-
2). At Jusante there was a decrease in Saguinus density from 1990 to
1991: 5.83 groups/km2 (20.4 ind/km2) in 1990, to 3.67 groups/km2 (13.6
ind/km2) in 1991 (Table 3-5, Fig. 3-2).
Callicebus bruneus. The density of Callicebus in the Reserve was
at its highest in 1989 with 3.61 groups/km2 (7.2 ind/km2). In 1990 the
density had decreased to 0.83 groups/km2 (1.3 ind/km2), and by 1991 it
was 0.62 groups/km2 (1.2 ind/km2)(Table 3-5, Fig. 3-2). At Jusante the
density of Callicebus went from 8.33 groups/km2 (20.0 ind/km2) in 1990,
to 4.85 groups/km2 (11.6 ind/km2) in 1991 (Table 3-5, Fig. 3-2).

42
Pithecia irrorata. Pithecia also had its highest density in the
Reserve in 1989, and then decreased thereafter. The density was 3.43
groups/km2 (10.3 ind/km2)in 1989, 2.07 groups/km2 (5.2 ind/km2) in 1990,
and 1.17 groups/km2 (3.5 ind/km2) in 1991(Table 3-5, Fig. 3-2). At
Jusante the density of Pithecia remained similar from 1990 to 1991: 2.60
groups/km2 (7.5 ind/km2), and 2.38 groups/km2 (7.6 ind/km2),
respectively (Table 3-5, Fig. 3-2).
Density and body weight of primates
When broad geographic regions are examined, differences in density
among primate species are related to body size. In general, population
densities decline with increasing body mass (Clutton-Brock and Harvey
1977, 1979; Eisenberg 1979; Robinson and Redford 1986; Kinnaird and
Eisenberg 1989). In the primate community at the Samuel Dam region, the
correlation between density and body weight was positive for primates in
the Reserve in 1989. This positive correlation is a contradiction of the
rule, because smaller animals should have higher densities than larger
ones. There was no significant correlation between density and body
weight in 1990 and 1991 (Fig. 3-3).
The correlation between density and body weight for primates at
Jusante was not significant in either year, but showed a tendency to be
negative in 1990, (Fig. 3-4).
Total primate density estimates
Primates at the Reserve comprised 61% of all mammalian sightings
in 1989, 69% in 1990, and 68% in 1991. When data for all primates are
pooled and density estimates were calculated for the area as a whole,
the result shows a 32% increase in group density from 1989 to 1990, and
a 23% decrease from 1990 to 1991 (Table 3-6, Fig. 3-5A) .

43
Table 3-5: Primate density estimates at the Reserve and at Jusante.
N = number of
sightings
, D =
group
density
, MGS
= mean group size.
SPECIE
SITE
YEAR
N
D
MGS
IND/XM1
95% D Cl
Ateles
Reserve
89
21
3.15
4.2
13.2
1.53-07.43
paniscus
Reserve
90
72
6.06
3.9
23.6
3.95-09.91
Reserve
91
66
3.69
3.0
11.1
2.14-07.06
Jusante
90
04
0.60
9.8
5.9
0.17-02.63
Jusante
91
01

1.0

Cebus
Reserve
89
54
5.42
3.7
20.1
3.21-09.22
apella
Reserve
90
71
6.63
4.3
28.5
4.15-10.31
Reserve
91
72
6.45
4.2
27.1
4.01-10.32
Jusante
90
32
2.92
6.0
17.5
1.59-06.15
Jusante
91
27
2.12
3.8
8.1
0.98-04.57
Pithecia
Reserve
89
33
3.43
3.0
10.3
1.76-06.59
irrorata
Reserve
90
17
2.07
2.5
5.2
0.75-05.06
Reserve
91
12
1.17
3.0
3.5
0.35-02.49
Jusante
90
21
2.60
2.9
7.5
1.16-05.57
Jusante
91
16
2.38
3.2
7.6
1.08-06.39
Callicebus
Reserve
89
15
3.61
2.0
7.2
1.57-08.34
bruneus
Reserve
90
05
0.83
1.6
1.3
0.20-03.03
Reserve
91
09
0.62
2.0
1.2
0.24-01.74
Jusante
90
59
8.33
2.4
20.0
4.77-14.72
Jusante
91
39
4.85
2.4
11.6
2.49-09.89
Saimir
Reserve
89
08
0.91
7.6
6.9
0.25-02.43
ustus
Reserve
90
05
3.68
7.8
28.7
0.89-05.68
Reserve
91
12
1.41
6.3
8.9
0.49-03.64
Jusante
90
15
1.69
14.1
23.8
0.62-04.55
Jusante
91
09
0.75
8.0
6.0
0.29-02.58
Saguinus
Reserve
89
14
1.85
3.3
6.1
0.71-03.88
fuscicollis
Reserve
90
24
4.08
3.5
14.3
1.92-09.18
Reserve
91
20
3.70
3.8
14.1
1.51-06.53
Jusante
90
34
5.83
3.5
20.4
2.88-11.23
Jusante
91
28
3.67
3.7
13.6
1.74-07.79
Callithrix
Reserve
89
10
1.41
2.9
4.1
0.38-03.82
emiliae
Reserve
90
15
3.06
3.3
10.1
1.31-06.69
Reserve
91
09
1.82
2.0
3.6
0.74-06.00
Jusante
90
20
3.13
2.9
9.1
1.34-05.97
Jusante
91
07
0.91
3.1
2.8
0.28-03.75

Groups/kilometer square
44
7
6
5
4
3
2
1
0
Cebus
7
6
5
4
3
2
1
0
Callithrix
1
7 n
6
5 -
4 -
3 -
2 -
1 -
0
Saimir
n
i
oo
cr>
o

00
CD
O
r
OO
CD
O
00
oo
CD
CD
OO
OO
CD
CD
OO
OO
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
7
6
5
4
3
2
1
0
Saguinus
\
\
_N
\
\
\
\
\
\
00
CD
O
t
00
CD
O
OO
OO
CD
CD
OO
00
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
9
8
7
6
5
4
3
2
1
0
Callicebus
J=t-
7
6
5
4
3
2
1
0
Pithecia
00
CD
O
T
00
CD
O
,
00
00
CD
CD
00
00
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
'

*
^
^
<
T
s
Year
Figure 3 2: Primate density estimates for the Samuel Ecological Station
during 1988 (Eletronorte 1989), 1989, 1990, 1991 (this study), and the
Jusante in 1990, and 1991 (this study).

45
Primates at Jusante comprised 71 and 58.5% of all mammal sightings
for 1990 and 1991, respectively. Primate group density for the area as a
whole decreased 44% from 1990 to 1991 (Table 3-6, Fig. 3-5A). Even
though a decrease in primate density occurred in both areas from 1990 to
1991, the decrease at Jusante was almost twice that of the Reserve. The
changes in densities of ind/km2 show the same pattern as the group
density changes, however, Jusante shows a more abrupt reduction in total
number of individuals than the Reserve (Fig. 3-5B).
Density changes between years at the Reserve
A cluster analysis comparing density results for all 4 years of
data for the Reserve shows that the 1988 densities had only a 25.27
degree of similarity with the 1989 densities (all comparisons excluded
Saimir because this species was not recorded during 1988) The 1990
community still only shows a 27.94 degree of similarity with 1988. By
1991, the degree of similarity with the 1988 community, increased to
61.92 (Fig. 3-6).
Table 3-6: Total primate densities for the Reserve and Jusante
for 1989, 1990, and 1991. N = number of sightings, D = group
density, MGS = mean group size (based on total number of
individuals sighted and total number of sightings per year).
SITE
YEAR
N
D
MGS
IND/KM2
95% Cl
Reserve
1989
161
17.59
3.6
63.3
13.86-22.81
Reserve
1990
209
23.25
3.9
90.7
18.92-29.02
Reserve
1991
200
18.03
3.6
64.9
13.71-23.68
Jusante
1990
185
24.51
4.4
107.8
17.62-33.46
Jusante
1991
127
13.83
3.5
48.4
8.86-23.14

Log10 density (ind/km2)
46
O Ateles A Cebus Pithecia Sai
Callicebus <0> Saguinus V Callithrix
1 .6 i- 1991
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
O

V *
O
i
2.0 2.5 5.0 5.5 4.0
Log10 body weight (g)
Figure 3-3: Relationship between density and body weight for
primates at the Samuel Ecological Station. (P = 0.0156 for
1989, 0.6338 for 1990, and 0.5726 for 1991).
Log10 body weight (g)
Figure 3-4: Relationship between density and body
weight for primates at Jusante. (P = 0.2135, and
0.6974 for 1990 and 1991 respectively). See Figure
3-3 for legends.

Ind/km2 Groups/knr
47
Reserve
Jusante
Figure 3-5: A) Total primate density (groups/km2) for the Reserve,
and Jusante for 1989, 1990, and 1991. B) Total primate density
(ind/km2) for the Reserve, and Jusante for 1989, 1990, and 1991.

48
25.27 -
3U.I0
75.09 -
19
88 19
91 IS
Years
90 19
89
Figure 3-6: Cluster analysis for all four years of density
data at the Reserve. Similarity levels = 61.92, 27.94, and
25.27, respectively (Minitab 10, Hierarchical cluster
analysis of observations).
Biomass
In the Reserve in 1989, the three largest primates (Ateles, Cebus,
and Pithecia, respectively) were also the three species that contributed
most to the crude and metabolic biomass. However, by 1990 Saimir
contributed more than Pithecia, despite weighing only a third of the
weight of Pithecia. By 1991, Cebus was the species that contributed most
to the metabolic biomass calculations, followed by Ateles and Saimir.
The total primate crude biomass for the three years showed a 55%
increase from 1989 to 1990, and then a decrease of 40% by 1991,
returning to levels similar to 1989 (Table 3-7). The total primate crude
biomass for 1988 was 154 kg/km2 (Eletronorte 1989); a value very similar
to the 1989 and 1991 values (Fig. 3-7).

49
There was a positive correlation between body weight and biomass
(kg km2) at the Reserve in all three years (Fig. 3-8). This correlation
is well documented (Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979;
Robinson and Redford 1986).
At Jusante, Cebus contributed most to the metabolic biomass
calculations in both years. Ateles was in second place in 1990, followed
by Saimir. By 1991, only one group of Ateles was recorded at Jusante
and the contribution by Saimir had decreased drastically, leaving
Pithecia as the second largest contributor. Total primate crude biomass
decreased by 61% from 1990 to 1991 (Table 3-8). The correlation between
biomass and body weight was also positive in both years at Jusante
(Fig. 3-9).
Table 3-7: Body weight, crude biomass, and metabolic biomass of primates
at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size.
SPECIES
BODY
WEIGHT
(kg)
CRUDE
BIOMASS
(kg km*)
METABOLIC
BIOMASS
(kgkm)
1989
1990
1991
1989
1990
1991
Ateles
6.299(29)
83.15
148.66
69.92
52.48
93.84
44.13
Cebus
2.304(142)
4 6.31
65.66
62.4 4
37.59
53.30
50.68
Pithecia
2.102(284)
21.65
10.93
7.36
17.98
9.08
6.11
Callicebus
0.798(279)
5.75
1.04
0.96
6.08
1.10
1.01
Saimir
0.739(277)
5.10
21.21
6.58
5.50
22.88
7.09
Ssguiuus
0.329(128)
2.01
4.70
4.64
2.65
6.21
6.13
Csllithrix
0.318(61)
1.30
3.21
1.14
1.74
4.28
1.52
TOTAL
165.27
255.41
153.04
124.02
190.69
160.80
(65.29)
( 109.37)
( 81.30)
< 44.75)
< 73.28)
( 56.10)

Crude Biomass (kg/km2)
50
Reserve
Jusante
Figure 3-7: Total primate crude biomass for the Samuel
Ecological Station during 1988 (Eletronorte 1989), 1989,
1990, and 1991 (this study), and for Jusante during
1990 and 1991 (this study).

Log10 biomass (g/km2)
51
O Ateles A Cebus Pithecia Saimir Callicebus O Saguinus V Callithrix
Log10 body weight (g)
Figure 3-8: Relationship between biomass and body weight for
primates at the Samuel Ecological Station. (P = 0.0001,
0.0403, and 0.0289 for 1989, 1990, and 1991, respectively).
Log10 body weight (g)
Figure 3-9: Relationship between biomass and body
weight for primates at Jusante. (P = 0.0131,
and 0.0237 for 1990, and 1991, respectively). See
Figure 3-8 for legends.

52
Table 3-8: Body weight, crude biomass, and metabolic biomass of primates
at Jusante during 1990, and 1991. ()= sample size.
SPECIES
BODY WEIGHT
(kg)
CRUDE
(kg km-*)
BIOMASS
METABOLIC
(kg0'15km2)
BIOMASS
1990
1991
1990
1991
Ateles
6.299(29)
37.16
23.46
Cebus
2.304(142)
40.32
18.66
32.73
15.15
Pitheda
2.102(284)
15.77
15.98
13.09
13.27
Calllcebus
0.798(279)
15.96
9.26
16.89
9.79
Saimir
0.739(277)
17.59
4.43
18.97
4.78
Saguinus
0.329(128)
6.71
4.47
8.86
5.91
Callithrix
0.318(61)
2.89
0.89
3.85
1.19
TOTAL
136.40
53.69
117.85
50.07
( 36.91)
( 15.8)
( 28.99)
( 13.55)
Terrestrial Diurnal Mammals
Of all the diurnal terrestrial mammals that occur in the region,
only the agouti and the two species of deer had sample sizes sufficient
for density estimation (Tab. 3-2). Because the deer species are
sometimes difficult to identify when seen for a brief moment moving in
the understory, and to increase sample size, the two species were pooled
(Mazama sp) for the analysis.
Density estimates prior to damming
During the 1987-1988 censuses at the Reserve, the density of
Dasyprocta fuliginosa was estimated at 3.33 ind/km2, and Mazama sp. at
0.34 ind/km2 (Fig. 3-10; Eletronorte 1989).
Individuals captured and released
Three hundred and three individuals of Dasyprocta fuliginosa were
captured during the rescue operation; 214 of those individuals were

53
released at the Reserve. Twenty-three individual Mazama sp. were
captured, and 19 were released at the Reserve (Eletronorte 1989).
Density estimates after damming
Dasyprocta fuliginosa. Agouti density at the Reserve was at its
highest in 1989 with 21.7 ind/km2. It decreased thereafter, declining to
12.1 ind/km2 in 1990, and to 7.3 ind/km2 in 1991 (Table 3-9, Fig. 3-10).
At Jusante, agouti density had a slight increase from 1990 to 1991: 4.3,
and 5.0 ind/km2, respectively (Table 3-9, Fig. 3-10).
Mazama sp.. Deer density in the Reserve increased 40% from 1989 to
1990, then decreased 22% from 1990 to 1991, returning to levels similar
to 1989. Density went from 3.3 to 4.6 to 3.6 ind/km2, respectively
(Table 3-9, Fig. 3-10). At Jusante, however, there was a 74% increase in
density from 1990 to 1991, going from 1.9 to 3.3 ind/km2, respectively
(Table 3-9, Fig. 3-10).
CM
E
jr
V)
Cl
3
o
o
20
18
16
14
12
10
8
6
4
2
0
Dasyprocta
\
\
\
10 -i
8
6
4
2
0
Mazama
J=^
]Reserve
y\v| Jusante
\
\
00
cr>
o
T
00
cn
O
t
oo
oo
CT)
CT)
oo
00
cn
CD
cn
CT>
cn
CD
CD
cn
cn
CD
.
.
*
^
.
.
*
*
Year
Figure 3-10: Density estimates for Dasyprocta and Mazama at the Samuel
Ecological Station for 1988 (Eletronorte 1989), 1989, 1990, and 1991
(this study), and at Jusante in 1990 and 1991 (this study).

54
Table 3-9 Terrestrial diurnal mammal density estimates at the Reserve
and at Jusante. N = number of sightings, D = group density, MGS = mean
group size.
SPECIE
SITE
YEAR
N
D
MGS
IND/KM2
95% Cl
Dasyprocta
Reserve
89
67
19.73
1.1
21.7
12.72-33.68
fuliginosa
Reserve
90
63
12.12
1.0
12.1
7.56-21.37
Reserve
91
50
7.29
1.0
7.3
4.60-11.82
Jusante
90
42
4.31
1.0
4.3
2.36-07.27
Jusante
91
47
4.96
1.0
5.0
2.81-08.38
Mazama sp.
Reserve
89
24
3.26
1.0
3.3
1.41-05.29
Reserve
90
18
4.58
1.0
4.6
1.98-09.77
Reserve
91
25
3.45
1.0
3.6
1.55-06.95
Jusante
90
16
1.92
1.0
1.9
0.76-05.37
Jusante
91
33
3.04
1.1
3.3
1.52-06.06
Total terrestrial mammal density estimates
Because only the agouti and the two deer species had sample sizes
large enough to calculate densities, all terrestrial mammal sightings
were pooled to arrive at a terrestrial mammal density for the Reserve as
a whole. Sightings for terrestrial mammals comprised 39, 31, and 32% of
all mammal sightings at the Reserve for 1989, 1990, and 1991,
respectively. From 1989 to 1990, the mean group density decreased only
slightly, however, from 1990 to 1991 the total terrestrial mammal
density at the Reserve decreased 28%. The number of ind/km2 shows a more
obvious decrease in density due to differences in mean group sizes
through the years (Table 3-10, Fig. 3-11).
At Jusante the sightings for terrestrial mammals comprised 29 and
41.5% of all sightings for 1990 and 1991, respectively. Group density
showed only a slight increase from 1990 to 1991 and the density of
terrestrial mammals a slight decline (Table 3-10, Fig. 3-11) .

55
Table 3-10: Total terrestrial mammal densities for the Reserve and
Jusante for 1989, 1990, and 1991. N.SP = number of species used in
calculations, N = number of sightings, D = group density, MGS = mean
group size (based on total number of individuals sighted and total
number of sightings per year). one sighting of 21 individuals of
T. pcari was eliminated from the MGS calculations in order to avoid
unrealistic increase in MGS.
SITE
TEAR
N.SP
N
D
MGS
IND/KM2
95% Cl
Reserve
1989
07
102
21.12
1.4
29.6
14.98-32.72
Reserve
1990
09
92
20.68
1.3*
26.9
14.87-31.07
Reserve
1991
09
93
14.93
1.3
19.4
10.96-21.41
Jusante
1990
08
74
8.65
1.4
12.1
5.34-14.13
Jusante
1991
08
90
8.86
1.2
10.6
5.79-12.16
Figure 3-11: Total terrestrial mammal density (groups/km2 and
ind/km2) for the Reserve, and Jusante for 1989, 1990, and 1991.

56
Biomass
At the Reserve, agouti biomass decreased 44% from 1989 to 1990,
and another 40% from 1990 to 1991. The biomass of the deer species had a
40% increase from 1989 to 1990, and then a 22% decrease in 1991 (Table
3-11). At Jusante, agouti biomass increased slightly from 1990 to 1991
(16%), however, the deer species had a 74% increase in biomass for the
same period. (Table 3-12).
There was a 688.5% increase in Mazama and Dasyprocta biomass from
1988 to 1989 (Fig. 3-12). From 1989 to 1990 the biomass remained high,
but by 1991 it had decreased 26.5%, although it was still much higher
than the 1988 biomass.
Table 3-11: Body weight, crude biomass, and metabolic biomass of agouti
and deer at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size, average weight for both species (adapted from Bodmer
1989).
SPECIES
BODY
WEIGHT (kg)
CRUDE
BIOMASS
(kg km*)
METABOLIC
BIOMASS
(kgkm*)
1989
1990
1991
1989
1990
1991
Dasyprocta
2.721(278)
59.05
32.92
19.86
45.97
25.63
15.47
Mazama sp.
20.000*
66.00
92.00
72.00
31.21
43.50
34.05
TOTAL
125.05
124.92
91.86
77.18
69.13
49.52
< 13.14)
< 5.7)
< 0)
( 10.24)
< 4.45)
< 0)
Table 3-12: Body weight, crude biomass, and metabolic biomass of
agouti and deer at Jusante during 1990, and 1991. ()= sample size,
* average weight for both species (adapted from Bodmer 1989).
SPECIES
BODY HEIGHT
(kg)
CRUDE
(kg km'*)
BIOMASS
METABOLIC
(kg 75km'2)
BIOMASS
1990
1991
1990
1991
Dasyprocta
2.721(278)
11.70
13.61
9.11
10.59
Mazama sp.
20.000*
38.00
66.00
17.97
31.21
TOTAL
49.70
79.61
27.08
41.80
< 1.65)
( 14.08)
( 1.29)
( 7.07)

Crude Biomass (kg/km2)
57
Reserve
Jusante
Figure 3-12: A: Biomass for both Mazama and Dasyprocta at the Reserve
and at Jusante (Eletronorte 1988, this study). B: Biomass for
Mazama and Dasyprocta at Jusante plotted separately (this study).

58
Discussion and Conclusions
The results suggest that the environmental changes created by the
construction of the dam altered the mammal community in the areas
adjacent to the reservoir.
Primates
Density estimates
Three different patterns can be seen with primate density changes
at the Reserve (Fig. 3-2). The first one, seen with Ateles, Callithrix,
and Saimir, is a large increase in density from 1989 to 1990, and then
a sharp decrease in 1991, returning to density levels similar to those
found in 1989. The second pattern, which involves Cebus, and Saguinus,
also is an increase in density from 1989 to 1990, however, the 1991
densities remaining high. The third pattern is a very different
situation, with the densities of Callicebus, and Pithecia being at their
highest in 1989, and declining in 1990, and 1991.
The increase in densities from 1989 to 1990 can be explained by
the migration of animals from the reservoir into the Reserve. Sixty
percent of the reservoir is, on average, only 3.5 m deep, which allowed
several tree species to survive for at least 8 months after the flooding
began. I observed flowering and fruiting trees inside the reservoir in
August 1989 (5 months after the completion of the filling of the
reservoir). Child (1968) observed the same phenomenon during the
formation of Lake Kariba as a result of the impoundment of the Zambezi
river in Zimbabwe. At Lake Kariba, species had different survival times
that varied from 4 to 12 months. He also observed that "most species
standing in water came into leaf and/or remained in leaf until they

59
died" (Child 1968, pg. 37). Because primates spend most of their time in
the mid to upper forest strata, it is reasonable to assume that the vast
majority of the primate population was still living inside the reservoir
when the 1989 survey was carried out at the Reserve. This assumption can
be corroborated by the fact that only 1,806 primates were captured
during the rescue operation in a 56,000 ha area (0.03 ind/ha)
(Eletronorte 1989). In contrast, at the Tucurui Dam site (which has a
much deeper reservoir), a total of 27,039 primates were captured in a
243,000 ha area (0.11 ind/ha) (Eletronorte 1985).
When I arrived at the Samuel Dam in May 1990 all vegetation inside
the reservoir was dead. The only exception was in the higher elevation
lands, which formed green islands inside the reservoir. Because the
reservoir is bordered by dikes on both sides, and by a paved highway on
the left river bank, a natural escape route for displaced animals was
the Reserve (Fig. 1-3, and 2-1) Thus, the increase in primate densities
in the Reserve was likely a result of the natural migration of animals
into the Reserve between August 1989 and May 1990, caused by the loss of
habitat inside the reservoir.
This scenario does not explain the third pattern of density
changes detected in Callicebus and Pithecia, whose population declined
steadily after 1989 However, Callicebus, Pithecia, and Saimir were
the most frequently captured species, with over 300 individuals released
(Table 3-3). This suggests that the 1989 census in the Reserve
documented the density increase in these species as a result of the
release operation. This argument is even more convincing when density
estimates from a study done in 1988 (prior to the flooding), is
examined (Fig. 3-2).

60
The densities for Ateles and Saguinus were almost identical in
1988 and 1989. Because only a few individuals were released in the
Reserve, there was no reason to expect otherwise. The higher 1989
estimates for Saimir, Callicebus, Pithecia, and to some extent Cebus,
reflect the increase in density caused by the released animals. Because
my study began only two months after the release was completed, it is
reasonable to assume that the animals were still inside the Reserve, and
that is why the 1989 densities were higher than the 1988. As for
Callithrix, the difference between the 1988 and 1989 densities may be a
reflection of the species characteristics. Callithrix are among the most
cryptic primates, in both pelage and habitat (Ferrari and Rylands 1994),
making them difficult to detect during transect samples. Because the
1988 censuses were not performed by my team, their low density might be
a consequence of differences in researcher's detection ability. Despite
having similar body weight, the same detectability differences do not
apply to Saguinus, "while relatively similar in body size and most
aspect of their ecology, S. fuscicollis invariably prefers lower forest
strata them its congeners" (Ferrari and Rylands 1994, pg. 82), which
makes them more visible during a census.
The decline in density estimates in 1991 for all species is most
likely a consequence of the dispersal of animals to adjacent areas or
death. The Reserve is located adjacent to an area of continuous forest,
without human inhabitation or access roads (Fig. 1-3), and dispersal
into those areas would be the expected behavior for animals in a
situation of crowding.
There was a sharp decline in density (between 37 and 71%) from
1990 to 1991 at Jusante with all primates, except Pithecia (Fig. 3-2).

61
The decline in density is not very apparent for Cebus and Saimir if we
only consider the number of groups per km2, however, these two species
were the only ones to show drastic reduction in mean group sizes over
the years (Table 3-5). Hence, if we consider the decline in the number
of individuals per km2, both species also show drastic decreases in
densities from 1990 to 1991.
Even though there was no sampling at Jusante during 1988 or 1989,
it is logical to assume a similar effect on primate communities in both
areas, as a consequence of the creation of the reservoir.
A buffer area for the protection of the dam turbines was created
inside the reservoir by clear cutting the forest closest to the dam,
which together with several construction projects and the concentration
of human activities near the dam could have inhibited animal migration
to Jusante. However, migration did occur, probably due to sheer
proximity of the area to the reservoir (animals stranded inside the
reservoir near the dam could probably see green forest on the other
side) (Fig. 2-1).
Despite the lack of data for Jusante in 1989, density estimates
for 6 of the 7 primates sampled at Jusante will most likely fit the
first pattern of density change described for primates at the Reserve
(density increase in 1990, followed by a sharp decrease in 1991) The
seventh species, Pithecia, fits the second pattern of density change
(density stays at similar levels from 1990 to 1991). Because no data
exist for 1989, it is not possible to determined if the third pattern
described at the Reserve (a continue decrease in density) was present at
Jusante. However, because the explanation given for this pattern in

62
density change was the active release of animals in the Reserve, such
pattern would not be expected to appear at Jusante.
Differences in crude densities and in the degree of density
decline among species and sites are most likely due to differences in
species behavior and/or habitat requirements.
Ateles. Ateles are frugivore-herbivores (Eisenberg 1981, Robinson
and Redford 1986, 1989), with 83 to 90% of their diet consisting of
fruits and the remainder of other plant parts (van Roosmalen and Klein
1988). Because the distribution of fruits in a forest is widely
scattered, Ateles density is probably restricted by the availability of
this food type (Robinson and Ramirez 1982). Home-range size increases as
group weight increases (Eisenberg 1979), and Ateles are the largest of
all primates in the area requiring, in Surinam, 12.2 ha per individual
(van Roosmalen 1980, Robinson and Janson 1987). The new arrivals at the
Reserve were most likely displaced to areas outside the Reserve by the
resident groups due to the unavailability of fruit crops large enough to
maintain the higher population density.
Their almost complete absence from Jusante can be explained by the
fact that they are restricted to, or occur in higher densities only in
primary forest, using upper levels of canopy and emergent trees
(Mittermeier and van Roosmalen 1981, Robinson and Ramirez 1982, van
Roosmalen and Klein 1988). Because the forest at Jusante was
characterized by having lower, more open canopy with fewer emergent
trees (Chapter 2), it is not surprising to find that the Reserve
represented a more suitable habitat for the species (Fig. 3-2).
Callithrix. Callithrix are insectivore-omnivores, with more than
50% of their diet consisting of invertebrates (Eisenberg 1981, Robinson

63
and Redford 1986, 1989). They are also adapted to feed on plant exudates
at certain times of the year to compensate for seasonal scarcities in
the availability of fruits (Ferrari and Lopes Ferrari 1989, Rylands and
Faria 1993, Ferrari 1993). They attain highest densities in second
growth forest and edge habitat.
Callithrix species tend to have larger group sizes and smaller
home-ranges than Saguinus species, and generally occur at higher
densities (Ferrari and Lopes Ferrari 1989, Rylands and Faria 1993).
Average group size and densities of Callithrix in both of my study sites
was lower than that of Saguinus (Table 3-5, Fig. 3-2), in contrast to
previous studies. This may be partly a consequence of their crypticity,
as described earlier.
Saimir. Saimir are classified as frugivore-omnivores, with more
than 50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989). They are habitat specialists, typical of flooded and
riverine forests (Eisenberg 1979, Freese et al 1982, Rylands and
Keuroghlian 1988). The species is known for its preference for more
open, secondary habitats, and they are most often encountered in liana
forests (Mittermeier and van Roosmalen 1981, Johns and Skoruppa 1987).
Neither of the study sites, the Reserve or Jusante, included flooded
forests. Despite the fact that the Jusante site is closer to the Jamari
river, and has a more open forest structure, transect censuses started
at a distance of 500 meters away from the river's edge. Therefore, high
densities of Saimir were not expected at either site. The high
densities in the Reserve in 1990, as well as the very high number of
individuals per km2 at Jusante (due to a larger mean group size; Table

64
3-5), probably occurred when animals living along the Jamari river
inside the reservoir moved to these areas in search of new suitable
habitat. Because neither ai;ea is suitable habitat, the animals most
likely dispersed along the Jamari river, causing the density decrease
seen in 1991 (Fig. 3-2).
Cebus. Cebus are also classified as frugivore-omnivore, with more
than 50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989). The species has a broad habitat tolerance (Eisenberg 1979) .
They are opportunistic, and usually well able to persist in disturbed
forest (Johns and Skoruppa 1987), which makes them the most adaptable
primate species in the neotropics (Mittermeier and van Roosmalen 1981) .
It is not surprising then that they were able to maintain high
population density in the Reserve. The reduction of 54% in the number of
individuals estimated per km2 at Jusante from 1990 to 1991 (Table 3-5)
seems inconsistent with their ecology. However, because Palmae species
were more abundant in the Reserve than at Jusante (Chapter 2), and
because Cebus apella rely heavily on palms in a number of different ways
(insect foraging, fruits, seeds, flowers, and many other plant parts
(Mittermeier and van Roosmalen 1981, Terborgh 1983)), it is reasonable
to suggest that the Reserve holds a higher carrying capacity for the
species, and hence an increased ability to maintain the increased
densities.
Saguinus. The species is classified as insectivore-omnivore, with
more than 50% of their diet consisting of invertebrates (Eisenberg 1981,
Robinson and Redford 1986, 1989). According to Rylands and Keuroghlian
(1988), optimal habitat for this species includes secondary forest and

