Citation
Niche overlap among Brown Brocket Deer, Pampas Deer, and cattle in the Pantanal of Brazil

Material Information

Title:
Niche overlap among Brown Brocket Deer, Pampas Deer, and cattle in the Pantanal of Brazil
Creator:
Pinder, Laurenz, 1958-
Publication Date:
Language:
English
Physical Description:
xi, 220 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Cattle ( jstor )
Deer ( jstor )
Dry seasons ( jstor )
Food ( jstor )
Forbs ( jstor )
Grasses ( jstor )
Pampas ( jstor )
Rainy seasons ( jstor )
Ranches ( jstor )
Species ( jstor )
Cattle -- Pantanal ( lcsh )
Cattle -- Habitat -- Pantanal ( lcsh )
Deer -- Pantanal ( lcsh )
Deer -- Habitat -- Pantanal ( lcsh )
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF ( lcsh )
Niche (Ecology) ( lcsh )
Wildlife Ecology and Conservation thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1997.
Bibliography:
Includes bibliographical references (leaves 202-219).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Laurenz Pinder.

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:
028637165 ( ALEPH )
38868715 ( OCLC )

Downloads

This item has the following downloads:


Full Text










NICHE OVERLAP AMONG BROWN BROCKET DEER, PAMPAS DEER, AND
CATTLE IN THE PANTANAL OF BRAZIL










BY

LAURENZ PINDER


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


1997




























This dissertation is dedicated to my parents, Angelina Di Giuseppe Pinder and Laurenz Heinrich Julius Pinder; to my wife, Euzelita Almeida Sousa Pinder; and to my grandmother, Olivia, who is very much alive in the heart of all of our family.














ACKNOWLEDGMENTS


I am very grateful to the institutions that have supported this study, and the many people who have encouraged me throughout the different phases of my doctoral study program. This research could not have been possible without the financial support of Conservation International, National Geographic Society, Tropical Conservation and Development Program (University of Florida), Wildlife Conservation Society, and World Wildlife Fund-US. The first 2.5 years of the study at the University of Florida were sponsored by the Conselho Nacional de Ci~ncia e Tecnologia (CNPq) of the Brazilian Ministry for Education.

Authorization to capture and handle pampas deer and brocket deer in the field were provided by the Instituto Brasileiro do Meio Ambiente (Agriculture Ministry). I especially thank Mr. Roberto Klabin, owner of Caiman Ranch, for allowing the research to be conducted on his lands and for providing logistical support during the field work.

In the Pantanal, I owe my gratitude to the people who contributed in one way or another to the data collection. Seu Celestino, a former professional hunter of big cats and hunting guide, provided valuable information and showed me the farthest corners of the ranch on horseback. His son, Cordeiro, and another cowboy kindly helped me in several occasions in the capture of brown


iii








brocket deer. Mauricio, a botany student, helped with the phytosociological and habitat survey data collection. In special, I would like to thank my wife Nininha for her companionship during the fieldwork, and later for helping in the preparation of the microhistological reference collection and in the drawings of the several histologic structures.

The immense difficulties in conducting a microhistologic analysis were overcome through the assistance offered by several key people. Drs. Arnildo Pott and Valle Pott of the Empresa Brasileira de Pesquisa Agropecuaria (Ministry of Agriculture, Brazil) identified the plants collected in the floodplain. Dr. Jane Kraus facilitated the use of the plant anatomy laboratories at the Universidade de Sao Paulo, and contributed strong technical support for the preparation of the microhistological reference collection. Statistical analyses of the large data-base generated in the field were facilitated greatly by the advice of Dr. Sergio Rosso of the Department of Ecology at the Universidade de Sao Paulo, and Jay Harrison at the University of Florida.

I am grateful to my supervisory committee for their advice, understanding, and support. I thank Dr. Ronald F. Labisky (Chair) for his extreme patience in revising and editing earlier versions of this dissertation, and for inspiring discussions. I am grateful to Dr. John F. Eisenberg for his words of encouragement when needed most. I am indebted to Dr. George Tanner for helpful comments on this dissertation. I thank Dr. Richard Bodmer for valuable discussions and comments that improved the quality of this dissertation. Dr. Kent


iv








H. Redford offered great support during the preparation of the proposal for the dissertation and early field work. Finally, I thank Dr. David Webb for stimulating conversation on the evolution of the cervids in South America.