65
forest edge mixed with tall primary forest. Several studies have
demonstrated that the species occurs in greater densities in secondary
forest near natural clearings than in mature forest(Eisenberg and
Thorington 1973, Mittermeier and van Roosmalen 1981, Robinson and
Ramirez 1982, Johns and Skoruppa 1987). Emmons (1984) concluded that
Saguinus density appeared to have increased in some areas where large
monkeys had been exterminated. They also overlap with Cebus in most
habitat and diet categories (Mittermeier and van Roosmalen 1981). The
lower densities of Ateles and Cebus at Jusante, and the increased edged
habitat at the Reserve created by the reservoir, were most likely
favorable factors influencing the maintenance of higher Saguinus
densities at both sites during 1990 and 1991.
Callicebus. This species is also classified as frugivore-omnivore,
with more than 50% of its diet composed of fruits, and the remainder
mostly invertebrates and vertebrates (Eisenberg 1981, Robinson and
Redford 1986, 1989). The species occurs in greatest densities in areas
characterized by forest openings with early successional vegetation, and
spends more time in the lower canopy levels and understory vegetation
(Kinsey 1981, Terborgh 1983, Robinson and Redford 1986, Robinson et all.
1987). The more open vegetation at the Jusante site (Chapter 2) most
likely created a more suitable habitat for Callicebus than in the
Reserve, which may explain their much higher densities at Jusante (Fig.
3-2). The possibility of interference competition with Cebus might also
affect Callicebus densities. Both species are catholic in their diet,
however, Cebus generally have larger group size, and are more aggressive
during interspecific encounters, possibly displacing Callicebus groups
from feeding trees. According to Emmons (1984), troops of larger

66
monkeys, such as Cebus, physically prevent access to fruit sources by
small ones, such as Saguinus and Callicebus. The lower densities of
Cebus at Jusante might benefit Callicebus, affecting their population
positively.
Pithecia. Pithecia are also frugivore-omnivores, with more than
50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989) They are usually found in the understory and lower to
middle parts of the canopy (Mittermeier and van Roosmalen 1981), and
they occur in gallery and both primary and secondary forest (Robinson
and Ramirez 1982). Pithecia are always rare (Mittermier and van
Roosmalen 1981, Robinson et all. 1987, Rylands and Keuroghlian 1988),
despite the fact that they have no distinct habitat preference. However,
their rarity may indicate that they are specialists within the forest
they occupy, or at least dependent on certain floristic
communities.(Rylands and Keuroghlian 1988). According to Johns and
Skoruppa (1987) Pithecia are able to feed on fruits from some of the
early colonizing trees, which might explain their higher densities at
Jusante.
Density changes between years at the Reserve
The low degree of similarity between the 1988 and 1989 densities
was most likely due to the increased densities of Pithecia, and
Callicebus, as a result of both the release of captured animals and the
movement of free ranging individuals into the area. The 1990 community
still show a low degree of similarity with 1988, possibly due to the
increase in densities as a consequence of the heavy migration of animals
to the Reserve. By 1991, the degree of similarity with the 1988

67
community, increased to 61.92. The community seems to be in the process
of returning to its original community structure, apparently restoring
its stability (Fig. 3-6).
Density comparisons with other western Amazonian sites
There is great variation in primate density among Amazonian sites
(Table 3-13). I limited the comparison to data from neighboring states
and/or countries in an attempt to avoid comparisons among areas with
distinctly different climate and vegetation. Only unhunted or slightly
hunted sites were used (n= 9, and n = 3, respectively). Mean primate
density for the neotropics, calculated by Robinson and Redford (1986),
is presented for reference.
It is clear that even light hunting strongly affects Ateles
density, and I cannot reject the possibility that Ateles has been hunted
at the Jusante site. There are rubber tappers in the area, and fisherman
sometimes hunt along the river. During the time I worked in the area I
observed a few hunting incidents, however, the main targets were
peccaries, deer, and the agouti. I never saw a captured or dead monkey,
and when questioned, the rubber tappers confirmed that they mainly
killed peccaries and agouties. I believe that the low density of Ateles
at Jusante was related to habitat and not to hunting.
Saimir shows an interesting situation where, in contrast to other
species, the number of groups per km2 is somewhat fixed around 2 despite
variation in the number of individuals per group.
Total density
Estimates of total primate density show a general trend of density
increase from 1989 to 1990 and then a decrease in 1991 (Table 3-6, Fig.
3-5). These changes are consistent with the assumption that the

Table 3-13: Primate densities in western Amazonia. Numbers per km2 (groups per km2). All sites are unhunted
with the exception of Ponta da Castanha, Yavari Miri, and Mamore, which were lightly hunted.
SITE
SPECIES
REFERENCE
Ateles
Cebus
Pithecia
Callicebus
Saimir
Saguinus
Callithrix
paniscus
apella
irrorata
bruneus
ustus
fuscicollis
emiliae
Samuel 1988
13.5
(2.6)
21.7
(4.0)
5.2
(1.4)
0.7 (0.3)
-
17.0
(2.0)
5.4
(0.4)
Eletronorte, 1989
Samuel 1989
13.2
(3.2)
20.1
(5.4)
10.3
(3.4)
7.2 (3.6)
6.9
(0.9)
6.1
(1.9)
4.1
(1.4)
This study
Samuel 1990
23.6
(6.1)
28.5
(6.6)
5.2
(2.1)
1.3 (0.8)
28.7
(3.7)
14.3
(4.1)
10.1
(3.1)
(4
Samuel 1991
11.1
(3.7)
27.1
(6.5)
3.5
(1.2)
1.2 (0.6)
8.9
(1.4)
14.1
(3.7)
3.6
(1.8)
U
Jusante 1990
5.9
(0.6)
17.5
(2.9)
7.5
(2.6)
20.0 (8.3)
23.8
(1.7)
20.4
(5.8)
9.1
(3.1)
6
Jusante 1991
-
8.1
(2.1)
7.6
(2.4)
11.6 (4.9)
6.0
(0.8)
13.6
(3.7)
2.8
(0.9)
44
Acaituba
8.8
(1.5)
7.8
(1.0)
-
-
-
-
-
Johns, 1986
Ponta da Cast.
1.3
(0.1)
11.5
(1.0)
-
-
32.0
(0.5)
-
-
Johns, 1986
Igarape Acu
-
-
-
-
-
8.9
-
Peres, 1990
Tefe
-
11.9
-
-
-
9.7
-
Peres, 1990
Urucu
-
32.2
-
-
10.2
15.0
-
Peres, 1990
Cocha Cashu
25.0
40.0
-
-
60.0
16.0
-
Terborgh, 1983
Cocha Cashu
22.4
(3.2)
36.0
(3.6)
-
-
84.0
(2.1)
10.8
(1.8)
-
Freese etal. 1982
Peru/Iquitos
-
25.0
(2.5)
-
-
72.0
(1.8)
15.0
(2.5)
-
Freese etal. 1982
Yavari Miri
2.5
11.3
37.8
-
Puertas & Bodmer
1993
Mamore
2.0
(0.4)
55.0
(5.5)
-
100.0
(2.5)
Freese etal. 1982
MEAN
11.3
12.4
-
62.3
26.9
Robinson &
Redford 1986

69
animals inside the reservoir moved to both study sites between August
1989 and May 1990, and then dispersed to adjacent forest between
November 1990 and May 1991. Even though 1989 data for Jusante do not
exist, I suspect that the total 1989 primate density for Jusante was,
like the Reserve, similar to its 1991 estimate.
The sharper density decrease at Jusante is most likely related to
lower capacity of the forest to support high primate densities. This is
perhaps most obvious for Cebus, whose density remained at high levels in
the Reserve, but decreased drastically at Jusante (Table 3-5).
Because mean group size for Cebus and Saimir at Jusante was much
higher in 1990 than in 1991, the number of individuals per km2 shows a
more abrupt decline in density than the group density (Fig. 3-5).
Biomass
Metabolic biomass takes into consideration the energy consumption
of the animal in relation to its body weight, therefore it makes a
better assessment of the ecological importance of the animal. The three
most important species accounted for 87, 89, and 63% of the metabolic
biomass in the Reserve during 1989, 1990, and 1991, respectively (Table
3-7). The three dominant species in the 1988 metabolic biomass
calculations were: Ateles = 53.68, Cebus = 40.58, and Pithecia = 9.08
kg0-75 km2, which are also the three largest species, and they accounted
for 91% of the total 1988 metabolic biomass (it is important to remember
that Saimir was not sighted in the Reserve during the 1988 census, and
hence that these calculations were based on six species instead of
seven) (Eletronorte 1989). This pattern repeated itself in the 1989
biomass calculations (Table 3-7).

70
There is a significant relationship between biomass and body
weight for primates, the largest the species the higher its biomass
(Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979; Robinson and
Redford 1986). This principle was observed in the 1989 biomass estimates
(Table 3-7). However, the same was not true for the following years. The
community went from having Ateles, Cebus, and Pithecia as the most
important species in 1988 and 1989 (also the three largest), to having
Ateles, Cebus, and Saimir in 1990, and Cebus, Ateles, and Saimir as
the three most important species in 1991.
Ateles and Cebus were the first and second most important species
during 1988, 1989 and 1990; however, in 1991 Cebus had greater metabolic
biomass than Ateles. This is probably due to the migration of Ateles to
adjacent forest between 1990 and 1991, and to the maintenance of the
increased density of Cebus in the Reserve. Pithecia came in third place
in 1988 and 1989, but it was replaced by Saimir in 1990 and 1991. The
high density increase of Saimir probably resulted from their migration
from the reservoir area to the Reserve.
Ateles and Cebus had similar biomass values in 1988 and 1989,
while Pithecia's biomass in 1989 was much higher than in 1988, most
likely a reflection of the increased density due to the number of
released animals. Despite the increase in Pithecia's biomass from 1988
to 1989, it is clear that the release operation did not change the
primate community significantly. The natural migration of the animals to
the Reserve, on the other hand, changed the community considerably.
The 7% increase in total primate biomass from 1988 to 1989 was
probably a consequence of the release operation; the 55% increase from
1990 to 1991 was most likely due to the natural migration of the

71
animals; and the 40% decrease in 1991, which returned the total biomass
to the same level found in 1988 (Fig. 3-7), was presumably due to
dispersal of the animals to adjacent forests, and to mortality. This
pattern of biomass changes strongly suggests that the primate carrying
capacity for the Reserve is around 150 kg km2.
Despite the lack of data for 1988 and 1989 at Jusante, it is clear
that the primate community at this site was also disrupted by the
migration of animals from the reservoir (Table 3-8).
The similarity of the 1990-1991 pattern between sites suggests
that the Jusante site was affected in a similar fashion to the Reserve,
and if so, the carrying capacity for the area would be around 50 kg km2.
The much lower biomass values for Jusante is expected because the
habitat favors smaller primate species, with Ateles occurring at very
low densities.
The most likely explanation for the great increase in biomass
during 1990 is the migration of animals from the flooded reservoir to
the Reserve and Jusante areas. The total lake area is 502 km2, of which
22 km2 remained green in the form of islands (measured from landsat
images 1:250,000 by Adolfo de La Pria Pereira, SEDAM-RO). Before
flooding, the area was undisturbed primary forest, with little or no
hunting pressure. The rescue operation only removed a small fraction of
the animals, the waters did not cover the tree tops, and there was no
case of high primate mortality inside the reservoir. Therefore, I can
safely state that the animals moved into adjacent areas.
In order to determine how many animals made it out of the
reservoir area alive, I used the biomass values estimated for the
Reserve and Jusante to estimate biomass inside the reservoir prior to

72
the disturbance. My sample areas at Jusante extended up to two
kilometers away from the river bank. However, the northeast corner of
the Reserve is located approximately two kilometers from the river bank
(Fig. 2-1), and the vegetation in the Reserve had a different structure,
and somewhat different composition than that of Jusante (Chapter 2).
Based on these observations I measured the area along two kilometers on
the right and left bank of the Jamari river, and considered it as
"Jusante habitat type" flooded by the reservoir. The remaining area was
then considered to be similar to the Reserve habitat type. The
measurements indicated that 50% of the reservoir consisted of Jusante
habitat, and the other 50% by the Reserve habitat. Then, I used the 1988
biomass estimates to calculate the biomass for 50% of the reservoir
area, and assuming that the 1988 biomass values for Jusante were similar
to the 1991 values (as is the case for the Reserve), I used a biomass
value of 50 kg km'2 to calculate the remaining 50% of the reservoir's
primate biomass. The difference in biomass increase from 1989 and 1990
at the Reserve and Jusante, divided by the 1988 estimated biomass for
the reservoir should then represent the proportion of the total primate
biomass that successfully fled the reservoir area. The Reserve is
officially 210 km2 in size, however, 30 km2 were also flooded by the
reservoir, so I used the value of 180 km2 to calculate biomass for the
Reserve. To be able to make a comparison, the same area was used to
calculate total biomass for Jusante, even though both areas are located
in a continuous forest. The result is presented in Table 3-14.
If all the assumptions above are correct, the results show that at least
65% of the biomass of primates in the reservoir reached "safe" grounds
within the Reserve and Jusante sites, even though they subsequently

73
Table 3-14: Estimation of primate biomass in the reservoir prior to
flooding, and the percentage of which reached the Reserve and Jusante
sites safely, (biomass for G and H calculations based on 50% of the
reservoir's total area).
Site
Year
Biomass
(kg km2)
Area
(km2)
Total Biomass
(kg km2)
a)Reserve
1989
165.27
180
29,748.60
b)Reserve
1990
255.41
180
45,973.80
c)(b minus a)
= Amount
increased
16,225.20
d)Jusante
1989
50.00
180
9,000.00
e)Jusante
1990
136.40
180
24,552.00
f)(e minus d)
= Amount
increased
15,552.00
g)Reservoir
1988
154.00
240
36,960.00
h)Reservoir
1988
50.00
240
12,000.00
i) (g plus h)
= Total
reservoir
biomass
48,960.00
PERCENT OF
RESERVOIR
PRIMATE
BIOMASS
ACCOUNTED FOR
[(c + f) -r i]
* 100 =
48,960.00
= 65%
of reservoir
dispersed to other areas or died (these calculations do not include the
night monkey, Aotus, which also inhabits the area). This high percentage
suggests that the migration of animals was not at random, but directed
to the Reserve and Jusante areas.
If the estimation of the reservoir's biomass is accurate, only
4.5% of the primate biomass was rescued during the operation. I observed
Cebus preying on snails and living inside the reservoir area during
1991, and I am certain that Cebus and other species still remain on some
of the islands. The 22 km2 area of islands could harbor another 7% of

74
the reservoir's original biomass. The total biomass accounted for would
then represent 76.5% of the total original biomass. That would leave
23.5% of the biomass unaccounted for. However, I have no doubt that some
animals died of starvation, and others were preyed upon inside the
reservoir (air born predators can easily kill an animal which cannot run
to the ground for cover, or hide among entangled vegetation), others
have certainly moved to other areas besides the Reserve and Jusante, and
that would account for a significant portion of the remaining biomass.
Terrestrial Diurnal Mammal
Density
The density of both terrestrial mammals censused at the Reserve
were much higher in 1989 than in 1988 (Fig. 3-10), presumably as a
result of the migration of animals to the area as a consequence of the
rising water levels in the reservoir, plus the active release of animals
captured inside the reservoir. Dasyprocta showed a much higher increase
in density from 1988 to 1989 than Mazama. This is not only due to the
fact that they naturally occur at higher densities than Mazama
(therefore more animals migrated into the Reserve), but also to the fact
that 214 individuals of Dasyprocta were released in the Reserve (of a
total of 303 captured), as opposed to only 19 individuals of Mazama
released (out of 23 captured) (Eletronorte 1989). The constant decrease
in Dasyprocta density may be due to predation rather than dispersal. The
Reserve has several species of carnivores, which certainly also
increased in numbers due to the rising of the water level in the
reservoir area. The apparent increase in Mazama density from 1989 to
1990 might be artificial due to restrictions of the TransAn program.

75
Even though the density estimate for 1990 only increased by about one
individual per km2, the 95% confidence interval for 1990 is 100% larger
than that for the 1989 estimate (Table 3-9).
Dasyprocta at Jusante showed a slight increase in density from
1990 to 1991, however, Mazama showed a 74% increase in density for the
same period (Table 3-9, Fig. 3-10). There is no other obvious
explanation for this increase, except perhaps that the species had a
very good reproductive year. The densities of Dasyprocta are much higher
at all times in the Reserve. The greater abundance of mature
Bertholletia excelsia (Brazil nut) and Orbignya barbosiana (babagu) at
the Reserve (Chapter 2) most likely makes this area a better habitat for
Dasyprocta. Mazama species had similar densities in both areas in 1991,
and if the 1990 density at the Reserve was overestimated (due to TransAn
restrictions, as mentioned above) then the 1990 densities for both areas
might also have been similar.
Total density
The density estimate for all terrestrial mammals sampled at the
Reserve was at its highest in 1989. Considering that only 313
individuals (belonging to nine species that were sampled during census)
were released in the Reserve (Eletronorte 1989), the migration of
animals to the Reserve is the most likely explanation for the higher
1989 density. Total group densities were virtually the same from 1989 to
1990, however, from 1990 to 1991 there was a 28% decrease in total group
density. The sharper decrease in the number of ind/km2 is due to a
reduction in mean group size from 1989 to 1990 (Table 3-10, Fig. 3-11).
The drastic density decrease from 1990 to 1991 was possibly a
result of such factors such as:

76
a) mortality: high population densities in 1989 and 1990 may have lead
to higher mortality due to inter- and intraspecific fighting and
interference competition. Assuming that the 1988 densities were much
lower for all terrestrial mammals (as they were for Dasyprocta and
Mazama sp.; Fig. 3-10), total mammal density in 1989 and 1990 were most
likely above the carrying capacity for the area, and a decline in
density would be expected through migration to other areas and/or higher
mortality rates. Small and mid-size mammals have relatively small home
ranges (home range increases with body size (Eisenberg 1979, Clutton-
Brook 1979)), and they are less likely to migrate long distances then
larger ones. Furthermore, migration might be expected to occur shortly
after the increase in density, and result in density changes similar to
that of the primate population (Fig. 3-5). Hence, high mortality rate is
more likely to have occurred than migration.
b)increased predation: even though I did not census the carnivore
community, it is natural to assume that their numbers also increased due
to migration into the Reserve. The increase in the number of animals in
the area, plus a probable increase in litter size, due to increased food
availability in 1989 and 1990, may have contributed to the decrease in
total terrestrial mammal densities as a consequence of higher carnivore
populations, and thus higher predation rates.
Even though group densities were virtually the same in 1990 and
1991 at Jusante, the number of ind/km2 had a 12% reduction during this
time as a result of smaller mean group size in 1991 (Table 3-10, Fig. 3-
11). The reasons for the decrease in density at Jusante are probably
similar to those for the Reserve.

77
Biomass
As expected, the increase in terrestrial mammal biomass in the
Reserve occurred between 1988 and 1989 with the rush of animals fleeing
the reservoir area (Fig. 3-12). Data for both deer species and the
agouti should exemplify the response of the entire community to the
flooding. The possible reasons for such changes have already been
discussed in the terrestrial mammal density section of this discussion.
At Jusante the density for all three species increased from 1990
to 1991, however, this is a reflection of the increase in Mazama
densities, and not an increase for the community as a whole (Fig. 3-12).

CHAPTER 4
UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL DAM
Introduction
The Amazonian region is recognized by its highly diverse bird
communities (Terborgh et al. 1984, Bierregaard 1990), with the greatest
concentration of species in its western portion (Haffer 1990). Bird
species richness is greater in the Amazon Basin where riverine habitats
occupy extensive areas (Remsen and Parker 1983).
The closing of the Samuel Hydroelectric Dam flood gates in the
Jamari River in November 1988 caused the submersion of approximately 107
km of pristine river bank, which represented a loss of approximately 214
km of riverine habitat. This figure considers only that portion of the
river that was completely submerged. However, an even larger stretch of
river was affected by the widening of the river caused by increased
depth as a result of the dam.
Riverine habitats differ in plant and animal composition from
adjacent habitats (Remsen and Parker 1983). Despite the fact that the
filling of the reservoir created a water-edge habitat this edge was a
terra firme forest before flooding, consequently its plant structure and
composition are different from that of the previously existing river-
edge. "The bird species composition of Amazonian river-created habitats
is generally distinct from that of adjacent terra firme forest" (Remsen
and Parker 1983, pg 226), therefore, the composition of bird species at
the reservoir's edge will be somehow different from that of river's
78

79
edge. Will the species displaced from river edge habitats be found at
the lakes's edge? Will the creation of edge habitat at the lake's
perimeter increase bird diversity? A better understanding of the
structure and composition of bird communities in such areas is essential
for the preservation of bird diversity in the tropics.
The purpose of this chapter is to investigate the differences
and/or similarities between understory bird communities in the newly
created habitat and in its previously existing riverine habitat. In
order to make comparisons, the understory bird community was sampled
with the use of mist nets in a pristine riverine habitat (located
downstream from the dam), as well as in an area located at the edge of
the newly formed lake.
Methods
Study Site
Birds were sampled in the same sites as mammals, with the
exception of a new transect line created at the Jusante site (line D;
Fig. 2-1). For a complete description of the study site see Chapter 3.
Study Design
Three 2-km transects at the Reserve (located in areas 1, 2, and
3), and 3, 2-km transects at Jusante (located in areas A, B, and C in
1990, and B, C, and D in 1991), were sampled for understory bird
community structure and composition (transect A was substituted for
transect D in 1991 due to the proximity of its distant end to the river)
(Fig. 2-1). Each transect ran from the edge of the water (lake in the

80
Reserve, and river at Jusante) to the forest interior at a 90 degree
angle to the water course. Transects were between 2 and 4 km apart. Mist
nets were placed along transects within three zones: (1) at the water's
edge, (2) at 1 km, and (3) at 2 km distance from the water. At each
zone, 15, 12-meter-long nets (2.6 m high, with 36 mm black nylon mesh),
were placed along a trail which ran parallel to the main transect. Mist
nets were placed with the lowest shelf at the ground, and in precisely
the same location during replicate surveys. A period of at least three
weeks separated samples of the same net-line.
This research was designed to sample small understory species; no
attempt was made to sample birds by visual or auditory methods. Despite
biases associated with the use of mist nets, such as the fact that
resource availability may affect the probability of a bird's capture
(Greenberg and Gradwohl 1986, Visscher 1981), and the fact that they may
only capture 40% of the species present (Terborgh et al. 1990, Thiollay
1994), nets are a widely used technique in the study of understory bird
communities (Karr 1981), and remain the most productive technique for
this type of study where comparisons are to be made. Furthermore, mist
nets provide an opportunity to collect reliable information in a
relatively short period of time for an area of tropical forest with high
species diversity (Karr and Freemark 1983).
Data Collection
Sites were sampled for 9 days each month (three days in each
transect: one day per zone/per transect) during the months of June
through October of 1990 and 1991, which represents the dry season in the

81
region (Chapter 2). Number of days sampled per zone/per site are listed
on Table 4-1.
Table 4-1: Number of mist-net days per site, year, and zone at the
Reserve and Jusante during 1990 and 1991.
ZONE
RESERVE
1990
1991
JUSANTE
1990
1991
EDGE
09
09
07
09
1 KM
09
09
07
09
2 KM
09
09
07
09
TOTAL
27
27
21
27
GRANDTOTAL
54
48
Nets were operated for 6 hours, from dawn to 1200/1300 hours. All
identified birds were sexed and aged (whenever possible), banded with
government-issued aluminum bands (except for hummingbirds), and released
at the site of capture. Measurements on body mass (g), lengths of wing,
tail, tarsus, bill (mm), and total body length (cm) were taken on all
birds, along with records of date, time, and place of capture and/or
recapture. Some specimens were collected for identification at the
Goeldi Museum in Belm, others were photographed alive for later
confirmation of field identification. Eight individual birds were not
identified at species level during the course of this study. They were
eliminated from the data set.
Data Analysis
The data set was initially described by species accumulation
curves, histograms, and tables of individuals and species captured per

82
site, year, and zone. Chi-square tests were used to compare number of
captures and species between years within zones and sites. This was done
in the interest of combining data from both years within each site.
Similarity among the species detected at the Reserve and Jusante,
and among species detected in each zone (within and across areas) was
quantified using Jaccard's Coefficient of Similarity, which equals zero
if no species are in common and one if the lists are identical. Cluster
analyses (group average strategy) were performed using the results of
the similarity index in order to identify any resemblance between sites
(Ludwig and Reynolds 1988).
To examine differences and/or similarities between understory bird
community structure at the Reserve and Jusante sites, Correspondance
Analysis (COA) were performed using the program CANOCO Version 3.12 (Ter
Braak 1987).
For the COA analysis, species were first grouped into 6 foraging
strata, 7 diet, and 8 substrata categories according to Karr et al.
(1990). The initial arrangement resulted in 40 different categories,
which were then condensed to 10 guild categories; insectivorous guilds
were based on primary foraging substrate. The guild categories included
six insectivore guilds (flycatchers=sallying/hawking/snatching [IA],
dead-leaf gleaners [ID], live-foliage gleaners [IL], terrestrial [IG],
bark-gleaners [IB], and army ant-followers [AF]), one nectarivore guild
(NI), one frugivore guild (FR), one omnivore guild (OM), and one guild
of birds which were directly associated with water (WA). The latter
guild includes piscivorous species, as well as insectivorous species
which catch insects over the water. The species Xenops minutus was the
only species that uses twigs as a substrate, so it was grouped with

83
those that use bark. The variables included in the COA consisted of the
number of species detected in each guild, and the sampling units
consisted of 18 survey areas. Information on the location of sampling
sites along transects is not used in the COA; therefore, if samples
become ordinated in a way that parallels this gradient, the COA will
lend credence to the ecological significance of this gradient
(Canaday 1991).
Results
Inter-Year Comparison
There were no significant differences between number of captures
or number of species between 1990 and 1991 samples in either area (Chi-
Square, p= 0.756, n= 546, p= 0.773, n= 603, for number of captures at
the Reserve, and Jusante, respectively, and p= 0.980, n= 207, p= 0.918,
n= 215, for number of species at the Reserve and Jusante, respectively)
(Tables 4-2, and 4-3) Thus, data from 1990 and 1991 were pooled.
Table 4-2: Number of captures per site, zone, and year. Chi-square
test performed for each area on number of captures per zone between
years. () percent of total capture.
ZONE
RESERVE
1990
1991
JUSANTE
1990
1991
EDGE
127 (50)
139
(47)
78
(28)
93
(29)
1 KM
68 (27)
86
(29)
103
(36)
107
(34)
2 KM
57 (23)
69
(24)
103
(36)
119
(37)
TOTAL
252 (100)
294
(100)
284
(100)
319
(100)
CHI-SQUARE
P=0.756
P
=0.773

84
Table 4-3: Number of species per site, zone, and year. Chi-square
test performed for each area on number of species per zone between
years. () percent of total species.
ZONE
RESERVE
1990
1991
JUS ANTE
1990
1991
EDGE
43
(69)
42
(71)
37 (56)
42
(55)
1 KM
30
(48)
31
(53)
30 (45)
39
(51)
2 KM
31
(50)
30
(51)
30 (45)
37
(49))
TOTAL
62
<84)
59
(80)
66 (68)
76
(78)
CHI-SQUARE
P=0.980
P=0.918
Overall Site Comparison
I found no significant difference in the number of species per
zone (Chi-Square, p = 0.979, n = 293). However, the sites were highly
differentiated when tested against the number of captures per zone (Chi-
Square, p = 0.000, n = 1,149 (Table 4-4, Fig. 4-1).
Table 4-4: Number of species and number of captures per site, and zone
(years combined). Chi-square test performed on number of species and
capture per zone between sites. () percent of species present and
total capture.
ZONE
SPECIES
RESERVE
JUSANTE
CAPTURES
RESERVE
JUSANTE
EDGE
56 (76)
61
(63)
266
(49)
171
(28)
1 KM
43 (58)
47
(48)
154
(28)
210
(35)
2 KM
40 (54)
46
(47)
126
(23)
222
(37)
TOTAL
74 (100)
97
(100)
546
(100)
603
(100)
CHI-SQUARE
P=0.979
P=
0.000

85
I I Reserve
1///I Jusante
Figure 4-1: Number of species and number of captures at each zone
at the Reserve and Jusante sites during 1990 and 1991.

86
Despite the slightly higher sampling effort at the Reserve (n = 54
days), than at Jusante (n = 48 days), both the number of total captures
and number of species were higher at Jusante than at the Reserve (n =
603 captures and 97 species, and n = 546 captures and n = 74 species,
respectively) (Fig. 4-2). The species accumulation curve at the Reserve
seems to be reaching an asymptote, while at Jusante the curve continues
to extend upwards (Fig. 4-3 and 4-4); however, the cumulative number of
captures and species do not differ significantly among sites
(Kolmogorov-Smirnov, p = 0.6325 and 0.1462, n = 102, for number of
species and number of captures, respectively). Within sites, the species
accumulation curves are almost identical for the 1 and 2 km zones, while
at the edge the number of species captured is higher than in the
interior of the forest (Fig. 4-3 and 4-4) The Jusante site differs from
the reserve in that there were fewer individuals captured at the edge
than at the 1 and 2 km zones (Fig. 4-1, Table 4-4).
Species Richness
During 1990 and 1991, a total of 1,149 individuals of 118 species
(Appendix B) were captured during 9,180 net-hours in both the Reserve
and Jusante sites (see Appendix C see for list of species with
occurrence confirmed). The two most speciose families for both areas
combined were the Formicaridae and Tyrannidae, which were represented by
25 and 20 species, respectively. At Jusante those two families were also
the most speciose, represented by 24 and 16 species, respectively. At
the Reserve the Formicaridae had the highest number of species (14),
while the Tyrannidae and the Furnaridae came in second represented by 10

Cumulative N of species
87
Figure 4-2: Cumulative number of captures and species at the
Reserve and Jusante sites for 1990 and 1991.

Cumulative N of species
88
80
70
60
50
40
30
20
10
Total O Edge 1 km 2 km
Cumulative N of captures
4 3: Cumulative number of captures and species per zone
for the Reserve during 1990 and 1991.

Cumulative N of species
89
Figure 4-4: Cumulative number of captures and species per zone
for Jusante during 1990 and 1991.

90
species each. Tropical forest communities have long been recognized for
their high diversity as well as their peculiar structure involving many
rare species (Richards 1952). This study shows that the 20 most common
species represent 70% of all captures at the Reserve, 67.8% at Jusante,
and 63% for both areas combined. No one species dominates the rank
abundance distribution, and a long tail of rare species is detectable in
both areas separately, as well as in the combined samples (Fig. 4-5).
Rarity is defined by Karr (1971) as those species making up less than 2%
of individuals captured. At the Reserve 57 species (77% of the sample)
occurred at rates lower than 2%; at Jusante, this proportion was higher
(83 species or 86% of the sample; Kolmogorov-Smirnov p = 0.0003, n =
171). The combined data set shows that of the 118 species captured, 105
occurred at rates lower than 2% (89% of the sample), and 27 (23%) were
represented by only one individual.
Guild Structure
Of the total 118 species, and 1,149 individuals captured, 77
species (65%) and 789 captures (69%) were insectivores. Guild
distribution for the Reserve and Jusante, and for both areas combined
are shown in Figure 4-6.
Insectivores represented 68 and 69% of all captures at the Jusante
and Reserve sites, respectively. The 2 guilds that had the greatest
number of species at Jusante were the live-foliage gleaner insectivores
(n = 30), and the omnivores (n = 11). At the Reserve the 2 guilds that
had the greatest number of species were the live-foliage gleaners (n =
20), and the bark-gleaner insectivores (n = 9). When number of captures
were considered, the 2 most common guilds at Jusante were the live-

Percent of captures
91
Rank-ordered abundances
Number of species
Number of species
Figure 4-5: Rank-ordered abundance for the Reserve, Jusante
and both areas combined for 1990 and 1991.

92
N of species
N of captures
Guilds
Figure 4-6: Guild structure of the community at the Samuel Ecological
Station, at Jusante, and both areas combined for 1990 and 1991.
Guild denominations are: IL = live-foliage gleaners, IB = bark-
gleaners, IA = flycatchers, IG = terrestrial insectivores, ID = dead-
leaf gleaners, AF = army ant-followers, OM = omnivores, FR =
frugivores, WA = birds which were directly associated with water, and
NI = nectarivores.

93
foliage gleaner insectivores with 24% of all captures, and the
frugivores with 20% of all captures. At the Reserve the 2 guilds with
the greatest numbers of captures were the live-foliage gleaner, and the
bark-gleaner insectivores with 22 and 21% of all captures, respectively.
Despite differences, community structure did not differ significantly
between sites (Kolmogorov-Smirnov, p = 0.7591 and 0.9883, n = 20, for
percentage of captures per guild, and percentage of species per guild,
respectively).
Faunal Similarity Between and Within Sites
The 12 most abundant species in each area are shown in Table 4-5.
The lists show great differences in abundance of species between sites.
Jaccard's Similarity Index for the Reserve and Jusante as a whole was
0.449, which indicates that the sites shared approximately half of the
species detected. Of the 118 species detected, the Reserve had 21
species that were unique, while Jusante had 44 unique species. The
number of species in common between sites was 52. The index was also
used to determine similarity of the species lists between zones (across
and within sites). The similarity indexes for all zones are shown in
Table 4-6.
Note that the edge zone at Jusante had very little in common with
any other zone (the least similar zones were the edge and the 2 km sites
at the Jusante). The edge at the Reserve had an index of approximately
0.35 across zones, while the 1 and 2 km zones for both sites had indices
ranging from 0.45 to 0.63 (with the highest similarity index between the
1 and 2 km zones at Jusante). A cluster analysis was performed using the

94
Table 4-5: The 12 most abundant species captured at the Reserve and
Jusante sites (years combined), and the percentage of the sample that
each one represents.
RESERVE
SPECIES
%
JUSANTE
SPECIES
%
Pipra nattereri
7.1
Pipra nattereri
10.1
Xiphorhynchus elegans
7.1
Hylophylax poecilinota
10.1
Glyphorhynchus spirurus
6.8
Glyphorhynchus spirurus
5.5
Pipra rubrocapilla
6.2
Habia rubica
4.2
Dendrocincla merula
3.7
Pipra fasciicauda
3.5
Onychorhynchus coronatus
3.7
Schiffornis turdinus
3.5
Deconychura stictolaema
3.3
Xiphorhynchus elegans
3.3
Hylophylax poecilinota
3.3
Dendrocincla merula
3.2
Phlegopsis nigromaculata
3.3
Myrmotherula longipennis
3.2
Myrmotherula hauxwelli
2.9
Myrmotherula hauxwelli
2.3
Schiffornis turdinus
2.9
Rhegmatorhina hoffmannsi
2.3
Synallaxis rutilans
2.7
Xenops minutus
2.3
% OF TOTAL CAPTURE
53.1
53.4

95
Table 4-6: Matrix showing results of Jaccard's Similarity Index
between zones. J = Jusante, R = Reserve. Numbers in bold identify
the least and the most similar zones.
J-EDGE
J-l KM
J-2 KM
R-EDGE
R-l KM
R-2 KM
J-EDGE
1.000
J-l KM
0.200
1.000
J-2 KM
0.189
0.603
1.000
R-EDGE
0.272
0.338
0.378
1.000
R-l KM
0.209
0.500
0.459
0.456
1.000
R-2 KM
0.247
0.450
0.458
0.371
0.566
1.000
matrix shown on Table 4-6, and the results clearly showed the grouping
of the 1 and 2 km zones at Jusante, and of the 1 and 2 km zones at the
Reserve. The edge zone at the Reserve is an intermediate area between
the four inland zones and the edge zone at Jusante, which is completely
different from all other sites (Fig. 4-7).
To determine similarity of species lists between individual mist
net lines Jaccard's Similarity Index was calculated for all transect
lines (with the exception of area A at Jusante which was sampled for 3
days only, in 1990). A cluster analysis was conducted using Jaccard
similarity values in order to explore fine patterns of affinity between
transect lines. The result, shown in Figure 4-8, grouped zones according
to sites and distances from the water, with only one exception: area C-E
at the Jusantes's edge was grouped with Jusantes's forest interior
zones.

96
Similarity
4.24 -
36.16 -
68.08 -
100.00
Jusante
Edge
Jusante
1 km
Jusante
2 km
Reserve Reserve Reserve
2 km 1 km Edge
Zones
Figure 4-7: Cluster analysis of Jaccard's Similarity indexes between
zones at the Reserve and Jusante.

97
Ecological Similarities Among Sites
Correspondence (COA) analysis is often used because it allows for
the examination of ecological interrelationships between sampling units
and species in a single analysis. However, COA "is good primarily for
one axis, the second axis tends to be an arch, and axes after the first
are often hard to interpret" (Ludwig and Reynolds 1988, pg. 254). The
results of a COA on my data show this to be true. The first axis of the
correspondance analysis explained 34.3% of the variation, and it
separated the sampling units according to zones (Fig. 4-9). The second
axis explained 19.6% of the variation, however, its interpretation is
not clear, and will not be discussed.
The bird guilds that showed strong positive association with the
first axis were IA, and WA, while the guilds that showed strong negative
association were AF, and IG (Fig. 4-10) Guilds of flycatchers and of
birds associated with water were, thus, more abundant at edge habitats,
while ant-followers and terrestrial insectivores were more abundant at
the forest interior.
Discussion and Conclusions
Overall Site Comparison
Species richness
Edges are often considered more diverse habitats than forest
interior due to a greater complexity of vegetation structure. I expected
a greater number of species to occur at the edge zones. However, the
number of species captured at the edge, and in the 1, and 2 km zones did
not differ statistically in either site (Table 4-3). Terborgh et al.

98
Similarity
Jusante Reserve Reserve Jusante Jusante
E0ge(*) Edge Inland Inland Inland
Net-lines
Figure 4-8: Cluster analysis of Jaccard's Similarity Index for 18
transect lines located at different zones at the Reserve and Jusante
sites.

-3
0
1
e
o
Jusante edge
A
Jusante 1 km
/

Jusante 2 km
1
0
Reserve edge
\
A
A
Reserve 1 km
0
A* A

Reserve 2 km

AXIS I
I
3


Edge
j
-3 -
Axis II
Forest interior
Figure 4-9: Correspondence analysis ordination of 18 transect lines at the Reserve and
Jusante sites. Dashed lines encircle interior forest transect that was grouped with
water edge habitat.

100
Figure 4-10: Guild association with correspondance analysis. Arrows show
guilds most strongly associated with first axis. IL = live-foliage
gleaners, IB = bark-gleaners, IA = flycatchers, IG = terrestrial
insectivores, ID = dead-leaf gleaners, AF = army ant-followers, OM =
omnivores, FR = frugivores, WA = birds which were directly associated
with water, and NI = nectarivores.

101
(1990) also established that the number of bird species did not increase
at edge habitats in Amazonia, regardless of the type of edge. Therefore,
the creation of the reservoir did not contribute to an increase in point
diversity, as might have been expected.
Species abundance
The number of captures at the Reserve's edge was significantly
greater than at the 1 and 2 km zones, and also greater than at the
Jusante's edge (Table 4-4). The higher density at the Reserve's edge was
most likely due to the settlement of displaced individuals from the
reservoir's area into the newly created habitat. The difference in
capture rates at the 1 and 2 km zones at the Reserve and Jusante was
probably due to the difference in understory vegetation structure. The
Jusante site had a more open canopy than the Reserve, and consequently,
a greater penetration of light and higher density of shrubs and
understory trees (Chapter 2). The Jusante site had, therefore, a more
complex and probably richer understory vegetation, which explains the
higher densities at the forest interior zones of Jusante.
Guild Structure
Even though live-foliage gleaning insectivores was the guild most
speciose, and with the greatest number of captures at both sites, ant-
followers and frugivores had the highest number of captures, compared to
the number of species present in each guild (Figure 4-6). Brawn et al.
(1995) found that frugivores have a significantly greater population
size than insectivores, and that, among insectivores, ant-followers have
the largest population size. The large number of frugivores and ant-
followers captured in this study is not a sampling bias, because

102
frugivores have the lowest capture probability, and flycatchers the
highest (flycatchers in this study had low capture rates) (Brawn et al.
1995). Mist net data, therefore, most likely reflected the actual guild
structure of the forests.
The 3 guilds that most differed structurally among sites were the
bark-gleaner, and ant-follower insectivores, and the guild of birds
associated with water. The latter had a greater number of species at the
Jusante site than at the Reserve, however, the percentage of captures
were similar. According to Remsen and Parker (1983), species restricted
to river-created habitats comprise 15% of the total land bird fauna of
the Amazon Basin, therefore, the higher number of species at Jusante
might be a reflection of the greater diversity at the river edge. The
bark-gleaner insectivores had greater number of captures at the Reserve,
while the ant-followers were more abundant at Jusante, which may be due
to differences in forest structure. As discussed earlier, the Jusante
site has probably a richer understory, resulting in a greater abundance
of ants, which, in turn, would account for the higher abundance of ant
following birds. The Reserve, on the other hand, has larger DBH trees
(Chapter 2) which means older trees, and probably a greater number of
dead trees, which could account for the greater densities of bark-
gleaner insectivores.
Faunal Similarities
Members of the Pipridae family are often the most frequently
captured individuals in tropical areas (Brawn et al. 1995, Terborgh et
al. 1990, Remsen and Parker 1983, Robinson et al. 1989) and, despite
differences in species abundance, Pipra nattereri was the most

103
frequently captured species in both areas. The 12 most abundant species
in both areas accounted for 53% of all captures, despite differences in
abundance among species. A few moderately dominant species and a
majority of rare species is a typical pattern of bird community
structure of most tropical forests (Thiollay 1994).
The forest interior in both sites was more similar in species
composition than were the edge zones (Fig. 4-7) The edge of the lake at
the Reserve had the vegetation structure of a terra firme forest, even
though it was at the water's edge. This may be why the area was similar
to an intermediate zone between forest interior and river edge.
Cluster analysis of individual transect lines, grouped 1 and 2 km
transects indistinctly (Fig. 4-8). The distance from the water's edge
appears not to influence species composition beyond a certain distance.
However, the Reserve and Jusante sites were separated by forest
interior, and edge zones. Transect C-E at Jusante was grouped with the
forest interior transects despite its location at the river edge. The
transect is located at a point on the river were the bank is very steep,
therefore, its vegetation is characteristic of a terra firme forest
instead of a river edge, or varzea vegetation.
Ecological Similarities
The arrangement of Jusante and Reserve water-edge transects on one
side of the first COA axis, and of the forest interior transects on the
opposite side, indicates that despite the difference in species
composition, both sites have similar community structure (Fig. 4-9).
This arrangement was due to a greater number of species in the WA, IA,

104
and NI guilds at the river and lake edges, and of a greater number of
species of IG, and AF guilds in the interior of the forest (Fig. 4-10).
River edge forests are unique in their species composition; only
7-28% of species found in river edges are found in terra firme forests
(Remsen and Parker 1983). The creation of the lake edge opened niches
associated with water at the Reserve, and structurally the edge at the
Reserve is similar to the edge at the Jusante. However, the species
composition of both areas is different due to the uniqueness of river
edge species. Forest interior, on the other hand, seems to be similar in
both species composition and structure at the Reserve and Jusante sites.

CHAPTER 5
RECOMMENDATIONS FOR FUTURE HYDROELECTRIC DAM CONSERVATION PROGRAMS
Overview
One of the first well documented animal rescue operation in
tropical habitats was associated with the formation of Lake Kariba in
Zimbabwe (Child 1968). The rescue of game animals was carried out on
islands that formed slowly during a 5-year period following the damming
of the Zambezi River. Kariba's operation was done primarily in
relatively open tree savannas, and animals were only handled when they
could not be driven to safety. A total of 3,800 large mammals were
rescued from 1958 to 1961. Eighty-five percent of the animals rescued
were driven to or released on the mainland (Fig. 5-1).
The first major rescue operation in South America was the
Brokopondo Project in Surinam in 1964 (Walsh and Gannon 1967). The
rescue operation lasted 18 months, and involved approximately 10,000
animals. Surinam's dense vegetation was a confounding factor, and there
was concern about overpopulating the lake's shore where animals were
released; however, all animals captured were released.
Rescue operations in Brazilian Amazonia Hydroelectric Dam sites
started in 1984 with Tucurui Dam Operation (state of Par). A total of
284,211 animals were captured and released at the lake's shore without
any consideration to potential overpopulation problems, or to impacts of
released animals on the behavior of resident animals (Eletronorte 1985).
The second rescue operation in Brazilian Amazonia was Balbina (state of
105

106
Amazonas) in 1987, where 19,536 animals were rescued, of which 12,687
were released at the river's edge, downstream from the dam (Eletronorte
1988b). The Samuel rescue operation was the first operation in which
only a small fraction (18%) of the animals rescued (16,000) were
Figure 5-1: Charts show percentage of animals captured which were:
ljreleased, 2)sent to research institutions, and 3)died during rescue
operations at each hydroelectric dam. Data on animal death during
rescue operation at Brokopondo were not available.

107
released (Fig. 5-1) (Eletronorte 1989). It also was the first time that
a land conservation unit, or preserve, was created to compensate for the
loss of habitat caused by flooding. Despite the improvements made in the
Samuel rescue operation in contrast to previous efforts, many aspects of
rescue operations need to be reconsidered.
Considerations
Rescue Operations
The cost-benefit ratio of rescue operations need to be reevaluated
before future rescue operations are initiated. The rationale behind
rescue operations has always been to save animals from unquestionable
death. However, animal mortality at hydroelectric dam sites is related
more to loss of habitat than to drowning. Several large mammals have
relatively good swimming capabilities (Child 1968). Kennerly (1963) has
even indicated the importance of swimming in ensuring gene flow between
populations separated by rivers. On the other hand, habitat loss is
irreversible. Kariba reservoir, the largest of all dams discussed above,
flooded an area of 5,462 km2 (ironically enough, its rescue operation
captured the least total number of animals, but most were ungulates
(Fig. 5-1)), Brokopondo flooded 1,683 km2, Tucurui 2,430 km2, Balbina
2,600 km2, and Samuel, the smallest of all, flooded 560 km2. Mammals,
amphibians, and some reptiles may survive by swimming to dry land, and
birds may fly away; however, the habitat lost can never be recovered.
Furthermore, animals captured during rescue operations are usually
released on the nearest piece of dry land, without any concern for the
animal community inhabiting the area of release. My results on mammal

108
density changes at the Samuel release site demonstrated not only that
they can move from the flooded area on their own, but also that the
site's carrying capacity will determine if they will stay on the release
area or not. Of the 3,729 mammals rescued from the Samuel reservoir,
only 2,374 were released (1,352 of these were primates). According to
biomass calculations (Table 3-7, Fig 3-7), the release of primates
caused only a slight increase in total primate biomass (7%) from 1988
(prior to the release) to 1989 (immediately after the release). From
1989 to 1990, total primate biomass increased 55%; this increase only
can be explained by the movement of primates from the flooded reservoir
to the release area. The movement of animals after the flooding
confirmed that primates can move on their own (provided that the filling
of the reservoir does not cover the tree tops). Biomass calculations for
1991 showed, however, that primates in general did not establish
residence at the release site. There was a 40% decrease of total primate
biomass from 1990 to 1991, with biomass returning to almost the same
level found in 1988 (154 kg km"2 in 1988, and 153 kg km"2 in 1991) .
Although the density of some primate species increased at the release
site, the increase included species that have great plasticity, such as
Cebus. The density of less robust species, such as Pithecia, decreased
from 1988 to 1991; therefore, total biomass for the area was maintained
(Fig. 3-3, and 3-5). Projections of total primate biomass in the
reservoir area before its filling, and of biomass increase in the study
sites (Table 3-14), suggest that at least 65% of all primates survived
the initial flooding. This study also demonstrated that many animals
moved through the release site, but it is unknown how far they had to go
to establish themselves, or even if they survived. The loss of habitat,

109
then, becomes the main issue in the survivorship of mammal populations
on hydroelectric dam sites.
The evidence for terrestrial mammals is less conclusive than
primates due to the fact that sample sizes for most species were small.
However, data on agouti and deer species suggest a similar pattern of
movement. Only 19 deer and 214 agoutis were released from November 1988
to March 1989; however, there was a 688.5% increase in deer and agouti
biomass at the release site from 1988 to 1989 (Fig. 3-12). The increase
can only be explained by the movement of animals fleeing from the rising
waters of the reservoir into the release site. Agouti and deer have,
therefore, swimming capabilities, and were able to escape the habitat
being flooded. Their biomass remained high for at least one year (from
1989 to 1990), but suffered a decrease from 1990 to 1991. The population
level at the release site may have been below carrying capacity for
those species, or their response to high density levels may be slower
than that of primates. But, again, the loss of habitat was of greater
importance than site mortality.
Data on bird community composition and structure demonstrated even
more clearly that the loss of habitat is the major concern associated
with the construction of dams. River-edge communities have a very unique
faunal species composition, with some species endemic to these sites.
The water edge created by the reservoir did not "substitute" for the
river edge, and as a consequence many species that were present at the
river's edge below the dam, were not present at the reservoir's edge
(Chapter 4). It is unknown if these species survived elsewhere, but the
fact remains that the habitat loss can never be replaced.

110
Rescue operations have become a public relations strategy used by
power companies to appease public opinion. The only way to make power
companies change their policies is to inform the public, and to give
power companies better options for rescue/conservation programs. Some
suggestions for future rescue programs are listed below:
1. Rescue operations should be confined to:
(a) endangered species. Several species classified as "endangered" by
IBAMA (Instituto Brasileiro do Meio Ambiente e Recursos Naturais
Renovveis), and "vulnerable" by IUCN (International Union for the
Conservation of Nature) were present at Samuel. These included giant
anteaters (Mirmecophaga tridactyla) giant armadillos (Priodontes
maximus), spider monkeys (Ateles paniscus), and many others. Species
known to be endangered and/or vulnerable should be rescued when
stranded.
(b) species that are unable to escape their flooded environment.
Sometimes females with infants are trapped on small islands unable to
swim with their young. An effort should be made to rescue those animals.
(c) species that could be used for research. At Samuel, poisonous
snakes, scorpions, and spiders were sent to several centers for
development and/or production of vaccines.
(d) species that could be used for re-establishing depleted populations
elsewhere. Some areas in the Amazon have been heavily hunted, and large
species are sometimes rare. Rescued animals such as deer, peccaries, and
large monkeys could be released at nearby sites where populations of
those species are low.

Ill
2. Only release animals on sites that have been studied
previously, and that have abnormally low densities, such as heavily
hunted sites.
3. Continue with the practice, established at Balbina, of donating
rescued animals to research institutions. Several health research
institutions utilize primates in their experiments. If rescued animals
were used to supply those institutions, there would be a decrease in the
number of wild animals removed from other sites for research purposes.
4. Create conservation units to compensate for the habitat loss
due to the creation of reservoirs.
5. Invest in professional staff to design, conduct, and supervise
conservation program at dam sites.
Creation of Conservation Units
Animal rescue operations are extremely expensive and for the most
part ineffective. If only a small portion of animals are rescued, the
funds usually reserved for such operations would be better used if
invested in the acquisition and maintenance of conservation units.
Creation of a conservation unit near the impacted area is of
extreme importance not only as a form of compensation for lost habitat,
but also as a place for displaced and/or released animals to move
through. As I have demonstrated, if given the opportunity, animals will
move from the flooded area into adjacent habitats. This practice may not
necessarily save their lives, but will give some individuals a better
chance of survival. A forest connection, however, must exist between the
impacted area and the conservation unit to enable the animals to move
from one area to another. Conservation units should be at least equal in

112
size as the forested area lost by the creation of the reservoir, and
should contain habitats similar to those lost.
The maintenance (including management, research, and protection)
of the conservation unit has to be the responsibility of the power
company in charge of the hydroelectric dam construction. Their profit
does not end with the construction of the power plant; therefore, their
responsibility to conservation should not end either, but rather
continue, at least, throughout the life span of the hydroelectric dam.
Preliminary Studies
Preliminary studies of plant and animal communities at future
reservoir sites, as well as at the conservation unit, are key
instruments for conservation strategies, and should be carried out well
in advance of dam constructions. Most researchers are called to dam
sites only a few months before the flooding occurs. It is imperative
that environmental studies begin at the same time as do feasibility
studies for the construction of a particular dam. Eletronorte started
the feasibility study at the Samuel Dam site some 10 years before
construction (Francisco Silveira Pereira, pers. communication); however,
ecological studies at the site started only 1 year before the filling of
the reservoir (Eletronorte 1989). Data on mammal density, and bird
community composition from the reservoir area, prior to its filling,
would have greatly helped in the understanding of the impact of the dam
in those communities.

113
Follow-up Studies
Long-term studies on community changes and or adaptations to new
environmental conditions are of the utmost importance in the
understanding of impact studies. My results demonstrated that several
years of sampling are necessary to detect, and understand changes in an
animal community. Long-term commitment from hydroelectric power
companies, as well as from the scientific communities, is essential to
an understanding of environmental impacts.
Eletronorte invested millions of dollars creating the infra
structure necessary for a series of follow-up studies at Balbina and at
Samuel, but the proposed research never happened, mainly due to lack of
sustained funding. As an example, large tanks were built at Balbina for
the purpose of studying manatees as potential macrophyte controllers in
the reservoir. Nutrient surplus in the reservoir's water, due to the
vegetation present at the bottom, usually causes an explosion of such
aquatic vegetation. The proposed study was never carried out, and until
recently the tanks were empty. Recently, they were converted into
aquaculture tanks. The same is true for the construction of laboratories
for water analysis at both Balbina and Samuel; the laboratories exist
but no research is promoted by Eletronorte. Hydroelectric power
companies must be responsible for promoting and sustaining research
before, during, and after the construction of dams. The aggregate
studies must focus at finding solutions for environmental problems
created by hydroelectric dam constructions.

114
Conclusions
It is not possible, or even desirable, to maintain the Amazon
region free from development and human impact. What must be done is to
integrate people and nature in a compatible co-existence. In order to do
so, we must study all forms of human disturbance in the Amazon and at
the same time try to minimize their consequences. Only then, may we be
able to preserve both nature and human integrity.
Eletronorte greatly improved its rescue operation policy between
its first rescue at Tucurui and the current effort at Samuel. The
practice of sending the bulk of animals to research institutions
(instead of releasing all of them), and the creation of a conservation
unit were commendable changes on the part of Eletronorte. Greater
changes, however, must be implemented in order to minimize impact and
maximize conservation of the Amazonian region; these changes will
necessitate funding both preliminary and follow-up ecological studies.

APPENDIX A
MAMMALIAN SPECIES RECORDED AT THE SAMUEL DAM SITE.*
ORDER
FAMILY SPECIES
Marsupialia
Primates
Edentata
Didelphidae
Callithrichidae
Cebidae
Choloepidae
Dasypodidae
Caluromys lanatus
Caluromys sp
Didelphis marsupialis
Marmosa sp
Metachirus nudicaudatus
Monodelphis sp
Philander opossum
Callithrix emiliae
Saguinus fuscicollis
Aotus azarae
Ateles paniscus
Callicebus bruneus
Cebus apella
Pithecia irrorata
Saimir ustus
Choloepus juruanus
Cabassous unicinctus
Dasypus kappleri
Dasypus novemcinctus
Priodontes maximus
115

116
ORDER
FAMILY SPECIES
Rodentia
Cetacea
Carnivora
Myrmecophagidae
Sciuridae
Hydrochaeridae
Dasyproctidae
Echimydae
Muridae
Erethizontidae
Delphinidae
Platanistidae
Felidae
Cyclopes didactylus
Tamanda tetradactyla
Myrmecophaga tridactyla
Sciurus langsdorffi
Sciurus sp
Hydrochaeris hydrochaeris
Agouti paca
Dasyprocta fuliginosa
Echymys macrurus
Echymys grandis
Isothrix bistriata
Lonchothrix emiliae
Proechimys longicaudatus
Proechimys sp
Mesomys hispidus
Rhipidomys leucodactylus
Nectomys squamipes
Oryzomys concolor
Oryzomys megacephalus
Coendou prehensilis
Sphiggurus insidiosus
Sotalia fluviatilis
Inia geoffrensis
Felis concolor
Felis pardalis

117
ORDER
FAMILY SPECIES
Mustelidae
Felis yagouaroundi
Panthera onca
Eira barbara
Procyonidae
Galictis cuja
Nasua nasua
Perissodactyla
Tapiridae
Potos flavus
Tapirus terrestris
Artiodactyla
Cervidae
Mazama americana
Tayassuidae
Mazama gouazoubira
Tayassu pcari
Tayassu tajacu
* Exclusive of Chiroptera
Source: Eletronorte 1993

APPENDIX B
BIRD SPECIES CAPTURED DURING THIS STUDY
ORDER/FAMILY SPECIES
J(N) R(N)
Gruiformes
Eurypygidae
Columbiformes
Columbidae
Caprimulgiformes
Caprimulgidae
Apodiformes
Trochilidae
Eurypyga helias
1 0
Geotrygon montana 11 6
Leptotila verreauxi 1 0
Nyctiphrynus ocellatus 0 1
Nytidromus albicollis 0 1
Campylopterus largipennis 2
Florisuga mellivora 0
Glaucis hirsuta 1
Phaethornis philippi 11
Phaethornis ruber 0
Phaethornis superciliosus 9
Thalurania furcata 0
4
2
0
11
1
0
5
Coraciiformes
Alcedinidae
Chloroceryle aenea 0 7
Chloroceryle americana 2 0
GUILD
W-SI-W
G-FR-G
G-FR-G
A-SI-A
A-SI-A
S-NI-F
C-NI-F
S-NI-F
S-NI-F
S-NI-F
S-NI-F
S-NI-F
W-FI-W
W-FI-W
118

119
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Chloroceryle inda
1
3
W-FI-W
Momotidae
Momotus momota
3
0
S-LO-F
Piciformes
Galbulidae
Galbula albirostris
0
4
U-LI-A
Galbula ruficauda
4
0
C-LI-A
Bucconidae
Malacoptila rufa
2
2
S-LI-F
Monasa nigrifrons
7
0
C-LI-A
Nonnula ruficapilla
4
1
S-LI-F
Capitonidae
Capito dayi
0
2
C-LO-F
Ramphastidae
Pteroglossus bitorquatus
1
1
C-LO-F
Picidae
Veniliornis affinis
1
1
U-SI-B
Passeriformes
Dendrocolaptidae
Deconychura longicauda
0
4
U-LI-B
Deconychura stictolaema
12
18
U-LI-B
Dendrocincla fuliginosa
0
3
S-LI-R
Dendrocincla merula
19
20
S-LI-R
Glyphorynchus spirurus
33
37
S-SI-B
Hylexetastes perroti
1
2
U-LI-B
Sittasomus griseicapillus
1
1
U-SI-B
Xiphorhynchus elegans
20
39
U-SI-B
Xiphorhynchus guttatus
1
2
S-LI-B
Xiphorhynchus obsoletus
2
0
U-SI-B
Xiphorhynchus picus
1
0
U-SI-B
Furnariidae
Automolus infuscatus
8
6
S-LI-D
Automolus ochrolaemus
3
1
S-LI-D

120
ORDER/FAMILY
Formicaridae
SPECIES
J(N)
R(N)
GUILD
Hyloctistes subulatus
9
7
U-LI-D
Philydor erythrocercus
6
2
U-LI-D
Philydor erythropterus
3
2
C-LI-D
Philydor ruficaudatus
0
3
U-LI-D
Sclerurus caudacutus
1
0
G-SI-G
Sclerurus mexicanus
0
1
G-SI-G
Sclerurus rufigularis
4
6
G-SI-G
Synallaxis rutilans
3
15
S-SI-F
Xenops minutus
14
12
U-LI-T
Cercomacra nigrescens
4
2
S-SI-F
Conopophaga aurita
6
0
G-SI-G
Dichrozona cincta
2
0
G-SI-G
Formicarius colma
1
3
G-LI-G
Hylophylax naevia
10
7
S-SI-F
Hylophylax poecilinota
61
18
S-LI-R
Hylophylax punctulata
1
10
S-SI-F
Hypocnemis cantator
2
12
S-SI-F
Hypocnemoides melanopogon
1
0
S-SI-F
Myrmeciza hemimelaena
2
10
S-LI-F
Myrmoborus leucophrys
3
0
S-SI-F
Myrmoborus myotherinus
7
0
S-SI-F
Myrmotherula axillares
1
0
U-SI-F
Myrmotherula hauxwelli
14
16
U-SI-F
Myrmotherula leucophthalma
1
0
S-SI-D
Myrmotherula longipennis
19
4
S-SI-F

121
ORDER/FAMILY
Pipridae
Cotingidae
Tyrannidae
SPECIES
J(N)
R(N)
GUILD
Percnostola leucostigma
0
4
S-LI-F
Phlegopsis nigromaculata
10
18
S-LI-R
Rhegmatorhina hoffmannsi
14
0
S-LI-R
Sclateria naevia
1
3
S-SI-G
Thamnomanes caesius
5
0
S-LI-F
Thamnomanes saturinus
13
2
S-LI-F
Thamnophilus aethiops
5
2
S-LI-F
Thamnophilus amazonicus
4
0
U-LI-F
Thamnophilus schistaceus
3
0
U-LI-F
Chiroxiphia parela
6
5
S-FR-F
Heterocercus flavivertex
0
1
S-FR-F
Heterocercus linteatus
1
0
S-FR-F
Manacus manacus
1
0
S-FR-F
Pipra fasciicauda
21
1
S-FR-F
Pipra nattereri
61
39
S-FR-F
Pipra rubrocapilla
11
34
S-FR-F
Schiffornis turdinus
21
16
S-SI-F
Tyranneutes stolzmanni
1
0
U-SO-F
Lipaugus vociferaos
0
3
C-LO-F
Attila cinnamomeus
0
1
U-LI-F
Attila spadiceus
2
5
U-LI-F
Cnemotriccus fuscatus
1
0
S-SI-A
Elaenea parvirostris
1
0
C-SO-F
Hemitriccus zosterops
0
4
U-SI-F
Laniocera hypopyrrha
0
3
C-LO-F

122
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Leptopogon amaurocephalus
2
0
S-SI-F
Mionectes oleaginea
3
14
S-SO-F
Myiarchus ferox
1
0
C-SI-A
Myiarchus tuberculifer
1
0
C-SI-F
Myiobius barbatus
2
2
S-SI-A
Ochthoeca littoralis
1
0
W-SI-A
Onychorhynchus coronatus
1
20
U-SI-A
Platyrinchus coronatus
2
0
S-SI-F
Platyrinchus saturatus
2
0
S-SI-F
Ramphotrigon ruficauda
0
2
U-SI-F
Rhynchocyclus olivaceus
2
0
U-SI-F
Rhytipterna simplex
2
3
C-LI-F
Terenotriccus erythrurus
4
3
S-SI-A
Tolmomyias poliocephalus
2
0
C-SI-F
Hirundinidae
Stelgidopteryx ruficollis
2
0
W-SI-A
Troglodytidae
Cyphorhinus arada
4
6
G-SI-G
Microcerculus marginatus
4
10
G-SI-G
Thryothorus genibarbis
6
2
S-SI-D
Thryothorus leucotis
1
0
S-SI-F
Turdidae
Turdus albicollis
2
13
S-SO-F
Turdus amaurochalinus
0
1
S-SO-F
Parulidae
Basileuterus fulvicauda
3
1
S-SI-F
Thraupidae
Euphonia laniirostris
2
0
C-FR-F
Haba rubica
25
0
S-SO-F
Tachyphonus cristatus
1
0
C-SO-F

123
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Tachyphonus luctuosus
2
0
C-SO-F
Tachyphonus surinamus
1
0
U-SO-F
Fringillidae
Arremon taciturnus
0
5
G-SO-G
Paroaria gularis
2
0
W-SI-T
Passerina cyanoides
3
8
S-SO-F
Pitylus grossus
1
0
C-SO-F
N = number of individuals captured
J = Jusante site
R = Reserve site
Guild = Forage Strata: A, above the canopy; C, canopy; U, understory;
S, shrub; G, ground.
Diet: SI, small insects; LI, large insects; SO, small insects,
fruits and small vertebrates; LO, large insects, fruits
and small vertebrates; FR, fruit; NI, nectar; FI, fish.
Foraging Substrate: A, air; F, live foliage; D, hanging dead
foliage; B, bark; R, near army ants; G, ground.