I wish to thank my parents and great friends, Laurenz Heinrich and

Angelina, for their love and support throughout all the difficulties I have faced. I am also deeply grateful to my aunts, Marie Anne and Ingeborg, who always have given me incentive and support when most needed. Nor could I forget my Aunt Marlene, who constantly had a word of advice and comfort. I would like to thank my great friend Sofia, and her many friends, for pointing me in the best direction. Finally, I am extremely grateful to all those people who, although not mentioned, have demonstrated faith in me and in my work.


v















TABLE OF CONTENTS


ACKNOW LEDGMENTS .............................................................................. iii

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

CHAPTERS

1 IN T R O D U C T IO N .................................................................................... 1

B a c k g ro u n d .................................................................................... ..... . 1
General Objectives of the Study......................................................... 5
Organization of the Dissertation ........................................................ 6

2 THE PANTANAL .................................................................................. 7

In tro d u ctio n .................................................................................. . ... .. 7
L o ca tio n ....................................................................................... . ..... 7
Geomorphology.................................................................................... 7
C lim a te ........................................................................................... ... .. 9
H yd ro lo g y ....................................................................................... ..... 9
V e g e ta tio n ...................................................................................... .... 1 0
Plant Phenology ................................................................................. 12
F a u n a .............................................................................................. ... . 1 3
Economy and Conservation ............................................................... 15

3 STUDY AREA ....................................................................................... 19

Rationale............................................................................. ........ 19
Location and Climate ......................................................................... 19
Flood Regime ............................................................. 20
Habitats ....................................................... 22
Land Use ........................................................... 23

4 CLASSIFICATION AND ORDINATION OF THE PLANT ORMATIONS
IN THE FLOODPLAIN................. ........................ 25

Introduction .................................................... 25
Objectives ...................................................... 25


vi









M e th o d s ...................................................................................................... 2 6
Rationale and Sampling Methodology ................................................. 26
Classification and Ordination of Plant Formations.............................. 29
Richness and Diversity of Plant Formations.................... 30
R e s u lts ................................................................................................... . ... 3 0
Cumulative Number of Species............................................................ 30
Species Associations .................................................................... 31
Classification of Plant Formations ................................................ 33
Ordination of Plant Formations ...................................................... 38
Richness and Diversity of Plant Formations ................................. 38
Characteristics of Plant Formations............................................... 41
M a rs h P o n d ............................................................................... 4 1
Moist Basin................................................................................ 44
Short Grass .............................................................................. 44
Tall Grass ................................................................................. 46
S c ru b ................................................................................... .... 4 6
Forest Edge .............................................................................. 46
D isc u s s io n ....................................................................................... ... 4 7

5 CATTLE DIET AND HABITAT SELECTION.......................................... 50
In tro d u c tio n ......................................................................................... 5 0
O bje c tiv e s .......................................................................................... 5 1
Methods ............................................................................................. 51
Fecal Sampling and Analysis ........................................................ 52
Reference Collection .................................................................... 53
Plant Species Availability ............................................................. 54
H a b ita t U s e ................................................................................... 5 4
R e s u lts ............................................................................................ ... . 5 6
Diet Composition .......................................................................... 56
H a b ita t U s e ................................................................................... 5 8
Seasonal Diet Diversity and Niche Breadth .................................. 58
D is c u s s io n ........................................................................................... 6 0

6 BROWN BROCKET DEER DIET AND HABITAT USE .......................... 67

In tro d u c tio n ......................................................................................... 6 7
O bje ctiv e s .......................................................................................... 6 9
Methods ............................................................................................. 69
Fecal Sampling and Direct Observations ..................................... 69
M icrohistologic Analyses................................................................ 70
Phenology of Trees ...................................................................... 72
Habitat Use ................................................................................... 73
R e s u lts ................................................................................................ 7 4
Diet Composition .......................................................................... 74
Availability of Fruits and Flowers .................................................. 79


vii









Habitat Use ................................................................................... 83
Discussion .......................................................................................... 84

7 PAM PAS DEER DIET AND HABITAT USE ........................................... 91

Introduction ......................................................................................... 91
Objectives .......................................................................................... 94
M ethods ............................................................................................. 94
Fecal Sam pling and Analyses ...................................................... 94
Habitat Use ................................................................................... 96
Results ................................................................................................ 96
Diet Com position .......................................................................... 96
Habitat Use .................................................................................... 103
Discussion .......................................................................................... 105
Fecal Analysis vs. Direct O bservations ......................................... 105
Pampas Deer Diet in Relation to the Habitat Characteristics ........ 106

8 ECOLOGICAL SEPARATION AMONG CATTLE, PAMPAS AND
BROW N BRO CKET DEER ................................................................. 114

Introduction ......................................................................................... 114
M e th o d s .............................................................................................. 1 1 7
Density Estim ates........................................................................... 117
Sim ilarity and Niche Overlap Indices.............................................. 117
Plant Species Availability and Selection......................................... 118
R e s u lts ................................................................................................ 1 1 9
Density and Biom ass of Cattle and Cervids .................................. 119
Dietary Sim ilarities.......................................................................... 121
Habitat Use ................................................................................... 124
Niche Overlap and Partitioning ...................................................... 127
Cattle and Deer Responses to Environmental Changes ............... 133
Discussion ........................................................................................... 136
Resource Partitioning .................................................................... 136
Classification of the Cervid Feeding Strategies.............................. 140
Conservation Im plications .............................. ............................. 141