APPENDIX C
BIRD SPECIES RECORDED AT THE SAMUEL DAM SITE
ORDER FAMILY SPECIES
Tinamiformes
Podicipediformes
Pelecaniformes
Ciciniiformes
Anseriformes
Falconiformes
Tinamidae Tinamus guttatus
Tinamus major
Crypturellus undulatus
Crypturellus noctivagus
Podicipedidae Podiceps dominicus
Anhingidae Anhinga anhinga
Phalacrocoracidae Phalacrocorax olivaceus
Ardeidae Pilherodius pileatus
Casmerodius albus
Bubulcus ibis
Ixobrychus exilis
Butoroides striatus
Tigrisoma lineatum
Egretta thula
Ciconidae Euxenura maguari
Threskiornithidae Mesembrinibis cayanensis
Anatidae Cairina moscata
Dendrocygna autumnalis
Carthartidae Sarcorhamphus papa
Cathartes aura
124

125
ORDER FAMILY
SPECIES
Cathartes burrovianus
Coragyps atratus
Accipitridae
Heterosnizias meridionalis
Buteogallus urubitinga
Leptodon cayanensis
Ictinia plmbea
Buteo nitidus
Leocopternis shistacea
Elanoides forficatus
Accipiter superciliosus
Falconidae
Daptrius americanus
Daptrius ater
Milvago chima chima
Polyborus plancus
Falco rufigularis
Herpetotheres cachinnans
Galliformes Phasianidae
Odontophorus gujanensis
Cracidae
Mi tu mitu
Notocrax urumutum
Opisthocomiformes Opisthocomidae
Notocrax sp
Opistochomus hoazin
Gruiformes Eurypygidae
Eurypyga helias
Psophiidae
Psophia viridis
Rallidae
Pomphinula martinika
Heliornithidae
Heliornis flica

126
ORDER FAMILY SPECIES
Charadriiformes
Charadriidae
Hoploxypterus cayanus
Scolopacidae
Tringa solitaria
Actitis macularia
Laridae
Sterna superciliaris
Phaetusa simplex
Jacanidae
Jacana jacana
Columbi formes
Columbidae
Columbina talpacoti
Columba cayannensis
Geotrygon montana
Leptotila rufaxilla
Leptotila verreauxi
Psittaciformes
Psittacidae
Ara ararauna
Ara chloroptera
Ara macao
Pionus menstruus
Amazona festiva
Amazona farinosa
Amazona ochrocephala
Pyrrhura picta
Pyrrhura melanura
Pionites leucogaster
Cuculiformes
Cuculidae
Piaya cayana
Piaya minuta
Crotophaga ani
Crotophaga major

127
ORDER
Caprimulgiformes
Apodiformes
Trogoniformes
Coraciiformes
Piciformes
FAMILY SPECIES
Caprimulgidae
Trochilidae
Trogonidae
Alcedinidae
Momotidae
Galbulidae
Gira gira
Neomorpha geoffroyi
Hydropsalis climatocerca
Nyctidromus albicolis
Nycphrynus ocellatus
Podager nanunda
Campylopterus largipennis
Florisuga mellivora
Phaethornis squalidus
Phaethornis sp
Phaethornis philippi
Phaethornis ruber
Thalurania furcata
Trogon melanurus
Trogon viridis
Trogon violaceus
Ceryle torquata
Chloroceryle amazona
Chloroceryle a enea
Chloroceryle americana
Chloroceryle inda
Bariphteingus sp
Momo tus momota
Glbula albirostris
Glbula ruficauda

128
ORDER FAMILY
SPECIES
Glbula dea
Bucconidae
Monasa nigrifons
Notharchus macrorhynchus
Nonnula ruficapilla
Capitonidae
Cap to dayi
Ramphastidae
Rhamphastus tucanus
Pteroglossus acari
Pteroglossus castanotis
Pteroglossus inscriptus
Pteroglossus bitorquatus
Picidae
Picumnus aurifrons
Melanerpes cruentatus
Campephilus melanoleucus
Colaptes campestris
Veniliornis affinis
Passeriformes Dendrocolaptidae
Xiphorhynchus pious
Xiphorhynchus guttatus
Xiphorhynchus elegans
Dendrocincla fuliginosa
Dendrocincla merula
Dendrocolaptes certhia
Glyphorynchus spirurus
Deconychura longicauda
Deconychura stictolaema
Helexetastes perro ti

129
ORDER FAMILY SPECIES
Sittasomus griseicapillus
Furnariidae Phylidor erythrocercus
Phylidor pyrrhodes
Phylidor ruficaudatus
Xenops minutus
Sclerulus mexicanus
Sclerulus rufigularis
Hyloctistes subulatus
Automulus infuscatus
Automulus ochrolaemus
Hypocnemoides maculicauda
Synallaxis rutilans
Formicaridae Thamnophilus schistaceus
Thamnophilus doliatus
Thamnophilus aethiops
Thamnomanes caesius
Thamnomanes saturinus
Cercomacra nigrescens
Phlegopsis nigromaculata
Cymbilaimus lineatus
Hylophylax poecilonota
Hylophylax naevia
Hylophylax punctulata
Hypocnemis cantator
Formicarius colma

130
ORDER
FAMILY SPECIES
Myrmeciza hemimelaena
Myrmoborus myotherinus
Myrmotherula hauxwelli
Myrmotherula longipennis
Myrmotherula ornata
Percnostola leucostigma
Sclateria naevia
Cotingidae Lipaugus vociferans
Iodopleura isabellae
Tityra cayana
Attila spadiceus
Pipridae Pipra rubrocapilla
Pipra nattereri
Pipra fasciicauda
Pipra sp
Machaeropterus pyrocephalus
Schiffornis turdinus
Chiroxiphia parela
Heterocercus flavivertex
Tyrannidae Xolmis irupero
Pitangus sulphuratus
Muscvora tyrannus
Tyrannus melancholicus
Megarhynchus pitanga
Myiozetetes similis

131
ORDER
FAMILY SPECIES
Myiarchus spp
Contopus virens
Tolmomyias sulphurescens
Todirostrum ciereum
Pipromorpha oleaginea
Terenotriccus erythrurus
Elaenia sp
Myiobius barbatus
Hemitriccus zosterops
Laniocerca hypopyrrha
Mionectes oleagines
Onychorhynchus coronatus
Platyrhinchus saturinus
Ramphotrigon ruficauda
Rhytiopterna simplex
Hirundinidae Atticora fasciata
Stelgidopterix ruficollis
Tachycineta albiventer
Progne chalybea
Troglodytidae Troglodytes aedon
Tryothorus genibarbis
Tryothorus coraya
Campylorhynchus turdinus
Cyphorhinus arada
Microcerculus marginatus

132
ORDER FAMILY SPECIES
Turdidae
Icteridae
Coerebidae
Thraupidae
Fringilidae
Turdus albicolis
Turdus amaurochalinus
Turdus rufiventris
Psarocolius decumanus
Psarocolius angustifrons
Cacicus cela
Cacicus haemorrhous
Molothrus bonariensis
Gymnostinops yuracares
Dacnes lineata
Cissopis leveriana
Ramphocelus carbo
Tachyphonus luctuosus
Tachyphonus cristatus
Euphonia laniirostris
Euphonia rufiventris
Tangara velia
Tangara chilensis
Thraupis episcopus
Thraupis palmarum
Thraupis sayaca
Haba rubica
Paroaria coronata
Paroaria gularis
Volatna jacaria

133
ORDER FAMILY SPECIES
Passerina cyanoides
Sporophila caerulescens
Sporophila americana
Sporophila castaneiventris
Sporophila plmbea
Myospiza aurifrons
Saltator maximus
Saltator sp
Pitylus grossus
Source: Eletronorte 1993

LIST OF REFERENCES
Baker, R. R. 1978. The Evolutionary Ecology of Animal Migration. Holmes
& Meier Publishers, Inc. New York, N.Y.
Bierregaard Jr., R. 0. 1990. Species composition and trophic
organization of the understory bird community in a central
Amazonian terra firme forest. In: Four Neotropical Rainforests. A.
H. Gentry (ed.). Yale University Press, London.
Bodmer, R. 1989. Ungulate biomass in relation to feeding strategy within
Amazonian forests. Oecologia 81:547-550.
Brawn, J. D., J. R. Karr, and J. D. Nichols. 1995. Demography of birds
in a neotropical forest: Effects of allometry, taxonomy, and
ecology. Ecology 76(1):4151.
Canaday, C. 1991. Effects of Encroachment by Industry and Agriculture on
Amazonian Forest Birds in the Cuyabeno Reserve, Equador. Master's
Thesis. University of Florida. Gainesville.
Child, G. 1968. Behaviour of Large Mammals During the Formation of Lake
Kariba. Trustees of the National Museums of Rhodesia, Mardon
Printers, Salisbury.
Clutton-Brock, T. H., P. H. Harvey. 1977. Species differences in feeding
and ranging behavior in primates. In: Primate Ecology: Studies of
Feeding and Ranging Behaviour in Lemurs, Monkeys and Apes. T. H.
Clutton-Brock (ed.). Academic Press, London.
. 1979. Home range size, population density and phylogeny in
Primates. In: Primate Ecology and Human Origins: Ecological
Influences and Social Organization. I. S. Bernstein and E. O.
Smith (eds.). Garland S. T. P. M. Press. New York.
Eisenberg, J. F. 1979. Habitat, economy, and society: some correlations
and hypothesis for the Neotropical primates. In: I. S. Bernstein
and E. 0. Smith (eds.). Ecological Influences on Social
Organization: Evolution and Adaptation. Garland S. T. P. M. Press,
New York, N.Y.
. 1981. The Mammalian Radiations: An Analysis of Trends in
Evolution, Adaptation and Behavior. University of Chicago Press,
Chicago, IL.
Eisenberg, J. F., and R. W. Thorington, Jr. 1973. A preliminary analysis
of a Neotropical mammal fauna. Biotropica 5 (3):150-161.
Eletronorte. 1985. Usina Hidroeltrica de Tucurui, Operagao Curupira -
Atividades Tcnicas (ATE-015/85). Brasilia, D.F.
. 1988a. Estudos de Impacto Ambiental (EIA), Vol. 2. Metodologa e
Diagnstico Ambiental. Brasilia, D.F.
. 1988b. Atividades Desenvolvidas pelo IBDF no Acompanhamento do
Resgate de Fauna na UHE Balbina, Relatrio Final. Brasilia.
1989. Relatrio da Operago Jamari. Brasilia.
134

135
. 1990. Usina Hidroeltrica de Samuel. Brasilia.
. 1993. Estagao Ecolgica de Samuel, Rondnia, Brasil Plano de
Manejo. Brasilia.
Emmons, L. H. 1984. Geographical variation in densities and diversities
of non-flying mammals in Amazonia. Biotropica 16(3):210-222.
Fearnside, P. M. 1989. Brazil's Balbina dam: Environment versus the
legacy of the pharaohs in Amazonia. Environmental Management
13(4):401-423.
Ferrari, S. F. 1993. Ecological differentiation in the Callitrichidae.
In: Marmosets and Tamarins: Systematics, Behaviour, and Ecology.
A. B. Rylands (ed.). Oxford University Press, Oxford.
Ferrari, S. F., and A. B. Rylands. 1994. Activity budgets and
differential visibility in field studies of three marmosets
(Callithrix spp.). Folia Primatologica 63:78-83.
Ferrari, S. F., and M. A. Lopes Ferrari. 1989. A re-evaluation of the
social organization of the Callitrichidae, with reference to the
ecological differences between genera. Folia Primatologica 52:132-
147.
Freese, C. H., P. G. Heltne, N. Castro, and G. Whitesides. 1982.
Patterns and determinants of monkey densities in Peru and Bolivia,
with notes on distribution. International Journal of Primatology
3(1):53-90.
Fyfe, D. A., and R. D. Routledge. 1991. User's Manual for the Program
TransAn Version 1.00. Master's thesis. Simon Fraser University.
Burnaby. Canada.
Greenberg, R., and J. Gradwohl. 1986. A comparative study of the social
organization of antwrens on Barro Colorado Island, Panama.
Ornithological Monographs 36:845-855.
Gribel, R. 1990. The Balbina disaster: The need to ask why? The
Ecologist 20(4):133-135.
Haffer, J. 1990. Avian species richness in tropical South America.
Studies on Neotropical Fauna and Environment 25:157-183.
Herrera, R. 1985. Nutrient cycling in Amazon forests. In: Amazonia. G.
T. Prance and T. E. Lovejoy (eds.). Pergamon Press, New Yook, N.Y.
IBGE. 1977. Geografia do Brasil. 1: Regiao Norte. I.B.G.E. Rio de
Janeiro, R.J.
Irion, G. 1984. Sedimentation and sediments of Amazonian rivers and
evolution of the Amazonian landscape since Pliocene times. In: The
Amazon: Limnology and Landscape Ecology of a Mighty Tropical River
and its Basin. H. Sioli (ed.). Dr. W. Junk Publishers, Dordrecht.
Johns. A. D. 1986. Effects of Habitat Disturbance on Rainforest Wildlife
in Brazilian Amazonia. Final Report. World Wildlife Fund U.S.
Washington, D.C.

136
. 1988. Economic development and wildlife conservation in Brazilian
Amazonia. Ambio 17(5):302-306.
Johns, A. D., and J. P. Skoruppa. 1987. Responses of rain-forest
primates to habitat disturbances: a review. International Journal
of Primatology 8 (2): 157-191.
Johnson, E. G., and R. D. Routledge. 1985. The line transect method: a
nonparametric estimator based on shape restrictions. Biometrics
41:669-679.
Jordan, C. F. 1985. Soils of the Amazon rainforest. In: Amazonia. G. T.
Prance and T. E. Lovejoy (eds.). Pergamon Press, New York, N.Y.
Junk, W. J., and J. A. S. Nunes de Mello. 1987. Impactos ecolgicos das
represas hidreltricas na bacia Amaznica Brasileira. Tubinger
Geographische Studien 95:367-385.
Karr, J. R., 1971. Structure of avian communities in selected Panama and
Illinois habitats. Ecological Monographs 41:207-231.
. 1981. Surveying birds with mist nets. Studies in Avian Biology
6:62-67.
Karr, J. R., and K. E. Freemark. 1983. Habitat selection and
environmental gradients: Dynamics in the stable" tropics.
Ecology 64(6):1481-1494.
Karr, J. R., S. K. Robinson, J. G. Blake, and R. O. Bierregaard, Jr.
1990. Birds of four netropical forests. In: Four Neotropical
Forests. A. H. Gentry (ed.). Yale University Press, London.
Kinnaird, M., and J. F. Eisenberg. 1989. A consideration of body size,
diet, and population biomass for neotropical mammals. In: Advances
in Neotropical Mammalogy. K. H. Redford and J. F. Eisenberg
(eds.). The Sandhill Crane Press Inc, Gainesville, FL.
Kinsey, W. G. 1977. Diet and feeding behavior of C. torquatus. In:
Primate Ecology; Studies of Feeding and Ranging Behaviour in
Lemurs, Monkeys and Apes. T. H. Clutton-Brock (ed.). Academic
Press, New York, N.Y.
. 1981. The titi monkeys, genus Callicebus. In: Ecology and
Behavior of Neotropical Primates, Vol. 1. A. F. Coimbra-Filho, and
R. A. Mittermeier (eds.). Academia Brasileira de Ciencias. Rio de
Janeiro, R.J.
Knight, D. H. 1978. Methods for sampling vegetation: An instructional
manual. Department of Botany, University of Wyoming, Laramie.
Liao, W., D. S. Bhargava, and J. Das. 1988. Some effects of dams on
wildlife. Environmental Conservation 15(l):68-70.
Lisboa, P. L. B. 1990. Rondnia: Colonizago e Floresta. Programa
Polonoroeste. Relatrio de Pesquisa n 9, SCT/PR-CNPq, Brasilia,
D.F.

137
Ludwig, J. A., and J. F. Reynolds. 1988. Statistical Ecology. John Wiley
& Sons, New York, N.Y.
Martinelli, L. A., R. L. Victoria, M. Z. Moreira, G. Arruda Jr., I. F.
Brown, C. A. C. Ferreira, L. F. Coelho, R. P. Lima, and W. W.
Thomas. 1988. Implantagao de Parcelas para Monitoreamento de
Dinmica Florestal na rea de Protegi Ambiental, UHE Samuel,
Rondnia. Unpublished Report.
Mittermeier, R. A., and M. G. M. van Roosmalen. 1981. Preliminary
observations on habitat utilization and diet in eight Surinam
monkeys. Folia Primatologica 36:1-39.
Mozetto, A. A., T. A. Stone, I. F. Brown, and D. L. Skole. 1990. O uso
de sistema geogrfico de informago e de sensoriamente remoto na
avaliago do impacto ambiental na Estagao Ecolgica da UHE Samuel,
Rondnia, Brasil. Intercincia 15(5):265-271.
National Research Council. 1981. Techniques for the Study of Primate
Populations Ecology. National Academy Press, Washington, D.C.
Payne, J. C. 1992. A Field Study of Techniques for Estimating Densities
of Duikers in Korup National Park, Cameroon. Master's thesis.
University of Florida. Gainesville.
Peres, C. A. 1990. Effects of hunting on Western Amazonian primate
communities. Biological Conservation 54:47-59.
Peters, R. H. 1983. The Ecological Implications of Body Size. Cambridge
University Press, Cambridge.
Pires, J. M. 1984. The Amazonian forest. In: The Amazon: Limnology and
Landscape Ecology of a Mighty Tropical River and its Basin. H.
Sioli (ed.). Dr. W. Junk Publishers, Dordrecht.
Pires, J. M., and G. T. Prance. 1985. The vegetation types of the
Brazilian Amazon. In: Amazonia. G. T. Prance and T. E. Lovejoy
(eds.). Pergamon Press, New York, N.Y.
Puertas, P., and R. E. Bodmer. 1993. Conservation of a high diversity
primate assemblage. Biodiversity and Conservation 2:586-593.
Remsen Jr., J. V., and T. A. Parker III. 1983. Contribution of river-
created habitats to bird species richness in Amazonia. Biotropica
15(3):223-231.
Richards, P. W. 1952. The Tropical Rain Forest. Cambridge University
Press, Cambridge.
Robinson, J. G., and C. H. Janson. 1987. Capuchins, squirrel monkeys,
and atelines: Socioecological convergence with old world primates.
In Primate Societies. B. B. Smuts, D. L. Cheney, S. M. Seyfarth,
R. W. Wrangham, and T. T. Struhsaker (eds.). University of Chicago
Press, Chicago, IL.
Robinson, J. G., and J. C. Ramirez. 1982. Conservation biology of
neotropical primates. In: Mammalian Biology in South America. M.
A. Mares and H. M. Genoways (eds.). Pymatuning Laboratory of
Ecology Special Publication no. 6. University of Pittsburg.
Linnesville, Pennsylvania.

138
Robinson, J. G., and K. H. Redford. 1986. Body size, diet, and
population density of Neotropical forest mammals. American
Naturalist 128:665-680.
. 1989. Body size, diet, and population variation in Neotropical
forest mammal species: Predictors of local extinctions? Advances
in Neotropical Mammalogy. K. H. Redford and J. F. Eisenberg
(eds.). The Sandhill Crane Press Inc, Gainesville, FL.
Robinson, J. G., P. C. Wright, and W. G. Kinzey. 1987. Monogamous cebids
and their relatives: Intergroup calls and spacing. In: Primate
Societies. B. B. Smuts, D. L. Cheney, S. M. Seyfarth, R. W.
Wrangham, and T. T. Struhsaker (eds.). University of Chicago
Press, Chicago, IL.
Robinson, S. K., J. Terborgh, and C. A. Munn. 1989. Lowland tropical
forest bird communities of a site in western Amazonia. In:
Biogeography and Ecology of Forest Bird Communities. A. Keast
(ed.). SPB Academic Publishing, Hague, The Netherlands.
Rylands, A. B., and A. Keuroghlian. 1988. Primate populations in
continuous forest and forest fragments in Central Amazonia. Acta
Amaznica 18 (3-4):291-307.
Rylands, A. B., and D. S. Faria. 1993. Habitats, feeding ecology and
home range size in the genus Callithrix. In: Marmosets and
Tamarins: Systematics, Behaviour, and Ecology. A. B. Rylands
(ed.). Oxford University Press, Oxford.
Salati, E. 1985. The climatology and hydrology of Amazonia. In:
Amazonia. G. T. Prance and T. E. Lovejoy (eds.). Pergamon Press,
New York, N.Y.
Sanchez, P. A. 1976. Properties and Management of Soils in the Tropics.
Wiley, New York, N.Y.
Serra, M. T. F. 1989. Planning Hydroelectric Projects. Hemisphere
2(2):80-83.
Ter Braak, C. J. F. 1987. Ordination. In: Data Analysis in Community and
Landscape Ecology. R. H. G. Jongman, C. F. G. Ter Braak, and O. F.
R. Tongeren (eds.). Pudoc. Wage.
Terborgh, J. 1983. Five New World Primates: A Study in Comparative
Ecology. Princeton University Press, Princeton, N.J.
Terborgh, J., J. W. Fitzpatrick, and L. Emmons. 1984. Annotated
checklist of bird and mammal species of Cocha Cashu Biological
Station, Manu National Park, Peru. Fieldiana (Zoology) 21:1-29.
Terborgh, J., S. K. Robinson, T. A. Parker III, C. A. Munn, and N.
Pierpont. 1990. Structure and organization of an Amazonian forest
bird community. Ecological Monographs 60(2):213-238.
Thiollay, J. M. 1994. Structure, density and rarity in an Amazonian
rainforest bird community. Journal of Tropical Ecology 10:449-481.

139
van Roosmallen, M. G. M. 1980. Habitat Preferences, Diet, Feeding
Strategy and Social Organization of the Black Spider Monkey
(Ateles p. paniscus Linnaeus 1758) in Surinam. Rijksinstituut voor
Natuurbeheer. Arnhem, The Netherlands.
van Roosmalen, M. G. M., and L. L. Klein. 1988. The spider monkeys,
genus Ateles. In: Ecology and Behavior of Neotropical Primates,
Vol. 2. R. A. Mittermeier, A. B. Rylands, A. Coimbra-Filho, and G.
A. B. Fonseca (eds.). World Wildlife Fund. Washington, D.C.
Visscher, M. N. 1981. Consideraciones sobre el uso de redes de neblina
en el anlisis de comunidades de aves en habitat tropical. Acta
Biolgica Venezuelica 11(2):89-107.
Walsh, J., and R. Gannon. 1967. Time Is Short and the Water Rises. E. P.
Dutton & Co., Inc. New York, N.Y.
Wilson, E. 0. 1988. Biodiversity. National Academy Press, Washington,
D.C.

BIOGRAPHICAL SKETCH
Rosa Maria Lemos de S was born in the city of Belo Horizonte,
Minas Gerais, Brazil. Being from a large city, and from a cosmopolitan
family, she remained unaware of nature until the images of "The Wild
Kingdom" reached her through a colored screen. It was love at first
sight! She packed her bags and journed to the University of Wisconsin-
Stevens Point, where she earned a B.S. degree in Wildlife Management.
From there, she returned to Brazil to get hands-on experience with the
"real" wildlife. She earned her Master's Degree from the University of
Brasilia working with the highly endangered species Brachyteles
arachnoides (woolly spider monkey) in one of the few remaining small
patches of the Atlantic Forest. She was then lured to the University of
Florida's Program for Studies in Tropical Conservation, where she was,
finally, introduced to the Amazon forest the ultimate place for
wildlife! The Amazon's magnificence has forever changed her image of
wilderness, and will keep her "busy" for the rest of her life with the
unending task of trying to preserve one of the few places on earth still
relatively unknown to mankind.
140

I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
K;
John F. Eisenberg
atQ^rine Ordway"
of Ecosystem Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Ronald F. Labisky /
Professor of Forest Resources
Conservation
and
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Melvin E. Sunquist
Associate Professor of Forest
Resources and Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Richard E. Bodmer
Assistant Professor of Forest
Resources and Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Douglas Jf. Levs
Associate Processor of
Zoology

This dissertation was submitted to the Graduate Faculty of the
School of Forest Resources and Conservation in the College of
Agriculture and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of Doctor of JPhilosophy.
December, 1995
Director/ Forest Resources and
Conservation
Dean, Graduate School



32
Methods
Study Site
The study was conducted in the Samuel Dam region located on the
Jamari river, in the state of Rondnia, approximately 50 km east of the
state's capital of Porto Velho (Fig. 1-1).
Prior to the filling of the reservoir, Eletronorte (the government
agency responsible for hydroelectric dam constructions in northern
Brazil) created a 21,000 ha Reserve (Estago Ecolgica de Samuel) to
compensate for the loss of 56,000 ha of forest due to the creation of
the reservoir. The Reserve is located east of the reservoir's
embankment, approximately 26 km from the dam. The other study site
(referred to as Jusante) was located 3 km below the dam, on the right
bank of the Jamari river (Fig. 2-1).
Data Collection
Five plots of 1 km2 were established in the Reserve in 1989, while
three plots were established at Jusante in 1990 (Fig. 2-1), creating 4
km of transect lines along each plot's perimeter. Transect surveys were
conducted by walking slowly (1 km/h), and stopping periodically to watch
and listen. Transects were conducted between 0630-0700 hours (depending
on the time of sunrise) and 1030-1100 hours in the morning, and between
1300 to 1700 hours in the afternoon. The number of transect samples in
each area was divided equally between morning and afternoon surveys.
Whenever possible, different transects were walked in the morning and
afternoon of the same day. If that was not, possible due to logistics,
the two daily surveys of a plot always began in the same direction to


135
. 1990. Usina Hidroeltrica de Samuel. Brasilia.
. 1993. Estagao Ecolgica de Samuel, Rondnia, Brasil Plano de
Manejo. Brasilia.
Emmons, L. H. 1984. Geographical variation in densities and diversities
of non-flying mammals in Amazonia. Biotropica 16(3):210-222.
Fearnside, P. M. 1989. Brazil's Balbina dam: Environment versus the
legacy of the pharaohs in Amazonia. Environmental Management
13(4):401-423.
Ferrari, S. F. 1993. Ecological differentiation in the Callitrichidae.
In: Marmosets and Tamarins: Systematics, Behaviour, and Ecology.
A. B. Rylands (ed.). Oxford University Press, Oxford.
Ferrari, S. F., and A. B. Rylands. 1994. Activity budgets and
differential visibility in field studies of three marmosets
(Callithrix spp.). Folia Primatologica 63:78-83.
Ferrari, S. F., and M. A. Lopes Ferrari. 1989. A re-evaluation of the
social organization of the Callitrichidae, with reference to the
ecological differences between genera. Folia Primatologica 52:132-
147.
Freese, C. H., P. G. Heltne, N. Castro, and G. Whitesides. 1982.
Patterns and determinants of monkey densities in Peru and Bolivia,
with notes on distribution. International Journal of Primatology
3(1):53-90.
Fyfe, D. A., and R. D. Routledge. 1991. User's Manual for the Program
TransAn Version 1.00. Master's thesis. Simon Fraser University.
Burnaby. Canada.
Greenberg, R., and J. Gradwohl. 1986. A comparative study of the social
organization of antwrens on Barro Colorado Island, Panama.
Ornithological Monographs 36:845-855.
Gribel, R. 1990. The Balbina disaster: The need to ask why? The
Ecologist 20(4):133-135.
Haffer, J. 1990. Avian species richness in tropical South America.
Studies on Neotropical Fauna and Environment 25:157-183.
Herrera, R. 1985. Nutrient cycling in Amazon forests. In: Amazonia. G.
T. Prance and T. E. Lovejoy (eds.). Pergamon Press, New Yook, N.Y.
IBGE. 1977. Geografia do Brasil. 1: Regiao Norte. I.B.G.E. Rio de
Janeiro, R.J.
Irion, G. 1984. Sedimentation and sediments of Amazonian rivers and
evolution of the Amazonian landscape since Pliocene times. In: The
Amazon: Limnology and Landscape Ecology of a Mighty Tropical River
and its Basin. H. Sioli (ed.). Dr. W. Junk Publishers, Dordrecht.
Johns. A. D. 1986. Effects of Habitat Disturbance on Rainforest Wildlife
in Brazilian Amazonia. Final Report. World Wildlife Fund U.S.
Washington, D.C.


114
Conclusions
It is not possible, or even desirable, to maintain the Amazon
region free from development and human impact. What must be done is to
integrate people and nature in a compatible co-existence. In order to do
so, we must study all forms of human disturbance in the Amazon and at
the same time try to minimize their consequences. Only then, may we be
able to preserve both nature and human integrity.
Eletronorte greatly improved its rescue operation policy between
its first rescue at Tucurui and the current effort at Samuel. The
practice of sending the bulk of animals to research institutions
(instead of releasing all of them), and the creation of a conservation
unit were commendable changes on the part of Eletronorte. Greater
changes, however, must be implemented in order to minimize impact and
maximize conservation of the Amazonian region; these changes will
necessitate funding both preliminary and follow-up ecological studies.


54
Table 3-9 Terrestrial diurnal mammal density estimates at the Reserve
and at Jusante. N = number of sightings, D = group density, MGS = mean
group size.
SPECIE
SITE
YEAR
N
D
MGS
IND/KM2
95% Cl
Dasyprocta
Reserve
89
67
19.73
1.1
21.7
12.72-33.68
fuliginosa
Reserve
90
63
12.12
1.0
12.1
7.56-21.37
Reserve
91
50
7.29
1.0
7.3
4.60-11.82
Jusante
90
42
4.31
1.0
4.3
2.36-07.27
Jusante
91
47
4.96
1.0
5.0
2.81-08.38
Mazama sp.
Reserve
89
24
3.26
1.0
3.3
1.41-05.29
Reserve
90
18
4.58
1.0
4.6
1.98-09.77
Reserve
91
25
3.45
1.0
3.6
1.55-06.95
Jusante
90
16
1.92
1.0
1.9
0.76-05.37
Jusante
91
33
3.04
1.1
3.3
1.52-06.06
Total terrestrial mammal density estimates
Because only the agouti and the two deer species had sample sizes
large enough to calculate densities, all terrestrial mammal sightings
were pooled to arrive at a terrestrial mammal density for the Reserve as
a whole. Sightings for terrestrial mammals comprised 39, 31, and 32% of
all mammal sightings at the Reserve for 1989, 1990, and 1991,
respectively. From 1989 to 1990, the mean group density decreased only
slightly, however, from 1990 to 1991 the total terrestrial mammal
density at the Reserve decreased 28%. The number of ind/km2 shows a more
obvious decrease in density due to differences in mean group sizes
through the years (Table 3-10, Fig. 3-11).
At Jusante the sightings for terrestrial mammals comprised 29 and
41.5% of all sightings for 1990 and 1991, respectively. Group density
showed only a slight increase from 1990 to 1991 and the density of
terrestrial mammals a slight decline (Table 3-10, Fig. 3-11) .


CHAPTER 2
CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION
Introduction
Amazonia occupies nearly 6,000,000 km2, with more than half of
this in Brazilian territory. Viewed from the air, the Amazon forest
appears quite homogeneous; however, when examined in detail,
considerable local variations of vegetation and floristic composition
are encountered (Pires and Prance 1985). The central plateau of the
Amazon Basin is limited to the north and to the south by moderate slopes
and to the west by the Andes, opening only to the east where it receives
hot and humid winds from the Atlantic. These unique characteristics plus
its latitudinal position give the Amazon region singular conditions,
such as almost constant daylength throughout the year, unchanged solar
energy at the limit of the earth's atmosphere, and little variation in
average monthly temperatures (Salati 1985).
Precipitation in the Amazon basin varies from 1,500 to > 3,000 mm
annually (IBGE 1977). A well defined dry season (from May through
September) is common in the central region.
Due to effects of high temperatures, high rainfall, and geology of
the region, the soils of Amazon rainforest have low potential for
supplying nutrients to plants. Intense weathering and leaching over
millions of years have removed the nutrients from the minerals which
form the parent material of the soil (Jordan 1985).
11


CHAPTER 4
UNDERSTORY BIRD COMMUNITY STRUCTURE AND COMPOSITION AT THE SAMUEL DAM
Introduction
The Amazonian region is recognized by its highly diverse bird
communities (Terborgh et al. 1984, Bierregaard 1990), with the greatest
concentration of species in its western portion (Haffer 1990). Bird
species richness is greater in the Amazon Basin where riverine habitats
occupy extensive areas (Remsen and Parker 1983).
The closing of the Samuel Hydroelectric Dam flood gates in the
Jamari River in November 1988 caused the submersion of approximately 107
km of pristine river bank, which represented a loss of approximately 214
km of riverine habitat. This figure considers only that portion of the
river that was completely submerged. However, an even larger stretch of
river was affected by the widening of the river caused by increased
depth as a result of the dam.
Riverine habitats differ in plant and animal composition from
adjacent habitats (Remsen and Parker 1983). Despite the fact that the
filling of the reservoir created a water-edge habitat this edge was a
terra firme forest before flooding, consequently its plant structure and
composition are different from that of the previously existing river-
edge. "The bird species composition of Amazonian river-created habitats
is generally distinct from that of adjacent terra firme forest" (Remsen
and Parker 1983, pg 226), therefore, the composition of bird species at
the reservoir's edge will be somehow different from that of river's
78


74
the reservoir's original biomass. The total biomass accounted for would
then represent 76.5% of the total original biomass. That would leave
23.5% of the biomass unaccounted for. However, I have no doubt that some
animals died of starvation, and others were preyed upon inside the
reservoir (air born predators can easily kill an animal which cannot run
to the ground for cover, or hide among entangled vegetation), others
have certainly moved to other areas besides the Reserve and Jusante, and
that would account for a significant portion of the remaining biomass.
Terrestrial Diurnal Mammal
Density
The density of both terrestrial mammals censused at the Reserve
were much higher in 1989 than in 1988 (Fig. 3-10), presumably as a
result of the migration of animals to the area as a consequence of the
rising water levels in the reservoir, plus the active release of animals
captured inside the reservoir. Dasyprocta showed a much higher increase
in density from 1988 to 1989 than Mazama. This is not only due to the
fact that they naturally occur at higher densities than Mazama
(therefore more animals migrated into the Reserve), but also to the fact
that 214 individuals of Dasyprocta were released in the Reserve (of a
total of 303 captured), as opposed to only 19 individuals of Mazama
released (out of 23 captured) (Eletronorte 1989). The constant decrease
in Dasyprocta density may be due to predation rather than dispersal. The
Reserve has several species of carnivores, which certainly also
increased in numbers due to the rising of the water level in the
reservoir area. The apparent increase in Mazama density from 1989 to
1990 might be artificial due to restrictions of the TransAn program.


86
Despite the slightly higher sampling effort at the Reserve (n = 54
days), than at Jusante (n = 48 days), both the number of total captures
and number of species were higher at Jusante than at the Reserve (n =
603 captures and 97 species, and n = 546 captures and n = 74 species,
respectively) (Fig. 4-2). The species accumulation curve at the Reserve
seems to be reaching an asymptote, while at Jusante the curve continues
to extend upwards (Fig. 4-3 and 4-4); however, the cumulative number of
captures and species do not differ significantly among sites
(Kolmogorov-Smirnov, p = 0.6325 and 0.1462, n = 102, for number of
species and number of captures, respectively). Within sites, the species
accumulation curves are almost identical for the 1 and 2 km zones, while
at the edge the number of species captured is higher than in the
interior of the forest (Fig. 4-3 and 4-4) The Jusante site differs from
the reserve in that there were fewer individuals captured at the edge
than at the 1 and 2 km zones (Fig. 4-1, Table 4-4).
Species Richness
During 1990 and 1991, a total of 1,149 individuals of 118 species
(Appendix B) were captured during 9,180 net-hours in both the Reserve
and Jusante sites (see Appendix C see for list of species with
occurrence confirmed). The two most speciose families for both areas
combined were the Formicaridae and Tyrannidae, which were represented by
25 and 20 species, respectively. At Jusante those two families were also
the most speciose, represented by 24 and 16 species, respectively. At
the Reserve the Formicaridae had the highest number of species (14),
while the Tyrannidae and the Furnaridae came in second represented by 10


100
Figure 4-10: Guild association with correspondance analysis. Arrows show
guilds most strongly associated with first axis. IL = live-foliage
gleaners, IB = bark-gleaners, IA = flycatchers, IG = terrestrial
insectivores, ID = dead-leaf gleaners, AF = army ant-followers, OM =
omnivores, FR = frugivores, WA = birds which were directly associated
with water, and NI = nectarivores.


28
type at the Samuel Dam region is the open or vine forest, commonly found
in Rondnia (Pires 1984). Furthermore, species most common in open
forests, according to Pires (1984), were described by Martinelli et al.
(1988) at the Reserve. The absence of epiphytes in the area, another
characteristic of open forest, can be related to the occurrence of a
marked dry season.
The lower basal area at Jusante may be due to the age of the
forest. According to Lisboa (1990), younger vegetational types are
located in areas that were submerged in the past. Elevation varies
between 60 and 100 m at Jusante, and between 90 and 150 m at the
Reserve. This difference in elevation supports the idea that the forest
at Jusante is younger, probably due to disturbance effects related to
its proximity to the river.
The more open forest at Jusante allows for greater penetration of
light, providing an opportunity for shrub and liana species to develop,
which creates a forest floor more densely covered by vegetation. The
denser vegetation is exemplified by the higher density of understory
trees in the 1-9 cm DBH category at Jusante, and the abundance of the
low growth musaceous palm-like species, Phanakospermum guianense, which
was encountered 88% more on the forest floor of Jusante than at the
Reserve.
Even though this study did not record species composition,
differences in the study sites were visually detected. The presence of
adult Bertholletia excelsia (Brazil nut), and Orbignya barbosiana
(babagu) were very.common at the Reserve, but rare at Jusante. On the
other hand, the species, Hevea brasiliensis (seringueira), was very
common at Jusante but never seen at the Reserve; the latter observation


41
3.68 groups/km2 (28.7 ind/km2), and by 1991 it had decreased to 1.41
groups/km2 (8.9 ind/km2) (Table 3-5, Fig. 3-2). The density of Saimir at
Jusante decreased by four fold if we consider the number of individuals
per km2. In 1990 the density was 1.69 groups/km2 (23.8 ind/km2), and by
1991 it was 0.75 groups/km2 (6.0 ind/km2) (Table 3-5, Fig. 3-2).
Cebus apella. Cebus density in the Reserve went from 5.42
groups/km2 (20.1 ind/km2) in 1989 to 6.63 groups/km2 (28.5 ind/km2) in
1990. The density in 1991 was similar to 1990 with 6.45 groups/km2 (27.1
ind/km2) (Table 3-5, Fig. 3-2). At Jusante Cebus density decreased from
1990 to 1991, going from 2.92 groups/km2 to 2.12 groups/km2. The
difference is greater if we consider the number of individuals per km2:
17.5 in 1990, and 8.1 in 1991. This difference is due to the fact that
mean group size went from 6.0 in 1990 to 3.8 in 1991 (Table 3-5,
Fig. 3-2).
Saguinus fuscicollis. The density of Saguinus in the Reserve more
than doubled from 1989 to 1990, going from 1.85 groups/km2 (6.1 ind/km2)
in 1989 to 4.08 groups/km2 (14.3 ind/km2) in 1990. Saguinus maintained a
high density in 1991: 3.7 groups/km2 (14.1 ind/km2 ) (Table 3-5, Fig. 3-
2). At Jusante there was a decrease in Saguinus density from 1990 to
1991: 5.83 groups/km2 (20.4 ind/km2) in 1990, to 3.67 groups/km2 (13.6
ind/km2) in 1991 (Table 3-5, Fig. 3-2).
Callicebus bruneus. The density of Callicebus in the Reserve was
at its highest in 1989 with 3.61 groups/km2 (7.2 ind/km2). In 1990 the
density had decreased to 0.83 groups/km2 (1.3 ind/km2), and by 1991 it
was 0.62 groups/km2 (1.2 ind/km2)(Table 3-5, Fig. 3-2). At Jusante the
density of Callicebus went from 8.33 groups/km2 (20.0 ind/km2) in 1990,
to 4.85 groups/km2 (11.6 ind/km2) in 1991 (Table 3-5, Fig. 3-2).


29
was expected because this particular species is known to occur along
water courses and not in terra firme.
The reason that the soils are so low in nutrients is because they
have been subjected to the intense weathering of the tropical climate
for many millions of years. However, a very slight difference in soil
quality within an area could be reflected in an entirely different
forest community through very fine adaptations of each community to
subtle differences in the soil (Jordan 1985). The latter could explain
the differences in species composition between the study sites.


10
Research Design
In order to document the changes that occurred after the
translocation of animals, I conducted research at the Samuel Ecological
Station during the months of June to October of 1989, 1990 and 1991. My
objectives were to sample both large diurnal mammals and understory
passerine birds both inside and outside the Reserve area.
My original experimental design included a control site for the
Reserve. The control was needed to represent mammal densities, and bird
community composition and structure in the Reserve prior to any flooding
because I started my work after the filling of the reservoir. However,
it was not possible to find an accessible area with the same
characteristics as the Reserve that was not affected by the flooding. I
then modified my original design and sampled an area downstream from the
dam (Jusante), which had similar characteristics to the area that was
flooded by the reservoir. Therefore, Jusante was a control site for the
area flooded by the reservoir, and not the Reserve.
This work represents the first attempt to understand community
changes brought about by hydroelectric dams in the Amazon. There is
still a lot of work to be done; however, I hope that the results of my
research will stimulate similar projects at future dam sites, and that
it will influence Eletronorte officials in their future environmental
decisions. Eletronorte should consider environmental impact studies and
rescue operations more seriously and make decisions based on scientific
facts rather than public indulgence.


-3
0
1
e
o
Jusante edge
A
Jusante 1 km
/

Jusante 2 km
1
0
Reserve edge
\
A
A
Reserve 1 km
0
A* A

Reserve 2 km

AXIS I
I
3


Edge
j
-3 -
Axis II
Forest interior
Figure 4-9: Correspondence analysis ordination of 18 transect lines at the Reserve and
Jusante sites. Dashed lines encircle interior forest transect that was grouped with
water edge habitat.


55
Table 3-10: Total terrestrial mammal densities for the Reserve and
Jusante for 1989, 1990, and 1991. N.SP = number of species used in
calculations, N = number of sightings, D = group density, MGS = mean
group size (based on total number of individuals sighted and total
number of sightings per year). one sighting of 21 individuals of
T. pcari was eliminated from the MGS calculations in order to avoid
unrealistic increase in MGS.
SITE
TEAR
N.SP
N
D
MGS
IND/KM2
95% Cl
Reserve
1989
07
102
21.12
1.4
29.6
14.98-32.72
Reserve
1990
09
92
20.68
1.3*
26.9
14.87-31.07
Reserve
1991
09
93
14.93
1.3
19.4
10.96-21.41
Jusante
1990
08
74
8.65
1.4
12.1
5.34-14.13
Jusante
1991
08
90
8.86
1.2
10.6
5.79-12.16
Figure 3-11: Total terrestrial mammal density (groups/km2 and
ind/km2) for the Reserve, and Jusante for 1989, 1990, and 1991.


15
Vegetation
Vegetation samples were taken for purpose of characterizing the
vegetation structure at both sites. A 1-km transect was sampled in each
of the 6 plots (soil and vegetation were sampled along the same 1-km
strip). The point-centered quarter method was used to sample large trees
(Knight 1978), which divides the area around a sample point into 4
imaginary quarters. Measurements of distance were taken from each point
to the center of the nearest tree in each quarter. Points in the
transect were 20 m apart, totaling 51 stations, and 204 trees measured
per transect. Measurements recorded included DBH (diameter at breast
height), and height (estimated). Only trees > 10 cm DBH were included.
Understory vegetation was sampled within 5 x 5 m quadrats at
alternate stations in each of the transects, totaling 26 quadrat sample
stations per transect. Within each quadrat all trees with DBH ranging
from 1 to 10 cm were measured, and all vegetation with DBH < 1 cm
(seedlings) was counted.
Results
Temperature and Precipitation
The average annual monthly temperature in the Samuel region is
27C; however, lower monthly temperatures are recorded during June,
July, and August due to a phenomenon called friagem, which is a cold
front sweeping over the continent from Antarctica (Table 2-1, Fig. 2-2).
This event, which lasts for only a few days, will lower temperatures
considerably, especially during nighttime.


24
in the categories between 10-15 m. However, average tree height found by
Martinelli et al. (17.8 m) was greater than in this study. The
difference may be due to different techniques of measuring tree height;
I estimated height visually, whereas Martinelli et al. used a clinometer
and a measuring tape. However, because both the Reserve and Jusante were
sampled similarly in this study, comparisons between categories of
height between sites should be valid.
Total tree density at the Reserve was 441.3 trees/ha (SD = 17.9,
n = 612), and 523.6 trees/ha (SD = 22.4, n = 612) at Jusante. However,
considering only trees with DBH ^ 30 cm, the density was higher at the
Reserve (90.9/ha; SD = 34.7, n = 126) than at Jusante (61.6/ha;
SD = 30.1, n = 72). Martinelli et al. (1988) reported similar findings
for the Reserve, total density was 483 trees/ha, and 80/ha for trees
l> 30 cm.
Understory
Average DBH for trees < 10 cm in both areas was very similar, 3.3
cm at the Reserve and 3.4 cm at Jusante, and did not differ in their
distribution (Chi-square, d.f. =8, p = 0.172, n = 573 and 798,
respectively) (Fig. 2-6). Similarly, average understory tree height did
not differ between sites; being 4.6 m and 5.1 m for the Reserve and
Jusante, respectively (Chi-Square, d.f. =7, p = 0.096, n = 573 and 798)
(Fig. 2-7). Seedling density was almost identical at the 2 sites, 15,015
individuals/ha (SD = 7,762, n = 78) at the Reserve, and 15,385
individuals/ha (SD = 7,953, n = 78) at Jusante. However, for trees with
DBH between 1 and 10 cm, density was 28% lower at the Reserve (2,933
individuals/ha; SD = 1,416, n = 78) than at Jusante (4,092


67
community, increased to 61.92. The community seems to be in the process
of returning to its original community structure, apparently restoring
its stability (Fig. 3-6).
Density comparisons with other western Amazonian sites
There is great variation in primate density among Amazonian sites
(Table 3-13). I limited the comparison to data from neighboring states
and/or countries in an attempt to avoid comparisons among areas with
distinctly different climate and vegetation. Only unhunted or slightly
hunted sites were used (n= 9, and n = 3, respectively). Mean primate
density for the neotropics, calculated by Robinson and Redford (1986),
is presented for reference.
It is clear that even light hunting strongly affects Ateles
density, and I cannot reject the possibility that Ateles has been hunted
at the Jusante site. There are rubber tappers in the area, and fisherman
sometimes hunt along the river. During the time I worked in the area I
observed a few hunting incidents, however, the main targets were
peccaries, deer, and the agouti. I never saw a captured or dead monkey,
and when questioned, the rubber tappers confirmed that they mainly
killed peccaries and agouties. I believe that the low density of Ateles
at Jusante was related to habitat and not to hunting.
Saimir shows an interesting situation where, in contrast to other
species, the number of groups per km2 is somewhat fixed around 2 despite
variation in the number of individuals per group.
Total density
Estimates of total primate density show a general trend of density
increase from 1989 to 1990 and then a decrease in 1991 (Table 3-6, Fig.
3-5). These changes are consistent with the assumption that the


APPENDIX C
BIRD SPECIES RECORDED AT THE SAMUEL DAM SITE
ORDER FAMILY SPECIES
Tinamiformes
Podicipediformes
Pelecaniformes
Ciciniiformes
Anseriformes
Falconiformes
Tinamidae Tinamus guttatus
Tinamus major
Crypturellus undulatus
Crypturellus noctivagus
Podicipedidae Podiceps dominicus
Anhingidae Anhinga anhinga
Phalacrocoracidae Phalacrocorax olivaceus
Ardeidae Pilherodius pileatus
Casmerodius albus
Bubulcus ibis
Ixobrychus exilis
Butoroides striatus
Tigrisoma lineatum
Egretta thula
Ciconidae Euxenura maguari
Threskiornithidae Mesembrinibis cayanensis
Anatidae Cairina moscata
Dendrocygna autumnalis
Carthartidae Sarcorhamphus papa
Cathartes aura
124


2
causes human and animal displacement and/or death, and can also bring
about extinction of species (Liao et al. 1988).
The Amazon region is well known for its high plant and animal
diversity. However, the distribution and densities of both plants and
animals are basically unknown in Amazonian regions. Despite their
negative impact on fauna and flora, hydroelectric dams provide good
opportunities for researchers to conduct detailed studies on local
distribution and densities of plants and animals if they are contacted
when the first feasibility studies begin, long before the creation of
the reservoir. Instead, the power companies invite scientist to research
the area 1 or 2 years prior the completion of the project, yielding only
a short-term evaluation. As a result of this policy, very little was
learned of the impact on wildlife resulting from the 3 dams constructed
most recently.
This study documents the impacts on avian and mammalian
communities created by the construction of the Samuel Hydroelectric
Power Plant on the Jamari River, Rondnia, Brazil.
The Samuel Hydroelectric Power Plant
The Samuel Hydroelectric Power Plant is located on the Jamari
River, a right bank tributary of the Madeira River, Rondnia, 8 45'S -
63 25'W. The site is 52 km east of the city of Porto Velho, the state's
capital, and 96 km from the confluence with the Madeira river (Fig. 1-
1). The Jamari River basin is entirely located in the State of Rondnia;
its watershed is located between 8 28' 11 07'S, and 62 36' 63
57'W. Its head waters are located in the Pacas Novos Mountain chain at


APPENDIX B
BIRD SPECIES CAPTURED DURING THIS STUDY
ORDER/FAMILY SPECIES
J(N) R(N)
Gruiformes
Eurypygidae
Columbiformes
Columbidae
Caprimulgiformes
Caprimulgidae
Apodiformes
Trochilidae
Eurypyga helias
1 0
Geotrygon montana 11 6
Leptotila verreauxi 1 0
Nyctiphrynus ocellatus 0 1
Nytidromus albicollis 0 1
Campylopterus largipennis 2
Florisuga mellivora 0
Glaucis hirsuta 1
Phaethornis philippi 11
Phaethornis ruber 0
Phaethornis superciliosus 9
Thalurania furcata 0
4
2
0
11
1
0
5
Coraciiformes
Alcedinidae
Chloroceryle aenea 0 7
Chloroceryle americana 2 0
GUILD
W-SI-W
G-FR-G
G-FR-G
A-SI-A
A-SI-A
S-NI-F
C-NI-F
S-NI-F
S-NI-F
S-NI-F
S-NI-F
S-NI-F
W-FI-W
W-FI-W
118


Crude Biomass (kg/km2)
50
Reserve
Jusante
Figure 3-7: Total primate crude biomass for the Samuel
Ecological Station during 1988 (Eletronorte 1989), 1989,
1990, and 1991 (this study), and for Jusante during
1990 and 1991 (this study).


132
ORDER FAMILY SPECIES
Turdidae
Icteridae
Coerebidae
Thraupidae
Fringilidae
Turdus albicolis
Turdus amaurochalinus
Turdus rufiventris
Psarocolius decumanus
Psarocolius angustifrons
Cacicus cela
Cacicus haemorrhous
Molothrus bonariensis
Gymnostinops yuracares
Dacnes lineata
Cissopis leveriana
Ramphocelus carbo
Tachyphonus luctuosus
Tachyphonus cristatus
Euphonia laniirostris
Euphonia rufiventris
Tangara velia
Tangara chilensis
Thraupis episcopus
Thraupis palmarum
Thraupis sayaca
Haba rubica
Paroaria coronata
Paroaria gularis
Volatna jacaria


6
Figure 1-3: Satellite image of the Samuel reservoir, and
the Samuel Ecological Station (From Instituto Nacional
de Pesquisas Espaciis INPE, August 1992).


40
densities in 1991 remained high and similar to the 1990 densities
instead of returning to values comparable to 1989. The densities of
Callicebus and Pithecia in the Reserve were at their highest in 1989,
and decreased steadily through 1990 and 1991(Table 3-5, Fig. 3-2).
With the exception of Pithecia, whose density was similar in 1990
and 1991, all other primate densities decreased substantially from 1990
to 1991 at the Jusante site (Table 3-5, Fig. 3-2).
Ateles paniscus. In 1989, the density of Ateles in the Reserve was
3.15 groups/km2 (13.2 individuals/km2), whereas in 1990 their density
increased to 6.06 groups/km2 (23.6 ind/km2). By 1991 their group density
was at 3.69 per km2 (higher than in 1989). However, the number of
individuals per km2 was 11.1 (less than in 1989) due to the fact that
the mean group size went from 4.2 in 1989 to 3.0 in 1991 (Table 3-5,
Fig. 3-2). At Jusante the density of Ateles was estimated at 0.60
groups/km2 (5.9 ind/km2) in 1990. In 1991 only one individual was
sighted during 252 km of census, making it impossible to estimate
density (Table 3-5, Fig. 3-2).
Callithrix emiliae. The density of Callithrix in the Reserve more
than doubled from 1989 to 1990, 1.41 groups/km2 (4.1 ind/km2) to 3.06
groups/km2 (10.1 ind/km2), respectively. By 1991, however, the density
had decreased to 1.82 groups/km2 (3.6 ind/km2), similar to 1989 (Table
3-5, Fig. 3-2). The density of Callithrix at Jusante decreased by 71%
from 1990 to 1991; from 3.13 groups/km2 (9.1 ind/km2) in 1990, to 0.91
groups/km2 (2.8 ind/km2) in 1991 (Table 3-5, Fig. 3-2).
Saimir ustus. The density of Saimiri in the Reserve increased
four fold from 1989 to 1990, then decreased by more than half by 1991.
The density in 1989 was 0.91 groups/km2 (6.9 ind/km2), by 1990 it was


I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
K;
John F. Eisenberg
atQ^rine Ordway"
of Ecosystem Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Ronald F. Labisky /
Professor of Forest Resources
Conservation
and
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Melvin E. Sunquist
Associate Professor of Forest
Resources and Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Richard E. Bodmer
Assistant Professor of Forest
Resources and Conservation
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Douglas Jf. Levs
Associate Processor of
Zoology


Crude Biomass (kg/km2)
57
Reserve
Jusante
Figure 3-12: A: Biomass for both Mazama and Dasyprocta at the Reserve
and at Jusante (Eletronorte 1988, this study). B: Biomass for
Mazama and Dasyprocta at Jusante plotted separately (this study).


52
Table 3-8: Body weight, crude biomass, and metabolic biomass of primates
at Jusante during 1990, and 1991. ()= sample size.
SPECIES
BODY WEIGHT
(kg)
CRUDE
(kg km-*)
BIOMASS
METABOLIC
(kg0'15km2)
BIOMASS
1990
1991
1990
1991
Ateles
6.299(29)
37.16
23.46
Cebus
2.304(142)
40.32
18.66
32.73
15.15
Pitheda
2.102(284)
15.77
15.98
13.09
13.27
Calllcebus
0.798(279)
15.96
9.26
16.89
9.79
Saimir
0.739(277)
17.59
4.43
18.97
4.78
Saguinus
0.329(128)
6.71
4.47
8.86
5.91
Callithrix
0.318(61)
2.89
0.89
3.85
1.19
TOTAL
136.40
53.69
117.85
50.07
( 36.91)
( 15.8)
( 28.99)
( 13.55)
Terrestrial Diurnal Mammals
Of all the diurnal terrestrial mammals that occur in the region,
only the agouti and the two species of deer had sample sizes sufficient
for density estimation (Tab. 3-2). Because the deer species are
sometimes difficult to identify when seen for a brief moment moving in
the understory, and to increase sample size, the two species were pooled
(Mazama sp) for the analysis.
Density estimates prior to damming
During the 1987-1988 censuses at the Reserve, the density of
Dasyprocta fuliginosa was estimated at 3.33 ind/km2, and Mazama sp. at
0.34 ind/km2 (Fig. 3-10; Eletronorte 1989).
Individuals captured and released
Three hundred and three individuals of Dasyprocta fuliginosa were
captured during the rescue operation; 214 of those individuals were


Percent of captures
91
Rank-ordered abundances
Number of species
Number of species
Figure 4-5: Rank-ordered abundance for the Reserve, Jusante
and both areas combined for 1990 and 1991.


36
Year
5
- 4.5
H-
3
a
kms walked
ind/km
J
Figure 3-1 :Number of kilometers walked and number of individual
animals seen per kilometer walked at the Reserve and at Tusante
during the study.
3.90 ( 1.80) ind/km in 1990, and 2.24 ( 0.72) ind/km in 1991 (n = 272
and 224 sightings, respectively; Fig. 3-1).
To evaluate bias in the sampling method, all sightings were
plotted according to the location on the trail where species were seen.
The result was an even distribution of sightings and species in each
area and in all three years. Therefore, there was no observer bias as to
where the animals were observed along the transects. There also was no
difference in the number of observations between morning and afternoon
censuses, indicating that animal sightings were independent of time of
day (Chi-square test = 6.85, 4 d.f., p= 0.05).


14
located 500 m from the River's edge in order to avoid flooded forest,
which has a unique but different set of vegetative characteristics.
Temperature and Precipitation
Temperature and precipitation data were collected by the
Engineering Department at the Samuel Hydroelectric Dam. The
climatological station at Samuel is located at 8 45'S 63 28'W, at an
altitude of 80 m, and has been operating since July 1977 (Eletronorte
1988a) .
Daily average temperature was recorded with the use of a maximum-
minimum temperature thermometer. Rainfall was recorded on a daily basis
with a rain gauge.
Nutrients in Soil
Soil samples were collected in 6, 1 km2 plots (plots 1, 2, 3, A,
B, and C; Fig. 2-1) Samples were taken along a straight line every 100
m for 1 km in each of the plots, totaling 11 stations in each plot. Soil
was collected, using a soil auger, at depths of 0-20 cm, 20-40 cm, and
40-60 cm in each of the stations, which yielded 33 samples/plot. Care
was taken to collect soil from a non-disturbed area (therefore top soil
would be intact), and to avoid contamination of lower depth samples by
upper samples.
Samples were kept in marked plastic bags and taken to Embrapa's
(Empresa Brasileira de Pesquisa Agropecuria) soil analysis laboratory
in Porto Velho for pH and nutrient (phosphorus [P], potassium (K],
calcium [Ca], aluminun [Al], and magnesium [Mg]) analysis within 2 days
of collection.


76
a) mortality: high population densities in 1989 and 1990 may have lead
to higher mortality due to inter- and intraspecific fighting and
interference competition. Assuming that the 1988 densities were much
lower for all terrestrial mammals (as they were for Dasyprocta and
Mazama sp.; Fig. 3-10), total mammal density in 1989 and 1990 were most
likely above the carrying capacity for the area, and a decline in
density would be expected through migration to other areas and/or higher
mortality rates. Small and mid-size mammals have relatively small home
ranges (home range increases with body size (Eisenberg 1979, Clutton-
Brook 1979)), and they are less likely to migrate long distances then
larger ones. Furthermore, migration might be expected to occur shortly
after the increase in density, and result in density changes similar to
that of the primate population (Fig. 3-5). Hence, high mortality rate is
more likely to have occurred than migration.
b)increased predation: even though I did not census the carnivore
community, it is natural to assume that their numbers also increased due
to migration into the Reserve. The increase in the number of animals in
the area, plus a probable increase in litter size, due to increased food
availability in 1989 and 1990, may have contributed to the decrease in
total terrestrial mammal densities as a consequence of higher carnivore
populations, and thus higher predation rates.
Even though group densities were virtually the same in 1990 and
1991 at Jusante, the number of ind/km2 had a 12% reduction during this
time as a result of smaller mean group size in 1991 (Table 3-10, Fig. 3-
11). The reasons for the decrease in density at Jusante are probably
similar to those for the Reserve.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E1PHX9K2F_OSO978 INGEST_TIME 2015-04-13T19:14:59Z PACKAGE AA00029831_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


72
the disturbance. My sample areas at Jusante extended up to two
kilometers away from the river bank. However, the northeast corner of
the Reserve is located approximately two kilometers from the river bank
(Fig. 2-1), and the vegetation in the Reserve had a different structure,
and somewhat different composition than that of Jusante (Chapter 2).
Based on these observations I measured the area along two kilometers on
the right and left bank of the Jamari river, and considered it as
"Jusante habitat type" flooded by the reservoir. The remaining area was
then considered to be similar to the Reserve habitat type. The
measurements indicated that 50% of the reservoir consisted of Jusante
habitat, and the other 50% by the Reserve habitat. Then, I used the 1988
biomass estimates to calculate the biomass for 50% of the reservoir
area, and assuming that the 1988 biomass values for Jusante were similar
to the 1991 values (as is the case for the Reserve), I used a biomass
value of 50 kg km'2 to calculate the remaining 50% of the reservoir's
primate biomass. The difference in biomass increase from 1989 and 1990
at the Reserve and Jusante, divided by the 1988 estimated biomass for
the reservoir should then represent the proportion of the total primate
biomass that successfully fled the reservoir area. The Reserve is
officially 210 km2 in size, however, 30 km2 were also flooded by the
reservoir, so I used the value of 180 km2 to calculate biomass for the
Reserve. To be able to make a comparison, the same area was used to
calculate total biomass for Jusante, even though both areas are located
in a continuous forest. The result is presented in Table 3-14.
If all the assumptions above are correct, the results show that at least
65% of the biomass of primates in the reservoir reached "safe" grounds
within the Reserve and Jusante sites, even though they subsequently


98
Similarity
Jusante Reserve Reserve Jusante Jusante
E0ge(*) Edge Inland Inland Inland
Net-lines
Figure 4-8: Cluster analysis of Jaccard's Similarity Index for 18
transect lines located at different zones at the Reserve and Jusante
sites.


APPENDIX A
MAMMALIAN SPECIES RECORDED AT THE SAMUEL DAM SITE.*
ORDER
FAMILY SPECIES
Marsupialia
Primates
Edentata
Didelphidae
Callithrichidae
Cebidae
Choloepidae
Dasypodidae
Caluromys lanatus
Caluromys sp
Didelphis marsupialis
Marmosa sp
Metachirus nudicaudatus
Monodelphis sp
Philander opossum
Callithrix emiliae
Saguinus fuscicollis
Aotus azarae
Ateles paniscus
Callicebus bruneus
Cebus apella
Pithecia irrorata
Saimir ustus
Choloepus juruanus
Cabassous unicinctus
Dasypus kappleri
Dasypus novemcinctus
Priodontes maximus
115


Cumulative N of species
87
Figure 4-2: Cumulative number of captures and species at the
Reserve and Jusante sites for 1990 and 1991.


104
and NI guilds at the river and lake edges, and of a greater number of
species of IG, and AF guilds in the interior of the forest (Fig. 4-10).
River edge forests are unique in their species composition; only
7-28% of species found in river edges are found in terra firme forests
(Remsen and Parker 1983). The creation of the lake edge opened niches
associated with water at the Reserve, and structurally the edge at the
Reserve is similar to the edge at the Jusante. However, the species
composition of both areas is different due to the uniqueness of river
edge species. Forest interior, on the other hand, seems to be similar in
both species composition and structure at the Reserve and Jusante sites.


45
Primates at Jusante comprised 71 and 58.5% of all mammal sightings
for 1990 and 1991, respectively. Primate group density for the area as a
whole decreased 44% from 1990 to 1991 (Table 3-6, Fig. 3-5A). Even
though a decrease in primate density occurred in both areas from 1990 to
1991, the decrease at Jusante was almost twice that of the Reserve. The
changes in densities of ind/km2 show the same pattern as the group
density changes, however, Jusante shows a more abrupt reduction in total
number of individuals than the Reserve (Fig. 3-5B).
Density changes between years at the Reserve
A cluster analysis comparing density results for all 4 years of
data for the Reserve shows that the 1988 densities had only a 25.27
degree of similarity with the 1989 densities (all comparisons excluded
Saimir because this species was not recorded during 1988) The 1990
community still only shows a 27.94 degree of similarity with 1988. By
1991, the degree of similarity with the 1988 community, increased to
61.92 (Fig. 3-6).
Table 3-6: Total primate densities for the Reserve and Jusante
for 1989, 1990, and 1991. N = number of sightings, D = group
density, MGS = mean group size (based on total number of
individuals sighted and total number of sightings per year).
SITE
YEAR
N
D
MGS
IND/KM2
95% Cl
Reserve
1989
161
17.59
3.6
63.3
13.86-22.81
Reserve
1990
209
23.25
3.9
90.7
18.92-29.02
Reserve
1991
200
18.03
3.6
64.9
13.71-23.68
Jusante
1990
185
24.51
4.4
107.8
17.62-33.46
Jusante
1991
127
13.83
3.5
48.4
8.86-23.14


Table 3-13: Primate densities in western Amazonia. Numbers per km2 (groups per km2). All sites are unhunted
with the exception of Ponta da Castanha, Yavari Miri, and Mamore, which were lightly hunted.
SITE
SPECIES
REFERENCE
Ateles
Cebus
Pithecia
Callicebus
Saimir
Saguinus
Callithrix
paniscus
apella
irrorata
bruneus
ustus
fuscicollis
emiliae
Samuel 1988
13.5
(2.6)
21.7
(4.0)
5.2
(1.4)
0.7 (0.3)
-
17.0
(2.0)
5.4
(0.4)
Eletronorte, 1989
Samuel 1989
13.2
(3.2)
20.1
(5.4)
10.3
(3.4)
7.2 (3.6)
6.9
(0.9)
6.1
(1.9)
4.1
(1.4)
This study
Samuel 1990
23.6
(6.1)
28.5
(6.6)
5.2
(2.1)
1.3 (0.8)
28.7
(3.7)
14.3
(4.1)
10.1
(3.1)
(4
Samuel 1991
11.1
(3.7)
27.1
(6.5)
3.5
(1.2)
1.2 (0.6)
8.9
(1.4)
14.1
(3.7)
3.6
(1.8)
U
Jusante 1990
5.9
(0.6)
17.5
(2.9)
7.5
(2.6)
20.0 (8.3)
23.8
(1.7)
20.4
(5.8)
9.1
(3.1)
6
Jusante 1991
-
8.1
(2.1)
7.6
(2.4)
11.6 (4.9)
6.0
(0.8)
13.6
(3.7)
2.8
(0.9)
44
Acaituba
8.8
(1.5)
7.8
(1.0)
-
-
-
-
-
Johns, 1986
Ponta da Cast.
1.3
(0.1)
11.5
(1.0)
-
-
32.0
(0.5)
-
-
Johns, 1986
Igarape Acu
-
-
-
-
-
8.9
-
Peres, 1990
Tefe
-
11.9
-
-
-
9.7
-
Peres, 1990
Urucu
-
32.2
-
-
10.2
15.0
-
Peres, 1990
Cocha Cashu
25.0
40.0
-
-
60.0
16.0
-
Terborgh, 1983
Cocha Cashu
22.4
(3.2)
36.0
(3.6)
-
-
84.0
(2.1)
10.8
(1.8)
-
Freese etal. 1982
Peru/Iquitos
-
25.0
(2.5)
-
-
72.0
(1.8)
15.0
(2.5)
-
Freese etal. 1982
Yavari Miri
2.5
11.3
37.8
-
Puertas & Bodmer
1993
Mamore
2.0
(0.4)
55.0
(5.5)
-
100.0
(2.5)
Freese etal. 1982
MEAN
11.3
12.4
-
62.3
26.9
Robinson &
Redford 1986


64
3-5), probably occurred when animals living along the Jamari river
inside the reservoir moved to these areas in search of new suitable
habitat. Because neither ai;ea is suitable habitat, the animals most
likely dispersed along the Jamari river, causing the density decrease
seen in 1991 (Fig. 3-2).
Cebus. Cebus are also classified as frugivore-omnivore, with more
than 50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989). The species has a broad habitat tolerance (Eisenberg 1979) .
They are opportunistic, and usually well able to persist in disturbed
forest (Johns and Skoruppa 1987), which makes them the most adaptable
primate species in the neotropics (Mittermeier and van Roosmalen 1981) .
It is not surprising then that they were able to maintain high
population density in the Reserve. The reduction of 54% in the number of
individuals estimated per km2 at Jusante from 1990 to 1991 (Table 3-5)
seems inconsistent with their ecology. However, because Palmae species
were more abundant in the Reserve than at Jusante (Chapter 2), and
because Cebus apella rely heavily on palms in a number of different ways
(insect foraging, fruits, seeds, flowers, and many other plant parts
(Mittermeier and van Roosmalen 1981, Terborgh 1983)), it is reasonable
to suggest that the Reserve holds a higher carrying capacity for the
species, and hence an increased ability to maintain the increased
densities.
Saguinus. The species is classified as insectivore-omnivore, with
more than 50% of their diet consisting of invertebrates (Eisenberg 1981,
Robinson and Redford 1986, 1989). According to Rylands and Keuroghlian
(1988), optimal habitat for this species includes secondary forest and


81
region (Chapter 2). Number of days sampled per zone/per site are listed
on Table 4-1.
Table 4-1: Number of mist-net days per site, year, and zone at the
Reserve and Jusante during 1990 and 1991.
ZONE
RESERVE
1990
1991
JUSANTE
1990
1991
EDGE
09
09
07
09
1 KM
09
09
07
09
2 KM
09
09
07
09
TOTAL
27
27
21
27
GRANDTOTAL
54
48
Nets were operated for 6 hours, from dawn to 1200/1300 hours. All
identified birds were sexed and aged (whenever possible), banded with
government-issued aluminum bands (except for hummingbirds), and released
at the site of capture. Measurements on body mass (g), lengths of wing,
tail, tarsus, bill (mm), and total body length (cm) were taken on all
birds, along with records of date, time, and place of capture and/or
recapture. Some specimens were collected for identification at the
Goeldi Museum in Belm, others were photographed alive for later
confirmation of field identification. Eight individual birds were not
identified at species level during the course of this study. They were
eliminated from the data set.
Data Analysis
The data set was initially described by species accumulation
curves, histograms, and tables of individuals and species captured per


27
individuals/ha; SD = 1,975, n = 78). Basal area was also lower at the
Reserve than at Jusante (3.4 m2 and 5.2 m2, respectively).
A striking difference in understory structure between the 2 sites
was the number of stemless palms. At the Reserve, the number of
individual stemless palms was 234, 65% of which belonged to the species
Orbignya barbosiana (babagu) which is of great economic importance in
the region as a source of oil. At Jusante, however, only 57 individual
stemless palms were counted, and only 8.8% of those were 0. barbosiana.
On the other hand, the palm-like species Phanakospermum guianenses
(sororoca, Musaceae) was recorded 58 times in the Reserve, but 469 times
at Jusante.
Discussion and Conclusions
The Amazon forest is heterogeneous both in the large numbers of
species within each community type, and in the large numbers of
community types in a given area. Structurally, the Reserve and Jusante
are different. DBH and tree height distribution revealed that the
Reserve has a higher and denser canopy than Jusante. Basal area and tree
density, for trees £ than 30 cm, were also much greater at the Reserve.
According to Pires and Prance,
similar types of vegetation have approximately the same biomass.
Biomass can be expressed by the basal area of trees per hectare,
using individuals of 30 cm or more in circumference [approximately
10 cm in diameter]. On this basis, the exceptionally large forests
can exceed 40 m2 of basal area. The open forests or vine forests
usually are between 18 and 24 m2. (Pires and Prance 1985, pg. 112)
Vine forests in the Marab region, on the Tocantins river, have basal
areas between 18-22 m2 (Pires 1984). In this study, basal area was 18.8
m2 at Jusante and 24.7 m2 at the Reserve, indicating that the forest


56
Biomass
At the Reserve, agouti biomass decreased 44% from 1989 to 1990,
and another 40% from 1990 to 1991. The biomass of the deer species had a
40% increase from 1989 to 1990, and then a 22% decrease in 1991 (Table
3-11). At Jusante, agouti biomass increased slightly from 1990 to 1991
(16%), however, the deer species had a 74% increase in biomass for the
same period. (Table 3-12).
There was a 688.5% increase in Mazama and Dasyprocta biomass from
1988 to 1989 (Fig. 3-12). From 1989 to 1990 the biomass remained high,
but by 1991 it had decreased 26.5%, although it was still much higher
than the 1988 biomass.
Table 3-11: Body weight, crude biomass, and metabolic biomass of agouti
and deer at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size, average weight for both species (adapted from Bodmer
1989).
SPECIES
BODY
WEIGHT (kg)
CRUDE
BIOMASS
(kg km*)
METABOLIC
BIOMASS
(kgkm*)
1989
1990
1991
1989
1990
1991
Dasyprocta
2.721(278)
59.05
32.92
19.86
45.97
25.63
15.47
Mazama sp.
20.000*
66.00
92.00
72.00
31.21
43.50
34.05
TOTAL
125.05
124.92
91.86
77.18
69.13
49.52
< 13.14)
< 5.7)
< 0)
( 10.24)
< 4.45)
< 0)
Table 3-12: Body weight, crude biomass, and metabolic biomass of
agouti and deer at Jusante during 1990, and 1991. ()= sample size,
* average weight for both species (adapted from Bodmer 1989).
SPECIES
BODY HEIGHT
(kg)
CRUDE
(kg km'*)
BIOMASS
METABOLIC
(kg 75km'2)
BIOMASS
1990
1991
1990
1991
Dasyprocta
2.721(278)
11.70
13.61
9.11
10.59
Mazama sp.
20.000*
38.00
66.00
17.97
31.21
TOTAL
49.70
79.61
27.08
41.80
< 1.65)
( 14.08)
( 1.29)
( 7.07)


38
1990 and 15 in 1991; a 16% decrease. Table 3-2 lists all species seen
during transect surveys in both sites during all three years. The list
of species seen is only a fraction of the total number of mammalian
species in the area and only represents the medium to large size
mammalian community (refer to Appendix A for a complete species list).
Of the 24 species seen during census, only 7 primate species, 2
deer, and the agouti had sample sizes large enough to estimate density.
Primates
Primate density estimates prior to damming
Primates censuses were performed at the Reserve by Eletronorte
researchers from September 1987 to February 1988 (Table 3-3, and Figure
3-2). Density estimates were based on 145 km of transect surveys.
Techniques used were comparable to the ones used in this study (National
Research Council 1981) .
Table 3-3: Primate density estimates at the Reserve prior
to the flooding of the reservoir. D = group density,
MGS = mean group size (Eletronorte 1988).Saimir were
not recorded during the 1988 censuses.
SPECIES
D
MGS
IND/KM2
Ateles paniscus
2.60
5.2
13.5
Cebus apella
4.02
5.4
21.7
Pithecia irrorata
1.38
3.8
5.2
Callicebus bruneus
0.26
2.6
0.7
Saimir ustus



Saguinus fuscicollis
2.00
8.5
17.0
Callithrix emiliae
0.36
15.0
5.4


80
Reserve, and river at Jusante) to the forest interior at a 90 degree
angle to the water course. Transects were between 2 and 4 km apart. Mist
nets were placed along transects within three zones: (1) at the water's
edge, (2) at 1 km, and (3) at 2 km distance from the water. At each
zone, 15, 12-meter-long nets (2.6 m high, with 36 mm black nylon mesh),
were placed along a trail which ran parallel to the main transect. Mist
nets were placed with the lowest shelf at the ground, and in precisely
the same location during replicate surveys. A period of at least three
weeks separated samples of the same net-line.
This research was designed to sample small understory species; no
attempt was made to sample birds by visual or auditory methods. Despite
biases associated with the use of mist nets, such as the fact that
resource availability may affect the probability of a bird's capture
(Greenberg and Gradwohl 1986, Visscher 1981), and the fact that they may
only capture 40% of the species present (Terborgh et al. 1990, Thiollay
1994), nets are a widely used technique in the study of understory bird
communities (Karr 1981), and remain the most productive technique for
this type of study where comparisons are to be made. Furthermore, mist
nets provide an opportunity to collect reliable information in a
relatively short period of time for an area of tropical forest with high
species diversity (Karr and Freemark 1983).
Data Collection
Sites were sampled for 9 days each month (three days in each
transect: one day per zone/per transect) during the months of June
through October of 1990 and 1991, which represents the dry season in the


79
edge. Will the species displaced from river edge habitats be found at
the lakes's edge? Will the creation of edge habitat at the lake's
perimeter increase bird diversity? A better understanding of the
structure and composition of bird communities in such areas is essential
for the preservation of bird diversity in the tropics.
The purpose of this chapter is to investigate the differences
and/or similarities between understory bird communities in the newly
created habitat and in its previously existing riverine habitat. In
order to make comparisons, the understory bird community was sampled
with the use of mist nets in a pristine riverine habitat (located
downstream from the dam), as well as in an area located at the edge of
the newly formed lake.
Methods
Study Site
Birds were sampled in the same sites as mammals, with the
exception of a new transect line created at the Jusante site (line D;
Fig. 2-1). For a complete description of the study site see Chapter 3.
Study Design
Three 2-km transects at the Reserve (located in areas 1, 2, and
3), and 3, 2-km transects at Jusante (located in areas A, B, and C in
1990, and B, C, and D in 1991), were sampled for understory bird
community structure and composition (transect A was substituted for
transect D in 1991 due to the proximity of its distant end to the river)
(Fig. 2-1). Each transect ran from the edge of the water (lake in the


Copyright 1995
by
Rosa Maria Lemos de S


61
The decline in density is not very apparent for Cebus and Saimir if we
only consider the number of groups per km2, however, these two species
were the only ones to show drastic reduction in mean group sizes over
the years (Table 3-5). Hence, if we consider the decline in the number
of individuals per km2, both species also show drastic decreases in
densities from 1990 to 1991.
Even though there was no sampling at Jusante during 1988 or 1989,
it is logical to assume a similar effect on primate communities in both
areas, as a consequence of the creation of the reservoir.
A buffer area for the protection of the dam turbines was created
inside the reservoir by clear cutting the forest closest to the dam,
which together with several construction projects and the concentration
of human activities near the dam could have inhibited animal migration
to Jusante. However, migration did occur, probably due to sheer
proximity of the area to the reservoir (animals stranded inside the
reservoir near the dam could probably see green forest on the other
side) (Fig. 2-1).
Despite the lack of data for Jusante in 1989, density estimates
for 6 of the 7 primates sampled at Jusante will most likely fit the
first pattern of density change described for primates at the Reserve
(density increase in 1990, followed by a sharp decrease in 1991) The
seventh species, Pithecia, fits the second pattern of density change
(density stays at similar levels from 1990 to 1991). Because no data
exist for 1989, it is not possible to determined if the third pattern
described at the Reserve (a continue decrease in density) was present at
Jusante. However, because the explanation given for this pattern in


Table 2-4: Average pH and nutrients per area, per category of depth
at Jusante ( s.d.)* Depth (cm), pH (H20) P and K (ppm), Ca, Mg,
and A1 (mEq/100 ml).
APEA
DEPTH
pH
P
K
Ca
Mg
Al
A
0-20
3.45(0.1)
3.27(1.4)
49.18(14.3)
0.13 (0.1)
0.17(0.1)
3.12(0.4)
20-40
3.73 (0.2)
1.36 (0.5)
14.73 ( 6.8)
0.04 (0.1)
0.12(0.1)
2.79 (0.3)
40-60
4.05 (0.2)
1.09 (0.3)
7.73 ( 2.6)
0.00 (0.0)
0.11 (0.1)
2.58 (0.1)
B
0-20
3.51 (0.2)
2.82 (1.3)
37.54 ( 8.2)
0.09(0.1)
0.15(0.1)
2.65 (0.8)
20-40
3.86 (0.2)
1.36 (0.7)
11.27 ( 2.3)
0.05(0.1)
0.12 (0.1)
2.12 (0.3)
40-60
4.15 (0.3)
1.73 (1.3)
10.09 ( 7.3)
0.02(0.1)
0.14(0.1)
1.90 (0.2)
C
0-20
3.71 (0.3)
5.18(2.7)
63.45 (28.4)
0.02 (0.0)
0.18(0.1)
2.92 (1.0)
20-40
4.11 (0.2)
2.64 (3.5)
29.91 (25.5)
0.01 (0.0)
0.11 (0.1)
2.41 (0.9)
40-60
4.43 (0.2)
1.27 (0.6)
13.09 ( 8.1)
0.01 (0.0)
0.11 (0.1)
2.13(0.9)
Note: Reference values for pH and nutrients (Source: EMBRAPAs Laboratory of Soil Analysis).
pH: < 4.3 = extremely acid
4.3 5.3 = strongly acid
5.4 6.5 = moderately acid
P: 0-10 ppm = low
11-30 ppm = medium
> 30 ppm = high
K: 0 45 ppm = low
46- 150 ppm = medium
> 150 ppm = high
Ca + Mg: 0.0 2.0 mEq = low
2.1 10.0 mEq = medium
> 10.0 mEq = high
Al: 0.0 0.3 mEq = low
> 0.3 mEq = high


CHAPTER 5
RECOMMENDATIONS FOR FUTURE HYDROELECTRIC DAM CONSERVATION PROGRAMS
Overview
One of the first well documented animal rescue operation in
tropical habitats was associated with the formation of Lake Kariba in
Zimbabwe (Child 1968). The rescue of game animals was carried out on
islands that formed slowly during a 5-year period following the damming
of the Zambezi River. Kariba's operation was done primarily in
relatively open tree savannas, and animals were only handled when they
could not be driven to safety. A total of 3,800 large mammals were
rescued from 1958 to 1961. Eighty-five percent of the animals rescued
were driven to or released on the mainland (Fig. 5-1).
The first major rescue operation in South America was the
Brokopondo Project in Surinam in 1964 (Walsh and Gannon 1967). The
rescue operation lasted 18 months, and involved approximately 10,000
animals. Surinam's dense vegetation was a confounding factor, and there
was concern about overpopulating the lake's shore where animals were
released; however, all animals captured were released.
Rescue operations in Brazilian Amazonia Hydroelectric Dam sites
started in 1984 with Tucurui Dam Operation (state of Par). A total of
284,211 animals were captured and released at the lake's shore without
any consideration to potential overpopulation problems, or to impacts of
released animals on the behavior of resident animals (Eletronorte 1985).
The second rescue operation in Brazilian Amazonia was Balbina (state of
105


21
In summary, both Jusante and Reserve sites had low pH values and
low levels of nutrients in the soil; however, the soil at Jusante had a
lower pH and a higher A1 content than at the Reserve indicating more
acid soils were characteristic of Jusante.
Vegetation
Canopy
Average DBH of trees was 22.3 cm (SD =14.8, n = 612) and 18.9 cm
(SD = 10, n = 612) at the Reserve and Jusante, respectively; 20.6% of
trees at the Reserve had diameters £ than 30 cm, compared to only 11.8%
at Jusante (Fig. 2-4). DBH differed between the 2 sites (Chi-Square, p <
0.006, n = 1224). The findings from the Reserve are in agreement with
those of Martinelli et al. (1988), who reported a mean DBH of 21.2 cm
with 16.6% of the trees with DBH > 30 cm from a survey prior to the
filling of the Reservoir.
Total basal area was larger at the Reserve than at Jusante; 24.7
m2, and 18.8 m2, respectively (n = 612 in both areas). The difference
was greater if only trees > 30 cm in DBH were considered; 16.7 m2 (n =
126) and 8.6 m2 (n = 72) in the Reserve and at Jusante, respectively.
The height of trees also differed between sites (Chi-Square, p <
0.0001, n = 1224). Despite the fact that average tree height at the
Reserve was only slightly greater than at Jusante, 13 m (SD =4.2, n =
612), and 12 m (SD =3.1, n = 612), respectively, 5.4% of trees at the
Reserve were i 21 m compared to 1.47% of trees at Jusante (Fig. 2-5).
Tree height distribution at the Reserve was also in conformity with
findings from Martinelli et al. (1988), with trees predominantly falling


101
(1990) also established that the number of bird species did not increase
at edge habitats in Amazonia, regardless of the type of edge. Therefore,
the creation of the reservoir did not contribute to an increase in point
diversity, as might have been expected.
Species abundance
The number of captures at the Reserve's edge was significantly
greater than at the 1 and 2 km zones, and also greater than at the
Jusante's edge (Table 4-4). The higher density at the Reserve's edge was
most likely due to the settlement of displaced individuals from the
reservoir's area into the newly created habitat. The difference in
capture rates at the 1 and 2 km zones at the Reserve and Jusante was
probably due to the difference in understory vegetation structure. The
Jusante site had a more open canopy than the Reserve, and consequently,
a greater penetration of light and higher density of shrubs and
understory trees (Chapter 2). The Jusante site had, therefore, a more
complex and probably richer understory vegetation, which explains the
higher densities at the forest interior zones of Jusante.
Guild Structure
Even though live-foliage gleaning insectivores was the guild most
speciose, and with the greatest number of captures at both sites, ant-
followers and frugivores had the highest number of captures, compared to
the number of species present in each guild (Figure 4-6). Brawn et al.
(1995) found that frugivores have a significantly greater population
size than insectivores, and that, among insectivores, ant-followers have
the largest population size. The large number of frugivores and ant-
followers captured in this study is not a sampling bias, because


4
an altitude of approximately 500 m; its total extension is approximately
560 km.
The reservoir, with a volume of 3.2 billion m3, occupies 560 km2
at normal maximum operating level of 87 m above sea level (Eletronorte
1990), and has the capacity of generating 216 MW of energy. The
reservoir, lying in a southeastern orientation, is 40 km in length; its
width spans 15 to 20 km in the first 25 km, and between 3 and 1 km for
the remaining 15 km. The construction of the hydroelectric power plant
started in 1982, and because of the flat terrain characteristic of the
region, 57 km of dikes were constructed along the right and left banks
in order to contain the river water (Fig. 1-2).
To guarantee conservation of the local fauna and flora,
Eletronorte created the Samuel Ecological Station, 20,865 ha in size,
adjacent to the area of influence by the Samuel Dam. However, roughly
20% of the Reserve was lost due to inundation after completion of the
reservoir (Mozeto et al. 1990, Eletronorte 1993) (Fig. 1-3). The
vegetation cover at the Reserve, before the flooding, consisted of 96.6%
terra firme, 2% temporarily flooded forest, and 1.4% secondary forest
(Eletronorte 1993). The only road into the Reserve area parallels the
dam (Fig. 1-3). Hunting does not occur within the Reserve because of
controlled access.
Rescue Operations... a History
To ameliorate the effects of habitat loss on wildlife, Eletronorte
employed "rescue operations" to save terrestrial animals once the
reservoir began to fill. Drowning animals is bad publicity, which


53
released at the Reserve. Twenty-three individual Mazama sp. were
captured, and 19 were released at the Reserve (Eletronorte 1989).
Density estimates after damming
Dasyprocta fuliginosa. Agouti density at the Reserve was at its
highest in 1989 with 21.7 ind/km2. It decreased thereafter, declining to
12.1 ind/km2 in 1990, and to 7.3 ind/km2 in 1991 (Table 3-9, Fig. 3-10).
At Jusante, agouti density had a slight increase from 1990 to 1991: 4.3,
and 5.0 ind/km2, respectively (Table 3-9, Fig. 3-10).
Mazama sp.. Deer density in the Reserve increased 40% from 1989 to
1990, then decreased 22% from 1990 to 1991, returning to levels similar
to 1989. Density went from 3.3 to 4.6 to 3.6 ind/km2, respectively
(Table 3-9, Fig. 3-10). At Jusante, however, there was a 74% increase in
density from 1990 to 1991, going from 1.9 to 3.3 ind/km2, respectively
(Table 3-9, Fig. 3-10).
CM
E
jr
V)
Cl
3
o
o
20
18
16
14
12
10
8
6
4
2
0
Dasyprocta
\
\
\
10 -i
8
6
4
2
0
Mazama
J=^
]Reserve
y\v| Jusante
\
\
00
cr>
o
T
00
cn
O
t
oo
oo
CT)
CT)
oo
00
cn
CD
cn
CT>
cn
CD
CD
cn
cn
CD
.
.
*
^
.
.
*
*
Year
Figure 3-10: Density estimates for Dasyprocta and Mazama at the Samuel
Ecological Station for 1988 (Eletronorte 1989), 1989, 1990, and 1991
(this study), and at Jusante in 1990 and 1991 (this study).


LIST OF REFERENCES
Baker, R. R. 1978. The Evolutionary Ecology of Animal Migration. Holmes
& Meier Publishers, Inc. New York, N.Y.
Bierregaard Jr., R. 0. 1990. Species composition and trophic
organization of the understory bird community in a central
Amazonian terra firme forest. In: Four Neotropical Rainforests. A.
H. Gentry (ed.). Yale University Press, London.
Bodmer, R. 1989. Ungulate biomass in relation to feeding strategy within
Amazonian forests. Oecologia 81:547-550.
Brawn, J. D., J. R. Karr, and J. D. Nichols. 1995. Demography of birds
in a neotropical forest: Effects of allometry, taxonomy, and
ecology. Ecology 76(1):4151.
Canaday, C. 1991. Effects of Encroachment by Industry and Agriculture on
Amazonian Forest Birds in the Cuyabeno Reserve, Equador. Master's
Thesis. University of Florida. Gainesville.
Child, G. 1968. Behaviour of Large Mammals During the Formation of Lake
Kariba. Trustees of the National Museums of Rhodesia, Mardon
Printers, Salisbury.
Clutton-Brock, T. H., P. H. Harvey. 1977. Species differences in feeding
and ranging behavior in primates. In: Primate Ecology: Studies of
Feeding and Ranging Behaviour in Lemurs, Monkeys and Apes. T. H.
Clutton-Brock (ed.). Academic Press, London.
. 1979. Home range size, population density and phylogeny in
Primates. In: Primate Ecology and Human Origins: Ecological
Influences and Social Organization. I. S. Bernstein and E. O.
Smith (eds.). Garland S. T. P. M. Press. New York.
Eisenberg, J. F. 1979. Habitat, economy, and society: some correlations
and hypothesis for the Neotropical primates. In: I. S. Bernstein
and E. 0. Smith (eds.). Ecological Influences on Social
Organization: Evolution and Adaptation. Garland S. T. P. M. Press,
New York, N.Y.
. 1981. The Mammalian Radiations: An Analysis of Trends in
Evolution, Adaptation and Behavior. University of Chicago Press,
Chicago, IL.
Eisenberg, J. F., and R. W. Thorington, Jr. 1973. A preliminary analysis
of a Neotropical mammal fauna. Biotropica 5 (3):150-161.
Eletronorte. 1985. Usina Hidroeltrica de Tucurui, Operagao Curupira -
Atividades Tcnicas (ATE-015/85). Brasilia, D.F.
. 1988a. Estudos de Impacto Ambiental (EIA), Vol. 2. Metodologa e
Diagnstico Ambiental. Brasilia, D.F.
. 1988b. Atividades Desenvolvidas pelo IBDF no Acompanhamento do
Resgate de Fauna na UHE Balbina, Relatrio Final. Brasilia.
1989. Relatrio da Operago Jamari. Brasilia.
134


39
Individuals captured and released
Primates represented 48.4% of all mammals captured, and 47% of all
animals released during the rescue/release operation (Eletronorte 1989).
The number of individuals per species captured during the operation, and
later released at the Reserve is shown in Table 3-4 (refer to Chapter 1
for details of operation). This was a relatively small number and
probably had an insignificant effect on most density shifts during the
years following flooding. Many changes in density derived form
emigration (see below).
Table 3-4: Number of primates captured at the Samuel
Dam reservoir, and number of primates released at the
Samuel Ecological Station (Eletronorte 1989).
SPECIES
NUMBERS
CAPTURED
NUMBERS
RELEASED
Aotus
104
60
Ateles
35
35
Callithrix
71
42
Saguinus
171
76
Cebus
207
180
Callicebus
348
309
Pithecia
369
324
Saimir
501
326
TOTAL
1, 806
1,352
Primate density estimates after damming
The densities of Ateles, Callithrix, and Saimir were high in 1990
and lower and approximately equal in 1989 and 1991. Cebus and Saguinus
densities were also high in 1990 in the Reserve, however, their


7
undoubtedly was a relevant factor in the decision by Eletronorte to
promote a rescue operation. However, the rescue operations created a
moral dilemma among Brazilian and foreign scientists. Was it worth
spending millions of dollars on such operations, which did not guarantee
the survival of the animals? Or was it more profitable, for
environmental conservation, to use the money to create new reserves and
to further maintain the already existing ones?
Eletronorte ignored such questions, and initiated the first rescue
operation at Tucurui in 1984-85; a total of 284,211 animals (vertebrates
and invertebrates) were captured and translocated (Eletronorte 1985).
Unfortunately, Eletronorte had not conducted a preliminary study to
define appropriate sites for release of the captured animals; thus, they
were released on the nearest piece of dry land (Johns 1986).
The Tucurui rescue operation cost US $30 million and employed 300
people (Johns 1986). The cost of rescuing 222,544 vertebrates was
$134.80 per individual; considering only the rescue of 107,094 birds and
mammals, the cost was $280.13 per individual. Furthermore, based on
crude estimates of primate densities, only 4% of the tamarins (Saguinus
midas), 6.4% of squirrel monkeys (Saimir sciureus), 6.9% of capuchin
monkeys (Cebus apella), 4.2% of saki monkeys (Chiropotes satanas), and
29.7% of howler monkeys (Alouatta belzebul) were rescued (Johns 1986).
According to Johns
these results suggest that rescue operations will remove only a
small proportion of primates. It is likely that there will be
critical overpopulation of lake fringe areas caused by the vast
majority of animals escaping unaided, which suggests that
releasing captured animals on the lake shore is worse than
useless. In fact, the real value of rescue operations is called
into question. (Johns 1986, pg. 20)


EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD
COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST
By
ROSA MARIA LEMOS DE S
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
1995


106
Amazonas) in 1987, where 19,536 animals were rescued, of which 12,687
were released at the river's edge, downstream from the dam (Eletronorte
1988b). The Samuel rescue operation was the first operation in which
only a small fraction (18%) of the animals rescued (16,000) were
Figure 5-1: Charts show percentage of animals captured which were:
ljreleased, 2)sent to research institutions, and 3)died during rescue
operations at each hydroelectric dam. Data on animal death during
rescue operation at Brokopondo were not available.


97
Ecological Similarities Among Sites
Correspondence (COA) analysis is often used because it allows for
the examination of ecological interrelationships between sampling units
and species in a single analysis. However, COA "is good primarily for
one axis, the second axis tends to be an arch, and axes after the first
are often hard to interpret" (Ludwig and Reynolds 1988, pg. 254). The
results of a COA on my data show this to be true. The first axis of the
correspondance analysis explained 34.3% of the variation, and it
separated the sampling units according to zones (Fig. 4-9). The second
axis explained 19.6% of the variation, however, its interpretation is
not clear, and will not be discussed.
The bird guilds that showed strong positive association with the
first axis were IA, and WA, while the guilds that showed strong negative
association were AF, and IG (Fig. 4-10) Guilds of flycatchers and of
birds associated with water were, thus, more abundant at edge habitats,
while ant-followers and terrestrial insectivores were more abundant at
the forest interior.
Discussion and Conclusions
Overall Site Comparison
Species richness
Edges are often considered more diverse habitats than forest
interior due to a greater complexity of vegetation structure. I expected
a greater number of species to occur at the edge zones. However, the
number of species captured at the edge, and in the 1, and 2 km zones did
not differ statistically in either site (Table 4-3). Terborgh et al.


70
There is a significant relationship between biomass and body
weight for primates, the largest the species the higher its biomass
(Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979; Robinson and
Redford 1986). This principle was observed in the 1989 biomass estimates
(Table 3-7). However, the same was not true for the following years. The
community went from having Ateles, Cebus, and Pithecia as the most
important species in 1988 and 1989 (also the three largest), to having
Ateles, Cebus, and Saimir in 1990, and Cebus, Ateles, and Saimir as
the three most important species in 1991.
Ateles and Cebus were the first and second most important species
during 1988, 1989 and 1990; however, in 1991 Cebus had greater metabolic
biomass than Ateles. This is probably due to the migration of Ateles to
adjacent forest between 1990 and 1991, and to the maintenance of the
increased density of Cebus in the Reserve. Pithecia came in third place
in 1988 and 1989, but it was replaced by Saimir in 1990 and 1991. The
high density increase of Saimir probably resulted from their migration
from the reservoir area to the Reserve.
Ateles and Cebus had similar biomass values in 1988 and 1989,
while Pithecia's biomass in 1989 was much higher than in 1988, most
likely a reflection of the increased density due to the number of
released animals. Despite the increase in Pithecia's biomass from 1988
to 1989, it is clear that the release operation did not change the
primate community significantly. The natural migration of the animals to
the Reserve, on the other hand, changed the community considerably.
The 7% increase in total primate biomass from 1988 to 1989 was
probably a consequence of the release operation; the 55% increase from
1990 to 1991 was most likely due to the natural migration of the


Percent of trees (%) Percent of trees (%)
25
DBH (cm)
DBH (cm)
Figure 2-6: Percentage of trees in each DBH (cm)
category sampled in the understory vegetation
at the Reserve and at Jusante.


3
Figure 1-1: Location of the Samuel Oam on the Jamari River


TABLE OF CONTENTS
ACKNOWLEDGMENTS iv
ABSTRACT viii
CHAPTERS
1 INTRODUCTION 1
Background 1
The Samuel Hydroelectric Power Plant 2
Rescue Operations... a History 4
The Samuel Rescue Operation 9
Research Design 10
2 CLIMATE, SOIL, AND VEGETATION IN THE SAMUEL DAM REGION 11
Introduction 11
Methods 12
Study Site 12
Temperature and Precipitation 14
Nutrients in Soil 14
Vegetation 15
Results 15
Temperature and Precipitation 15
Nutrients in Soil 18
Vegetation 21
Canopy 21
Understory 24
Discussion and Conclusions 27
3 INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A
CONSEQUENCE OF DAMMING 30
Introduction 30
Methods 32
Study Site 32
Data Collection 32
Data Analysis 33
Biomass estimation 35
Results 35
Sightings per Kilometer Walked 35
Species Observed 37
Primates 38
Primate density estimates prior to damming 38
Individuals captured and released 39
Primate density estimates after damming 39
Density and body weight of primates 42
Total primate density estimates 42
Density changes between years at the Reserve 45
Biomass 48
Terrestrial Diurnal Mammals 52
Density estimates prior to damming 52
vi


34
shape-restricted, density estimator (Payne 1992). The shape-restricted
estimator, first introduced by Johnson and Routledge (1985) and later
modified by Fyfe and Routledge (1991), involves modeling the probability
of detecting an individual as a function of its perpendicular distance
from the transect line.
TransAn requires sightings from at least four independent transect
lines to calculate confidence limits. Because I had only three transects
per site, the data were divided between morning and afternoon transects
to increase the number of transects to six (for the 1989 data the total
number of transects was 10, because there were five different plots at
the Reserve). Despite inherent biases, the transect censuses are
currently the most cost effective method to evaluate large mammal
densities in rainforests (Emmons 1984).
Table 3-1: Total number of kilometers walked during surveys, per
site, plot, and year.
KM WALKED PER YEAR
SITE
PLOT
1989
1990
1991
TOTAL
Reserve
1
48
88
84
2
48
96
84
3
48
80
84
4
36


5
36


Total
216
264
252
732
Jusante
A

80
84
B

80
84
C

80
84
Total

240
252
492
GRAND TOTAL
1,224


12
Forest physiognomy in the Amazon basin is influenced not only by
soil but also by the age of the vegetation at specific sites (Lisboa
1990) Water levels in the past and in the present have great influence
in the formation of the vegetation. Younger vegetation formations,
present today, are located in areas that were submerged in the past.
The objective of this chapter is to characterize the climate,
soil, and vegetation in the area, in order to understand the differences
and/or similarities of the study sites and their associated fauna.
Methods
Study Site
The Samuel Hydroelectric Dam is located in the state of Rondnia,
50 km east of the state capital of Porto Velho. Two sites were selected
for the study. The first site was the Samuel Ecological Station with an
area of 21,000 ha, located at the southeast border of the Samuel
Reservoir approximately 26 km (straight line) from the dam. The Reserve
was created by Eletronorte (the Brazilian Agency for Hydroelectric Power
Development in the Amazon Region) in an attempt to establish a protected
site for the animals rescued from the reservoir area during flooding.
The second site was located approximately 3 km below the dam, and will
be referred to as Jusante. The Jusante site was comparable,
floristically as well as faunistically, to 50% of the area that was
flooded by the reservoir (Fig. 2-1).
Within each of these sites, 3, 1 km2 plots were used to sample
soil and vegetation (Fig. 2-1). Plots at Jusante were deliberately


ACKNOWLEDGMENTS
Most students consider themselves lucky if they find the right
academic advisor; I consider myself blessed for having found not one but
two! John Robinson was my first advisor, who helped define my work and
gave support during the initial field work. John Eisenberg took over
when Robinson left the University, and he has provided valuable advice
and moral support throughout this lengthy portion of my work. I thank
both of them for their help and guidance! Kent Redford was greatly
responsible for my interest in the University of Florida, and together
with his family made life in Gainesville even more enjoyable. His
knowledge and keen interest in the Neotropics has often motivated me,
and for that I am thankful. I am also extremely thankful to Ron Labisky,
Jay Malcolm, Richard Bodmer, Doug Levey, and Mel Sunquist for their
comments, advice, and editorial help. Special thanks go to Richard
Bodmer for agreeing to substitute for John Robinson at such short
notice.
Field work in Rondnia would not have been possible without the
technical support from ELETRONORTE. Several people have greatly helped
with the bureaucracy, allowing me to do field work: Edgar Menezes
Cardoso, Bruno Payolla, Rubens Guilhardi, and Carlos Fabbris. I also had
iv


31
My assumption at the beginning of this study was that the
mammalian community in the Reserve could have been affected in three
ways: 1) by the release of rescued animals, 2) by the migration of
animals fleeing from the flooded reservoir (I have used the term
migration as defined by Baker 1977 throughout the text), and 3) by a
combination of both. During the rescue operation, from November 1988 to
March 1989, 2,374 mammals were released inside the Reserve (Eletronorte
1989). In addition to the release of rescued animals in the Reserve, I
expected migration of animals from the reservoir to occur, because the
forest at the reservoir was continuous with the forest at the Reserve. I
hypothesized that the Reserve would experience animal overcrowding for
an undetermined length of time, possibly surpassing the carrying
capacity for the area. My hypothesis could be tested by estimating
mammalian densities in the Reserve at different points in time and by
examining differences in biomass values for the community.
The Jusante site, on the other hand, could only have been affected
by the migration of animals from the reservoir area (because there was
no release of animals in the area), or not affected at all.
If my assumptions were correct, the noted responses in density
changes, regardless of site, would be immediate in the case of
terrestrial mammals (because they would have to flee from the rising
water), but possibly delayed for arboreal species (because they could
stay on top of trees while the vegetation was still alive). The time
frame in which density changes would occur was unknown. To increase the
probability of documenting such changes (completely or partially), the
sites were sampled repeatedly.


18
Figure 2-3: Total monthly Precipitation (mm),
at the Samuel Dam site for 1989,1990, and 1991.
Nutrients in Soil
Surface soils were extremely acid in all plots. The soils were
less acid at progressively lower depths. Amounts of phosphorus (P),
potassium (K), calcium (Ca), magnesium (Mg), and aluminum (Al) decreased
from surface to lower depths. Levels of P, Ca, and Mg were low, whereas
levels of K ranged from low to medium. The levels of Al, however, were
high in all plots and at all depths (Tables 2-3, and 2-4 ). Among sites,
levels of Al were different at all depths (ANOVA, n = 2, d.f. = 1,
p < 0.0264, 0.0108, and 0.0100, for 0-20, 20-40, and 40-60 cm depth,
respectively), and pH was different at 0-20, and 20-40 cm depth (ANOVA,
n = 2, d.f. = 1, p < 0.0450, and 0.0239, respectively). Levels of P, K,
Ca, and Mg did not differ among sites.


109
then, becomes the main issue in the survivorship of mammal populations
on hydroelectric dam sites.
The evidence for terrestrial mammals is less conclusive than
primates due to the fact that sample sizes for most species were small.
However, data on agouti and deer species suggest a similar pattern of
movement. Only 19 deer and 214 agoutis were released from November 1988
to March 1989; however, there was a 688.5% increase in deer and agouti
biomass at the release site from 1988 to 1989 (Fig. 3-12). The increase
can only be explained by the movement of animals fleeing from the rising
waters of the reservoir into the release site. Agouti and deer have,
therefore, swimming capabilities, and were able to escape the habitat
being flooded. Their biomass remained high for at least one year (from
1989 to 1990), but suffered a decrease from 1990 to 1991. The population
level at the release site may have been below carrying capacity for
those species, or their response to high density levels may be slower
than that of primates. But, again, the loss of habitat was of greater
importance than site mortality.
Data on bird community composition and structure demonstrated even
more clearly that the loss of habitat is the major concern associated
with the construction of dams. River-edge communities have a very unique
faunal species composition, with some species endemic to these sites.
The water edge created by the reservoir did not "substitute" for the
river edge, and as a consequence many species that were present at the
river's edge below the dam, were not present at the reservoir's edge
(Chapter 4). It is unknown if these species survived elsewhere, but the
fact remains that the habitat loss can never be replaced.


66
monkeys, such as Cebus, physically prevent access to fruit sources by
small ones, such as Saguinus and Callicebus. The lower densities of
Cebus at Jusante might benefit Callicebus, affecting their population
positively.
Pithecia. Pithecia are also frugivore-omnivores, with more than
50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989) They are usually found in the understory and lower to
middle parts of the canopy (Mittermeier and van Roosmalen 1981), and
they occur in gallery and both primary and secondary forest (Robinson
and Ramirez 1982). Pithecia are always rare (Mittermier and van
Roosmalen 1981, Robinson et all. 1987, Rylands and Keuroghlian 1988),
despite the fact that they have no distinct habitat preference. However,
their rarity may indicate that they are specialists within the forest
they occupy, or at least dependent on certain floristic
communities.(Rylands and Keuroghlian 1988). According to Johns and
Skoruppa (1987) Pithecia are able to feed on fruits from some of the
early colonizing trees, which might explain their higher densities at
Jusante.
Density changes between years at the Reserve
The low degree of similarity between the 1988 and 1989 densities
was most likely due to the increased densities of Pithecia, and
Callicebus, as a result of both the release of captured animals and the
movement of free ranging individuals into the area. The 1990 community
still show a low degree of similarity with 1988, possibly due to the
increase in densities as a consequence of the heavy migration of animals
to the Reserve. By 1991, the degree of similarity with the 1988


96
Similarity
4.24 -
36.16 -
68.08 -
100.00
Jusante
Edge
Jusante
1 km
Jusante
2 km
Reserve Reserve Reserve
2 km 1 km Edge
Zones
Figure 4-7: Cluster analysis of Jaccard's Similarity indexes between
zones at the Reserve and Jusante.


Figure 2-1: The Samuel Dam Reservoir area showing study sites (Reserve and Jusante), and study
plots (1, 2, 3, 4, 5, A, B, C, and D) .


37
Species Observed
The number of species recorded at the
similar for all years. Sixteen species were
1990, and 17 in 1991. At Jusante the number
Reserve during census was
recorded in 1989, 18 in
of species seen was 19, in
Table 3-2: List of species observed during all transect surveys for both
sites. NR = number of sightings for 1989, 1990, and 1991 at the Reserve
NJ = number of
sightings for 1990
and 1991 at Jusante.
ORDER
FAMILY
SPECIES
NR
NJ
Primata
Cebidae
Aotus azarae
01
01
Callicebus bruneus
29
98
Pithecia irrorata
62
37
Cebus apella
197
59
Saimiri ustus
25
24
Ateles paniscus
165
05
Callithrichidae
Callithrix emiliae
34
27
Saguinus fuscicollis
58
62
Edentata
Myrmecophagidae
Tamanda tetradactyla
02
05
Myrmecophaga tridactyla
01
00
Bradypodidae
Choloepus didactylus
00
03
Dasypodidae
Cabassous sp.
00
02
Carnivora
Procyonidae
Nasua nasua
10
12
Mustelidae
Eira barbara
11
01
Lutra longicaudis
00
01
Felidae
Felis pardalis
01
01
Panthera onca
01
00
Perissodactyla
Tapiridae
Tapirus terrestris
03
02
Artiodactyla
Tayassuidae
Tayassu tajacu
11
03
Tayassu pcari
01
00
Cervidae
Mazama americana
Mazama guazoubira "
67
49
Rodentia
Sciuridae
Sciurus sp.
10
15
Dasyproctidae
Dasyprocta fuliginosa
180
89


133
ORDER FAMILY SPECIES
Passerina cyanoides
Sporophila caerulescens
Sporophila americana
Sporophila castaneiventris
Sporophila plmbea
Myospiza aurifrons
Saltator maximus
Saltator sp
Pitylus grossus
Source: Eletronorte 1993


CHAPTER 3
INTER-YEAR DIFFERENCES IN DENSITIES AND BIOMASS OF MAMMALS AS A
CONSEQUENCE OF DAMMING
Introduction
The importance of tropical rain forests to global biodiversity is
clearly appreciated when one realizes that they cover only 7% of the
earth's land surface, but contain more than half the species of the
world's biota (Wilson 1988). Despite the importance of tropical forests,
and the fact that very little is known about their fauna and flora,
development of tropical areas is occurring at a rapid pace, and will
bring about the extinction of species. To avoid mass extinction and to
be able to guide developing agencies, a better understanding of the
communities and their responses to environmental changes is needed. The
purpose of this chapter is to document the response of mammalian
communities to environmental changes resulting from the construction of
the Samuel Hydroelectric Dam in the Amazon.
Two sites were monitored after the filling of the Samuel
Hydroelectric Dam reservoir. In the first site (a reserve), animals
captured during a rescue operation were released. This site was
monitored during 1988 (before the flooding of the reservoir), 1989,
1990, and 1991. The second site (referred to as Jusante) was an
undisturbed area located downstream from the dam, on the right bank of
the Jamari river, and it was monitored during the 1990 and 1991 field
seasons.
30


103
frequently captured species in both areas. The 12 most abundant species
in both areas accounted for 53% of all captures, despite differences in
abundance among species. A few moderately dominant species and a
majority of rare species is a typical pattern of bird community
structure of most tropical forests (Thiollay 1994).
The forest interior in both sites was more similar in species
composition than were the edge zones (Fig. 4-7) The edge of the lake at
the Reserve had the vegetation structure of a terra firme forest, even
though it was at the water's edge. This may be why the area was similar
to an intermediate zone between forest interior and river edge.
Cluster analysis of individual transect lines, grouped 1 and 2 km
transects indistinctly (Fig. 4-8). The distance from the water's edge
appears not to influence species composition beyond a certain distance.
However, the Reserve and Jusante sites were separated by forest
interior, and edge zones. Transect C-E at Jusante was grouped with the
forest interior transects despite its location at the river edge. The
transect is located at a point on the river were the bank is very steep,
therefore, its vegetation is characteristic of a terra firme forest
instead of a river edge, or varzea vegetation.
Ecological Similarities
The arrangement of Jusante and Reserve water-edge transects on one
side of the first COA axis, and of the forest interior transects on the
opposite side, indicates that despite the difference in species
composition, both sites have similar community structure (Fig. 4-9).
This arrangement was due to a greater number of species in the WA, IA,


49
There was a positive correlation between body weight and biomass
(kg km2) at the Reserve in all three years (Fig. 3-8). This correlation
is well documented (Eisenberg 1979; Clutton-Brock and Harvey 1977, 1979;
Robinson and Redford 1986).
At Jusante, Cebus contributed most to the metabolic biomass
calculations in both years. Ateles was in second place in 1990, followed
by Saimir. By 1991, only one group of Ateles was recorded at Jusante
and the contribution by Saimir had decreased drastically, leaving
Pithecia as the second largest contributor. Total primate crude biomass
decreased by 61% from 1990 to 1991 (Table 3-8). The correlation between
biomass and body weight was also positive in both years at Jusante
(Fig. 3-9).
Table 3-7: Body weight, crude biomass, and metabolic biomass of primates
at the Samuel Ecological Station during 1989, 1990, and 1991.
()= sample size.
SPECIES
BODY
WEIGHT
(kg)
CRUDE
BIOMASS
(kg km*)
METABOLIC
BIOMASS
(kgkm)
1989
1990
1991
1989
1990
1991
Ateles
6.299(29)
83.15
148.66
69.92
52.48
93.84
44.13
Cebus
2.304(142)
4 6.31
65.66
62.4 4
37.59
53.30
50.68
Pithecia
2.102(284)
21.65
10.93
7.36
17.98
9.08
6.11
Callicebus
0.798(279)
5.75
1.04
0.96
6.08
1.10
1.01
Saimir
0.739(277)
5.10
21.21
6.58
5.50
22.88
7.09
Ssguiuus
0.329(128)
2.01
4.70
4.64
2.65
6.21
6.13
Csllithrix
0.318(61)
1.30
3.21
1.14
1.74
4.28
1.52
TOTAL
165.27
255.41
153.04
124.02
190.69
160.80
(65.29)
( 109.37)
( 81.30)
< 44.75)
< 73.28)
( 56.10)


113
Follow-up Studies
Long-term studies on community changes and or adaptations to new
environmental conditions are of the utmost importance in the
understanding of impact studies. My results demonstrated that several
years of sampling are necessary to detect, and understand changes in an
animal community. Long-term commitment from hydroelectric power
companies, as well as from the scientific communities, is essential to
an understanding of environmental impacts.
Eletronorte invested millions of dollars creating the infra
structure necessary for a series of follow-up studies at Balbina and at
Samuel, but the proposed research never happened, mainly due to lack of
sustained funding. As an example, large tanks were built at Balbina for
the purpose of studying manatees as potential macrophyte controllers in
the reservoir. Nutrient surplus in the reservoir's water, due to the
vegetation present at the bottom, usually causes an explosion of such
aquatic vegetation. The proposed study was never carried out, and until
recently the tanks were empty. Recently, they were converted into
aquaculture tanks. The same is true for the construction of laboratories
for water analysis at both Balbina and Samuel; the laboratories exist
but no research is promoted by Eletronorte. Hydroelectric power
companies must be responsible for promoting and sustaining research
before, during, and after the construction of dams. The aggregate
studies must focus at finding solutions for environmental problems
created by hydroelectric dam constructions.


Ill
2. Only release animals on sites that have been studied
previously, and that have abnormally low densities, such as heavily
hunted sites.
3. Continue with the practice, established at Balbina, of donating
rescued animals to research institutions. Several health research
institutions utilize primates in their experiments. If rescued animals
were used to supply those institutions, there would be a decrease in the
number of wild animals removed from other sites for research purposes.
4. Create conservation units to compensate for the habitat loss
due to the creation of reservoirs.
5. Invest in professional staff to design, conduct, and supervise
conservation program at dam sites.
Creation of Conservation Units
Animal rescue operations are extremely expensive and for the most
part ineffective. If only a small portion of animals are rescued, the
funds usually reserved for such operations would be better used if
invested in the acquisition and maintenance of conservation units.
Creation of a conservation unit near the impacted area is of
extreme importance not only as a form of compensation for lost habitat,
but also as a place for displaced and/or released animals to move
through. As I have demonstrated, if given the opportunity, animals will
move from the flooded area into adjacent habitats. This practice may not
necessarily save their lives, but will give some individuals a better
chance of survival. A forest connection, however, must exist between the
impacted area and the conservation unit to enable the animals to move
from one area to another. Conservation units should be at least equal in


85
I I Reserve
1///I Jusante
Figure 4-1: Number of species and number of captures at each zone
at the Reserve and Jusante sites during 1990 and 1991.


119
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Chloroceryle inda
1
3
W-FI-W
Momotidae
Momotus momota
3
0
S-LO-F
Piciformes
Galbulidae
Galbula albirostris
0
4
U-LI-A
Galbula ruficauda
4
0
C-LI-A
Bucconidae
Malacoptila rufa
2
2
S-LI-F
Monasa nigrifrons
7
0
C-LI-A
Nonnula ruficapilla
4
1
S-LI-F
Capitonidae
Capito dayi
0
2
C-LO-F
Ramphastidae
Pteroglossus bitorquatus
1
1
C-LO-F
Picidae
Veniliornis affinis
1
1
U-SI-B
Passeriformes
Dendrocolaptidae
Deconychura longicauda
0
4
U-LI-B
Deconychura stictolaema
12
18
U-LI-B
Dendrocincla fuliginosa
0
3
S-LI-R
Dendrocincla merula
19
20
S-LI-R
Glyphorynchus spirurus
33
37
S-SI-B
Hylexetastes perroti
1
2
U-LI-B
Sittasomus griseicapillus
1
1
U-SI-B
Xiphorhynchus elegans
20
39
U-SI-B
Xiphorhynchus guttatus
1
2
S-LI-B
Xiphorhynchus obsoletus
2
0
U-SI-B
Xiphorhynchus picus
1
0
U-SI-B
Furnariidae
Automolus infuscatus
8
6
S-LI-D
Automolus ochrolaemus
3
1
S-LI-D


Cumulative N of species
88
80
70
60
50
40
30
20
10
Total O Edge 1 km 2 km
Cumulative N of captures
4 3: Cumulative number of captures and species per zone
for the Reserve during 1990 and 1991.


123
ORDER/FAMILY
SPECIES
J(N)
R(N)
GUILD
Tachyphonus luctuosus
2
0
C-SO-F
Tachyphonus surinamus
1
0
U-SO-F
Fringillidae
Arremon taciturnus
0
5
G-SO-G
Paroaria gularis
2
0
W-SI-T
Passerina cyanoides
3
8
S-SO-F
Pitylus grossus
1
0
C-SO-F
N = number of individuals captured
J = Jusante site
R = Reserve site
Guild = Forage Strata: A, above the canopy; C, canopy; U, understory;
S, shrub; G, ground.
Diet: SI, small insects; LI, large insects; SO, small insects,
fruits and small vertebrates; LO, large insects, fruits
and small vertebrates; FR, fruit; NI, nectar; FI, fish.
Foraging Substrate: A, air; F, live foliage; D, hanging dead
foliage; B, bark; R, near army ants; G, ground.


62
density change was the active release of animals in the Reserve, such
pattern would not be expected to appear at Jusante.
Differences in crude densities and in the degree of density
decline among species and sites are most likely due to differences in
species behavior and/or habitat requirements.
Ateles. Ateles are frugivore-herbivores (Eisenberg 1981, Robinson
and Redford 1986, 1989), with 83 to 90% of their diet consisting of
fruits and the remainder of other plant parts (van Roosmalen and Klein
1988). Because the distribution of fruits in a forest is widely
scattered, Ateles density is probably restricted by the availability of
this food type (Robinson and Ramirez 1982). Home-range size increases as
group weight increases (Eisenberg 1979), and Ateles are the largest of
all primates in the area requiring, in Surinam, 12.2 ha per individual
(van Roosmalen 1980, Robinson and Janson 1987). The new arrivals at the
Reserve were most likely displaced to areas outside the Reserve by the
resident groups due to the unavailability of fruit crops large enough to
maintain the higher population density.
Their almost complete absence from Jusante can be explained by the
fact that they are restricted to, or occur in higher densities only in
primary forest, using upper levels of canopy and emergent trees
(Mittermeier and van Roosmalen 1981, Robinson and Ramirez 1982, van
Roosmalen and Klein 1988). Because the forest at Jusante was
characterized by having lower, more open canopy with fewer emergent
trees (Chapter 2), it is not surprising to find that the Reserve
represented a more suitable habitat for the species (Fig. 3-2).
Callithrix. Callithrix are insectivore-omnivores, with more than
50% of their diet consisting of invertebrates (Eisenberg 1981, Robinson


excellent field assistance from Carry Ann Cadmam, Rodrigo Mariano,
Barroso, Antonio, and Chico. David Oren, Jos Maria Cardoso, and Chris
Canaday identified bird specimens and provided exceptional friendship.
This study was generously supported by the World Wildlife Fund -
WWF/US, Conservation International Cl, Lincoln Park Zoo "Scott
Neotropical Fund", Tropical Conservation and Development Program TCD,
Tinker Foundation, and the Program for Study in Tropical Conservation -
PSTC. Various individuals from these institutions simplified my life by
providing easy flow between finances and field work: Gustavo Fonseca,
and Sonia Rigueira from Cl, Cleber Alho, and LouAnn Dietz from WWF-US,
Steven Thompson from the Lincoln Zoo, and Kent Redford, Steven
Sanderson, and Peter Polshek from TCD. The Conselho Nacional de
Desenvolvimento Cientfico e Tecnolgico CNPq granted me a scholarship
for my academic work. The Empresa Brasileira de Pesquisa Agro-Pecuria -
EMBRAPA analyzed my soil samples. The Centro de Pesquisas para
Conservagao das Aves Silvestres CEMAVE/IBAMA supplied bird banding
permits and aluminum bands.
Life in Gainesville has been enriched by the ephemeral presence of
friends such as Jay Malcolm, Justina Ray, Joe Fragoso, Chris Canaday,
Wendy Townsend, John Payne, Ann Edwards, Miriam Marmontel, Andres
Navarro, Susan Walker, Denise Imbroise, Dener Martins, Peter Crawshaw,
Damian Rumiz, Cludio and Suzana Pdua, Rajanathan Rajaratnam, and many
others.
I am much grateful to Joao Paulo Viana for his endless help
throughout all phases of this work. His love, patience, and assistance
during the final weeks were specially cherished.
v


63
and Redford 1986, 1989). They are also adapted to feed on plant exudates
at certain times of the year to compensate for seasonal scarcities in
the availability of fruits (Ferrari and Lopes Ferrari 1989, Rylands and
Faria 1993, Ferrari 1993). They attain highest densities in second
growth forest and edge habitat.
Callithrix species tend to have larger group sizes and smaller
home-ranges than Saguinus species, and generally occur at higher
densities (Ferrari and Lopes Ferrari 1989, Rylands and Faria 1993).
Average group size and densities of Callithrix in both of my study sites
was lower than that of Saguinus (Table 3-5, Fig. 3-2), in contrast to
previous studies. This may be partly a consequence of their crypticity,
as described earlier.
Saimir. Saimir are classified as frugivore-omnivores, with more
than 50% of their diet composed of fruits, and the remainder mostly
invertebrates and vertebrates (Eisenberg 1981, Robinson and Redford
1986, 1989). They are habitat specialists, typical of flooded and
riverine forests (Eisenberg 1979, Freese et al 1982, Rylands and
Keuroghlian 1988). The species is known for its preference for more
open, secondary habitats, and they are most often encountered in liana
forests (Mittermeier and van Roosmalen 1981, Johns and Skoruppa 1987).
Neither of the study sites, the Reserve or Jusante, included flooded
forests. Despite the fact that the Jusante site is closer to the Jamari
river, and has a more open forest structure, transect censuses started
at a distance of 500 meters away from the river's edge. Therefore, high
densities of Saimir were not expected at either site. The high
densities in the Reserve in 1990, as well as the very high number of
individuals per km2 at Jusante (due to a larger mean group size; Table


Log10 density (ind/km2)
46
O Ateles A Cebus Pithecia Sai
Callicebus <0> Saguinus V Callithrix
1 .6 i- 1991
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
O

V *
O
i
2.0 2.5 5.0 5.5 4.0
Log10 body weight (g)
Figure 3-3: Relationship between density and body weight for
primates at the Samuel Ecological Station. (P = 0.0156 for
1989, 0.6338 for 1990, and 0.5726 for 1991).
Log10 body weight (g)
Figure 3-4: Relationship between density and body
weight for primates at Jusante. (P = 0.2135, and
0.6974 for 1990 and 1991 respectively). See Figure
3-3 for legends.


120
ORDER/FAMILY
Formicaridae
SPECIES
J(N)
R(N)
GUILD
Hyloctistes subulatus
9
7
U-LI-D
Philydor erythrocercus
6
2
U-LI-D
Philydor erythropterus
3
2
C-LI-D
Philydor ruficaudatus
0
3
U-LI-D
Sclerurus caudacutus
1
0
G-SI-G
Sclerurus mexicanus
0
1
G-SI-G
Sclerurus rufigularis
4
6
G-SI-G
Synallaxis rutilans
3
15
S-SI-F
Xenops minutus
14
12
U-LI-T
Cercomacra nigrescens
4
2
S-SI-F
Conopophaga aurita
6
0
G-SI-G
Dichrozona cincta
2
0
G-SI-G
Formicarius colma
1
3
G-LI-G
Hylophylax naevia
10
7
S-SI-F
Hylophylax poecilinota
61
18
S-LI-R
Hylophylax punctulata
1
10
S-SI-F
Hypocnemis cantator
2
12
S-SI-F
Hypocnemoides melanopogon
1
0
S-SI-F
Myrmeciza hemimelaena
2
10
S-LI-F
Myrmoborus leucophrys
3
0
S-SI-F
Myrmoborus myotherinus
7
0
S-SI-F
Myrmotherula axillares
1
0
U-SI-F
Myrmotherula hauxwelli
14
16
U-SI-F
Myrmotherula leucophthalma
1
0
S-SI-D
Myrmotherula longipennis
19
4
S-SI-F


94
Table 4-5: The 12 most abundant species captured at the Reserve and
Jusante sites (years combined), and the percentage of the sample that
each one represents.
RESERVE
SPECIES
%
JUSANTE
SPECIES
%
Pipra nattereri
7.1
Pipra nattereri
10.1
Xiphorhynchus elegans
7.1
Hylophylax poecilinota
10.1
Glyphorhynchus spirurus
6.8
Glyphorhynchus spirurus
5.5
Pipra rubrocapilla
6.2
Habia rubica
4.2
Dendrocincla merula
3.7
Pipra fasciicauda
3.5
Onychorhynchus coronatus
3.7
Schiffornis turdinus
3.5
Deconychura stictolaema
3.3
Xiphorhynchus elegans
3.3
Hylophylax poecilinota
3.3
Dendrocincla merula
3.2
Phlegopsis nigromaculata
3.3
Myrmotherula longipennis
3.2
Myrmotherula hauxwelli
2.9
Myrmotherula hauxwelli
2.3
Schiffornis turdinus
2.9
Rhegmatorhina hoffmannsi
2.3
Synallaxis rutilans
2.7
Xenops minutus
2.3
% OF TOTAL CAPTURE
53.1
53.4


137
Ludwig, J. A., and J. F. Reynolds. 1988. Statistical Ecology. John Wiley
& Sons, New York, N.Y.
Martinelli, L. A., R. L. Victoria, M. Z. Moreira, G. Arruda Jr., I. F.
Brown, C. A. C. Ferreira, L. F. Coelho, R. P. Lima, and W. W.
Thomas. 1988. Implantagao de Parcelas para Monitoreamento de
Dinmica Florestal na rea de Protegi Ambiental, UHE Samuel,
Rondnia. Unpublished Report.
Mittermeier, R. A., and M. G. M. van Roosmalen. 1981. Preliminary
observations on habitat utilization and diet in eight Surinam
monkeys. Folia Primatologica 36:1-39.
Mozetto, A. A., T. A. Stone, I. F. Brown, and D. L. Skole. 1990. O uso
de sistema geogrfico de informago e de sensoriamente remoto na
avaliago do impacto ambiental na Estagao Ecolgica da UHE Samuel,
Rondnia, Brasil. Intercincia 15(5):265-271.
National Research Council. 1981. Techniques for the Study of Primate
Populations Ecology. National Academy Press, Washington, D.C.
Payne, J. C. 1992. A Field Study of Techniques for Estimating Densities
of Duikers in Korup National Park, Cameroon. Master's thesis.
University of Florida. Gainesville.
Peres, C. A. 1990. Effects of hunting on Western Amazonian primate
communities. Biological Conservation 54:47-59.
Peters, R. H. 1983. The Ecological Implications of Body Size. Cambridge
University Press, Cambridge.
Pires, J. M. 1984. The Amazonian forest. In: The Amazon: Limnology and
Landscape Ecology of a Mighty Tropical River and its Basin. H.
Sioli (ed.). Dr. W. Junk Publishers, Dordrecht.
Pires, J. M., and G. T. Prance. 1985. The vegetation types of the
Brazilian Amazon. In: Amazonia. G. T. Prance and T. E. Lovejoy
(eds.). Pergamon Press, New York, N.Y.
Puertas, P., and R. E. Bodmer. 1993. Conservation of a high diversity
primate assemblage. Biodiversity and Conservation 2:586-593.
Remsen Jr., J. V., and T. A. Parker III. 1983. Contribution of river-
created habitats to bird species richness in Amazonia. Biotropica
15(3):223-231.
Richards, P. W. 1952. The Tropical Rain Forest. Cambridge University
Press, Cambridge.
Robinson, J. G., and C. H. Janson. 1987. Capuchins, squirrel monkeys,
and atelines: Socioecological convergence with old world primates.
In Primate Societies. B. B. Smuts, D. L. Cheney, S. M. Seyfarth,
R. W. Wrangham, and T. T. Struhsaker (eds.). University of Chicago
Press, Chicago, IL.
Robinson, J. G., and J. C. Ramirez. 1982. Conservation biology of
neotropical primates. In: Mammalian Biology in South America. M.
A. Mares and H. M. Genoways (eds.). Pymatuning Laboratory of
Ecology Special Publication no. 6. University of Pittsburg.
Linnesville, Pennsylvania.


Cumulative N of species
89
Figure 4-4: Cumulative number of captures and species per zone
for Jusante during 1990 and 1991.


138
Robinson, J. G., and K. H. Redford. 1986. Body size, diet, and
population density of Neotropical forest mammals. American
Naturalist 128:665-680.
. 1989. Body size, diet, and population variation in Neotropical
forest mammal species: Predictors of local extinctions? Advances
in Neotropical Mammalogy. K. H. Redford and J. F. Eisenberg
(eds.). The Sandhill Crane Press Inc, Gainesville, FL.
Robinson, J. G., P. C. Wright, and W. G. Kinzey. 1987. Monogamous cebids
and their relatives: Intergroup calls and spacing. In: Primate
Societies. B. B. Smuts, D. L. Cheney, S. M. Seyfarth, R. W.
Wrangham, and T. T. Struhsaker (eds.). University of Chicago
Press, Chicago, IL.
Robinson, S. K., J. Terborgh, and C. A. Munn. 1989. Lowland tropical
forest bird communities of a site in western Amazonia. In:
Biogeography and Ecology of Forest Bird Communities. A. Keast
(ed.). SPB Academic Publishing, Hague, The Netherlands.
Rylands, A. B., and A. Keuroghlian. 1988. Primate populations in
continuous forest and forest fragments in Central Amazonia. Acta
Amaznica 18 (3-4):291-307.
Rylands, A. B., and D. S. Faria. 1993. Habitats, feeding ecology and
home range size in the genus Callithrix. In: Marmosets and
Tamarins: Systematics, Behaviour, and Ecology. A. B. Rylands
(ed.). Oxford University Press, Oxford.
Salati, E. 1985. The climatology and hydrology of Amazonia. In:
Amazonia. G. T. Prance and T. E. Lovejoy (eds.). Pergamon Press,
New York, N.Y.
Sanchez, P. A. 1976. Properties and Management of Soils in the Tropics.
Wiley, New York, N.Y.
Serra, M. T. F. 1989. Planning Hydroelectric Projects. Hemisphere
2(2):80-83.
Ter Braak, C. J. F. 1987. Ordination. In: Data Analysis in Community and
Landscape Ecology. R. H. G. Jongman, C. F. G. Ter Braak, and O. F.
R. Tongeren (eds.). Pudoc. Wage.
Terborgh, J. 1983. Five New World Primates: A Study in Comparative
Ecology. Princeton University Press, Princeton, N.J.
Terborgh, J., J. W. Fitzpatrick, and L. Emmons. 1984. Annotated
checklist of bird and mammal species of Cocha Cashu Biological
Station, Manu National Park, Peru. Fieldiana (Zoology) 21:1-29.
Terborgh, J., S. K. Robinson, T. A. Parker III, C. A. Munn, and N.
Pierpont. 1990. Structure and organization of an Amazonian forest
bird community. Ecological Monographs 60(2):213-238.
Thiollay, J. M. 1994. Structure, density and rarity in an Amazonian
rainforest bird community. Journal of Tropical Ecology 10:449-481.


126
ORDER FAMILY SPECIES
Charadriiformes
Charadriidae
Hoploxypterus cayanus
Scolopacidae
Tringa solitaria
Actitis macularia
Laridae
Sterna superciliaris
Phaetusa simplex
Jacanidae
Jacana jacana
Columbi formes
Columbidae
Columbina talpacoti
Columba cayannensis
Geotrygon montana
Leptotila rufaxilla
Leptotila verreauxi
Psittaciformes
Psittacidae
Ara ararauna
Ara chloroptera
Ara macao
Pionus menstruus
Amazona festiva
Amazona farinosa
Amazona ochrocephala
Pyrrhura picta
Pyrrhura melanura
Pionites leucogaster
Cuculiformes
Cuculidae
Piaya cayana
Piaya minuta
Crotophaga ani
Crotophaga major


117
ORDER
FAMILY SPECIES
Mustelidae
Felis yagouaroundi
Panthera onca
Eira barbara
Procyonidae
Galictis cuja
Nasua nasua
Perissodactyla
Tapiridae
Potos flavus
Tapirus terrestris
Artiodactyla
Cervidae
Mazama americana
Tayassuidae
Mazama gouazoubira
Tayassu pcari
Tayassu tajacu
* Exclusive of Chiroptera
Source: Eletronorte 1993


128
ORDER FAMILY
SPECIES
Glbula dea
Bucconidae
Monasa nigrifons
Notharchus macrorhynchus
Nonnula ruficapilla
Capitonidae
Cap to dayi
Ramphastidae
Rhamphastus tucanus
Pteroglossus acari
Pteroglossus castanotis
Pteroglossus inscriptus
Pteroglossus bitorquatus
Picidae
Picumnus aurifrons
Melanerpes cruentatus
Campephilus melanoleucus
Colaptes campestris
Veniliornis affinis
Passeriformes Dendrocolaptidae
Xiphorhynchus pious
Xiphorhynchus guttatus
Xiphorhynchus elegans
Dendrocincla fuliginosa
Dendrocincla merula
Dendrocolaptes certhia
Glyphorynchus spirurus
Deconychura longicauda
Deconychura stictolaema
Helexetastes perro ti


84
Table 4-3: Number of species per site, zone, and year. Chi-square
test performed for each area on number of species per zone between
years. () percent of total species.
ZONE
RESERVE
1990
1991
JUS ANTE
1990
1991
EDGE
43
(69)
42
(71)
37 (56)
42
(55)
1 KM
30
(48)
31
(53)
30 (45)
39
(51)
2 KM
31
(50)
30
(51)
30 (45)
37
(49))
TOTAL
62
<84)
59
(80)
66 (68)
76
(78)
CHI-SQUARE
P=0.980
P=0.918
Overall Site Comparison
I found no significant difference in the number of species per
zone (Chi-Square, p = 0.979, n = 293). However, the sites were highly
differentiated when tested against the number of captures per zone (Chi-
Square, p = 0.000, n = 1,149 (Table 4-4, Fig. 4-1).
Table 4-4: Number of species and number of captures per site, and zone
(years combined). Chi-square test performed on number of species and
capture per zone between sites. () percent of species present and
total capture.
ZONE
SPECIES
RESERVE
JUSANTE
CAPTURES
RESERVE
JUSANTE
EDGE
56 (76)
61
(63)
266
(49)
171
(28)
1 KM
43 (58)
47
(48)
154
(28)
210
(35)
2 KM
40 (54)
46
(47)
126
(23)
222
(37)
TOTAL
74 (100)
97
(100)
546
(100)
603
(100)
CHI-SQUARE
P=0.979
P=
0.000


127
ORDER
Caprimulgiformes
Apodiformes
Trogoniformes
Coraciiformes
Piciformes
FAMILY SPECIES
Caprimulgidae
Trochilidae
Trogonidae
Alcedinidae
Momotidae
Galbulidae
Gira gira
Neomorpha geoffroyi
Hydropsalis climatocerca
Nyctidromus albicolis
Nycphrynus ocellatus
Podager nanunda
Campylopterus largipennis
Florisuga mellivora
Phaethornis squalidus
Phaethornis sp
Phaethornis philippi
Phaethornis ruber
Thalurania furcata
Trogon melanurus
Trogon viridis
Trogon violaceus
Ceryle torquata
Chloroceryle amazona
Chloroceryle a enea
Chloroceryle americana
Chloroceryle inda
Bariphteingus sp
Momo tus momota
Glbula albirostris
Glbula ruficauda


9
rescued animals, so that a follow-up would bring some understanding to
the changes brought about by the dam construction.
The Samuel Rescue Operation
The rescue operation started in November 1988, immediately after
the closing of flood gates. The objective of this operation was to
recover every animal found stranded on small islands created by the
filling of the reservoir. The operation lasted 4.5 months, during which
time 16,000 animals were rescued. The rescued animals included 6,590
arthropods, 3,729 mammals, 3,504 reptiles, 2,099 amphibians, and 78
birds (Eletronorte 1989). Of the 16,000 animals rescued, 11,417 were
sent alive to research institutions, and 1,729, which were sacrificed or
died during the operation, were sent to research institutes or museums.
Only 2,854 were released inside the Reserve. The number of animals
released at the Samuel Ecological Station was significantly less than
the number of animals released at previous dam sites. However, these
animals, combined with the animals that moved on their own to the
Reserve site, have probably caused some impact on the populations
previously inhabiting the site.
The influence of the dam and its reservoir is not a one-time
phenomena. During the dry season, the reservoir shrinks to an estimated
40% of its fullest extent. The approximately 300 km2 area which will be
exposed each year has the potential of altering adjacent biological
communities in a major way.


CHAPTER 1
INTRODUCTION
Background
Socioeconomic demands on land in Brazilian Amazonia are very high.
Land officially designated for development totals 2,100,000 km2, or 65%
of the total Amazonian area; of this area 4.4% is to be flooded by
hydroelectric development (Johns 1988). Eletrobras, the Brazilian
electric company, identified 80 potential dam sites in the Amazon region
in its Plan 2010 (Serra 1989). The objective of the agency is to
stimulate development in the region by attracting investors with
inexpensive energy sources.
Until 1980, only 2 small hydroelectric dams were operating in the
Amazon: Curu-Una, near Santarm, and Paredao, in Amap state. Each dam
impacted an area less than 100 km2 (Junk and Nunes de Mello 1987). Since
then, 3 large dams have been added to the region and are operating in
the Amazon; Tucurui, near Belm, Balbina, near Manaus, and Samuel, near
Porto Velho. Collectively, these 3 dams have flooded an area of 5,350
km2. If Eletronorte, the Brazilian Agency for Hydroelectric Power
Development in the Amazon region, succeeds in completing all the dams
projected for the Amazon in the 2010 plan, an area of roughly 100,000
km2 will be flooded (Fearnside 1989).
The flooding of such large areas has a tremendous impact on humans
and wildlife. The most significant effect is the loss of land, which
1


8
Despite the unsuccessful results of the Tucurui rescue operation,
a similar rescue operation was carried out by Eletronorte at Balbina in
1987. Although the area of Tucurui and Balbina were similar, the number
of animals rescued at Balbina was only about 10% of the number rescued
at Tucurui (Gribel 1990). The Balbina reservoir is much shallower, which
limited boat travel and thus, reduced the number of animals rescued. As
at Tucurui, the rescued animals were released with the same carefree
attitudes as before, and most likely did not survive the pressures of
high densities and hunting.
Only in 1988, after much pressure from the scientific community in
Brazil, did Eletronorte change its rescue operation policies. A new
rescue plan was developed because it was obvious that releasing animals
at random did not benefit the animals rescued or the community into
which they were released.
Museums, research institutions, universities, and zoological
gardens were contacted throughout Brazil, before the Samuel Dam rescue
operation started, and offered live specimens, study skins, and display
skins of species rescued in the Samuel dam reservoir. Priority was given
to research institutions that could provide housing facilities for the
animals, and to well-known researchers willing to work at the Samuel
site during the operation. Most importantly, an ecological station was
created by Eletronorte to be the focus of conservation studies, and the
only site of animal release.
Several inventories were carried out inside the Reserve prior to
the flooding of the reservoir, including vegetation, mammal, reptile and
bird surveys, and soil analyses. The projects were conducted in an
attempt to understand the community before the translocation of the


35
Biomass estimation
Crude biomass was calculated using average adult body weight (BW)
and population density (D) of a species: BW*D (kg km2). The average
animal weight used for biomass calculations was obtained from data
gathered by biologists working at the rescue operation during 14
November 1988 to 29 March 1989 (Eletronorte 1989) Because the rescue
operation was conducted as the reservoir was being flooded, and because
flooding was accomplished within four months (a relatively short period
of time), I assumed that the average animal weight recorded reflected
the true weight of the animals in the wild before any stressful
situation could cause weight loss.
Metabolic biomass also was calculated. It is an important
measurement because species sampled varied greatly in size, and
metabolic biomass takes into account that energy expenditure increases
allometrically with body weight to the power of 0.75 (Peters 1983). It
was then calculated as BW'75*D (kg0'75 Km2) The density, biomass, and
body weight estimates were ln-transformed to linearize the data.
Results
Sightings per Kilometer Walked
The number of individuals sighted per km in the Reserve was very
similar for all three years. The mean number of individuals sighted per
km walked in the Reserve was 3.34 (SD = 0.49; n = 269 sightings) in
1989, 3.59 (SD = 0.47; n = 304 sightings) in 1990, and 3.34 (SD = 0.39;
n = 296 sightings) in 1991 (Fig. 3-1). At Jusante, the mean number of
individuals sighted per km walked was 43% lower in 1991 than in 1990:


59
died" (Child 1968, pg. 37). Because primates spend most of their time in
the mid to upper forest strata, it is reasonable to assume that the vast
majority of the primate population was still living inside the reservoir
when the 1989 survey was carried out at the Reserve. This assumption can
be corroborated by the fact that only 1,806 primates were captured
during the rescue operation in a 56,000 ha area (0.03 ind/ha)
(Eletronorte 1989). In contrast, at the Tucurui Dam site (which has a
much deeper reservoir), a total of 27,039 primates were captured in a
243,000 ha area (0.11 ind/ha) (Eletronorte 1985).
When I arrived at the Samuel Dam in May 1990 all vegetation inside
the reservoir was dead. The only exception was in the higher elevation
lands, which formed green islands inside the reservoir. Because the
reservoir is bordered by dikes on both sides, and by a paved highway on
the left river bank, a natural escape route for displaced animals was
the Reserve (Fig. 1-3, and 2-1) Thus, the increase in primate densities
in the Reserve was likely a result of the natural migration of animals
into the Reserve between August 1989 and May 1990, caused by the loss of
habitat inside the reservoir.
This scenario does not explain the third pattern of density
changes detected in Callicebus and Pithecia, whose population declined
steadily after 1989 However, Callicebus, Pithecia, and Saimir were
the most frequently captured species, with over 300 individuals released
(Table 3-3). This suggests that the 1989 census in the Reserve
documented the density increase in these species as a result of the
release operation. This argument is even more convincing when density
estimates from a study done in 1988 (prior to the flooding), is
examined (Fig. 3-2).


82
site, year, and zone. Chi-square tests were used to compare number of
captures and species between years within zones and sites. This was done
in the interest of combining data from both years within each site.
Similarity among the species detected at the Reserve and Jusante,
and among species detected in each zone (within and across areas) was
quantified using Jaccard's Coefficient of Similarity, which equals zero
if no species are in common and one if the lists are identical. Cluster
analyses (group average strategy) were performed using the results of
the similarity index in order to identify any resemblance between sites
(Ludwig and Reynolds 1988).
To examine differences and/or similarities between understory bird
community structure at the Reserve and Jusante sites, Correspondance
Analysis (COA) were performed using the program CANOCO Version 3.12 (Ter
Braak 1987).
For the COA analysis, species were first grouped into 6 foraging
strata, 7 diet, and 8 substrata categories according to Karr et al.
(1990). The initial arrangement resulted in 40 different categories,
which were then condensed to 10 guild categories; insectivorous guilds
were based on primary foraging substrate. The guild categories included
six insectivore guilds (flycatchers=sallying/hawking/snatching [IA],
dead-leaf gleaners [ID], live-foliage gleaners [IL], terrestrial [IG],
bark-gleaners [IB], and army ant-followers [AF]), one nectarivore guild
(NI), one frugivore guild (FR), one omnivore guild (OM), and one guild
of birds which were directly associated with water (WA). The latter
guild includes piscivorous species, as well as insectivorous species
which catch insects over the water. The species Xenops minutus was the
only species that uses twigs as a substrate, so it was grouped with


33
give an interval of six hours between the morning and the afternoon
survey (i.e. the same point in the trail would be traversed in the
afternoon six hours after the morning survey). Transect surveys on
different days began at opposite ends of the route to reduce potential
biases resulting from direction of travel by the observer. Each transect
had equal numbers of surveys originating in both directions. The time,
transect identification, location on the trail, species, number of
individuals sighted, angle of sighting, and distance from the observer
to the animal when first seen was recorded for every non-flying mammal
encountered. Surveys were conducted by myself, an assistant, and two
field helpers.
The study was conducted from May to August 1989 and from May to
October of 1990 and 1991. These months correspond to the dry season in
the region (refer to Chapter 2 for detail). During 1989, due to
logistical problems and to the lesser amount of time spent at the study
site, data were gathered only at the Samuel Reserve.
Two of the five plots sampled in 1989 in the Reserve were
abandoned in 1990 and 1991 due to logistic difficulties. However,
observations collected in these areas were included in the 1989 analysis
to arrive at a density estimate for the entire site. Data from the
various plots at a site were pooled within years to give an overall
density estimate. The number of kilometers walked per site, plot, and
year are listed on Table 3-1.
Data Analysis
Data were analyzed using the computer program TransAn, version
1.00. TransAn is a flexible computer program that uses a non-parametric,


108
density changes at the Samuel release site demonstrated not only that
they can move from the flooded area on their own, but also that the
site's carrying capacity will determine if they will stay on the release
area or not. Of the 3,729 mammals rescued from the Samuel reservoir,
only 2,374 were released (1,352 of these were primates). According to
biomass calculations (Table 3-7, Fig 3-7), the release of primates
caused only a slight increase in total primate biomass (7%) from 1988
(prior to the release) to 1989 (immediately after the release). From
1989 to 1990, total primate biomass increased 55%; this increase only
can be explained by the movement of primates from the flooded reservoir
to the release area. The movement of animals after the flooding
confirmed that primates can move on their own (provided that the filling
of the reservoir does not cover the tree tops). Biomass calculations for
1991 showed, however, that primates in general did not establish
residence at the release site. There was a 40% decrease of total primate
biomass from 1990 to 1991, with biomass returning to almost the same
level found in 1988 (154 kg km"2 in 1988, and 153 kg km"2 in 1991) .
Although the density of some primate species increased at the release
site, the increase included species that have great plasticity, such as
Cebus. The density of less robust species, such as Pithecia, decreased
from 1988 to 1991; therefore, total biomass for the area was maintained
(Fig. 3-3, and 3-5). Projections of total primate biomass in the
reservoir area before its filling, and of biomass increase in the study
sites (Table 3-14), suggest that at least 65% of all primates survived
the initial flooding. This study also demonstrated that many animals
moved through the release site, but it is unknown how far they had to go
to establish themselves, or even if they survived. The loss of habitat,


Percent of trees (%) Percent of trees (%)
23
Height (m)
Height (m)
Figure 2-5: Percentage of trees in each height (m) category
at the Samuel Reserve and Jusante.


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
EFFECTS OF THE SAMUEL HYDROELECTRIC DAM ON MAMMAL AND BIRD
COMMUNITIES IN A HETEROGENEOUS AMAZONIAN LOWLAND FOREST
By
ROSA MARIA LEMOS DE S
December 1995
Chairman: Dr. John F. Eisenberg
Major Department: Forest Resources and Conservation (Wildlife Ecology
and Conservation)
Two sites located near a hydroelectric dam reservoir, in the
northwestern region of Brazil, were monitored in order to document
changes in mammal and bird communities brought about by: a) the creation
of the reservoir, and b) by the release of rescued animals into one of
the study areas. Mammals were sampled with transect surveys, and
understory bird communities with mist nets.
Primate densities and biomass at the Samuel Ecological Station
(located at the southeast border of the Samuel Dam reservoir and
hereafter referred to as reserve) increased after the Samuel Dam flood
gates closed. Primate biomass in the reserve was 154 kg km'2 in 1988,
prior to damming, and increased to 165 ( 65) kg km'2 in 1989,
immediately after flooding, due to the release of animals captured in
the area flooded by the reservoir. However, because of shallow water,
most primates remained in the flooded forest for the first year since
the vegetation in the area was still alive. By 1990, all woody
viii


95
Table 4-6: Matrix showing results of Jaccard's Similarity Index
between zones. J = Jusante, R = Reserve. Numbers in bold identify
the least and the most similar zones.
J-EDGE
J-l KM
J-2 KM
R-EDGE
R-l KM
R-2 KM
J-EDGE
1.000
J-l KM
0.200
1.000
J-2 KM
0.189
0.603
1.000
R-EDGE
0.272
0.338
0.378
1.000
R-l KM
0.209
0.500
0.459
0.456
1.000
R-2 KM
0.247
0.450
0.458
0.371
0.566
1.000
matrix shown on Table 4-6, and the results clearly showed the grouping
of the 1 and 2 km zones at Jusante, and of the 1 and 2 km zones at the
Reserve. The edge zone at the Reserve is an intermediate area between
the four inland zones and the edge zone at Jusante, which is completely
different from all other sites (Fig. 4-7).
To determine similarity of species lists between individual mist
net lines Jaccard's Similarity Index was calculated for all transect
lines (with the exception of area A at Jusante which was sampled for 3
days only, in 1990). A cluster analysis was conducted using Jaccard
similarity values in order to explore fine patterns of affinity between
transect lines. The result, shown in Figure 4-8, grouped zones according
to sites and distances from the water, with only one exception: area C-E
at the Jusantes's edge was grouped with Jusantes's forest interior
zones.