9 SUM MARY AND CO NCLUSIO NS ......................................................... 145

APPENDICES

A LIST OF COMMON PLANT SPECIES IN THE FLOODPLAIN .............. 152

B UNG ULATE DIET CO M PO SITIO N ........................................................ 160


viii









C ANTLER CYCLE AND REPRODUCTION ............................................. 169

Brown Brocket Deer and Pampas Deer Antler Cycles ........................ 169
R e p ro d u ctio n ....................................................................................... 16 9

D H O M E R A N G E S IZ E ............................................................................... 177

Brown Brocket Deer Home Range ...................................................... 177
Pampas Deer Home Range ................................................................ 178

E BONFERRONI TESTS FOR HABITAT USE AND AVAILABILITY.......... 183

F KRUSKAL-WALLIS TESTS FOR DIET COMPOSITION........................ 187

G STP AND SPEARMAN RANK CORRELATION TESTS......................... 191

R E F E R E N C E S ............................................................................................ 2 0 2

BIO G RA PH IC A L S KETC H .......................................................................... 220


ix














Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Philosophy NICHE OVERLAP AMONG BROWN BROCKET DEER, PAMPAS DEER, AND CATTLE IN THE PANTANAL OF BRAZIL By

Laurenz Pinder

December, 1997



Chairman: Dr. Ronald F. Labisky
Major Department: Wildlife Ecology and Conservation


Niche partitioning among brown brocket deer (Mazama gouazoubira gouazoubira), pampas deer (Ozotoceros bezoarticus leucogaster), and cattle was studied in the floodplains of the Pantanal (190 57' S, 56* 25' W), Brazil. The climate was seasonal, with wet, flood, and dry seasons. The vegetation in the study area consisted of six plant formations that were determined through the classification and ordination of random sampling plots: Marsh Ponds, Moist Basins, Short Grass, Tall Grass, Scrub, and Forest Edge.

Pampas deer used all plant formations except forest patches, but used Short Grass most frequently. Moist Basin was an important source of food for pampas deer during the dry season, when 44% of their diet consisted of a forb (Melochia simplex), which grows in this habitat. The progressive reduction of

x









dietary diversity and niche breadth from rainy to dry seasons suggested that availability of new-growth was important for pampas deer. Graminoids, forbs, and browse dominated their diet during the rainy season; forbs and browse were prominent during the flood season; and forbs alone constituted 79% of the diet in the dry season.

Brown brocket deer preferred Scrub and Forest Edge habitats, and were affected little by seasonal changes in food availability. Brown brocket deer were mostly browsers throughout the year and consumed the greatest proportion of forbs (55%) in the dry season, when they foraged more frequently in Moist Basin and Short Grass, as compared to other seasons. Brown brocket deer managed to maintain dietary diversity and niche breadth throughout the year by increasing their use of different types of habitat when food was limited in the selected habitat. Neither fruits and flowers of trees contributed significantly to the diet of brown brocket deer although they were consumed in the rainy season and in the dry season respectively.

Eighty percent of the cattle's diet consisted of graminoids, reducing

probabilities of competition between cattle and cervids. Niche partitioning among the three ungulates was achieved principally by the selection of different plant species and by the type of plant formation selected.


xi














CHAPTER 1
INTRODUCTION


Background

Wet areas are important to global biodiversity preservation besides providing substantial resources of food and water for people (UNEP, 1991/1992). The Pantanal, the world's largest wetland, constitutes a substantial component of South American biodiversity which has floristic and faunal elements from the Amazonian, Cerrado, and Chaco biogeographical regions (Adamoli, 1984; Pott, 1988). The region has been relatively free from effects caused by economic activities, such as skin trade, animal traffic, gold mining, deforestation, and cattle production until recent years. Currently, the decline of large ranches, and a future waterway conversion of the Paraguay River, are threatening the ecological balance in the region (Gomes, 1997). As a result, several faunal species are in danger of a future extinction in the wild (Alho et al., 1988).

Additionally, there is social pressure for land utilization and for

optimization of current agricultural methods; therefore, large mammals, as a group, will become difficult to conserve. Collectively, the vastness of the region, low density of people, and large ranches (hundreds of square kilometers each) ensured the survival of the entire fauna in natural condition until thel 970s.


1







2


However, South American large mammals occur in low numbers naturally (Eisenberg, 1989; Redford & Eisenberg, 1992), which potentially exposes them to greater chances of local extinction when subjected to environmental changes or excessive hunting pressure (Redford, 1997). Therefore, conservation of local fauna will depend either on large reserves containing several habitat types, which do not exist currently, or on adequate land management practices that are compatible with the local ecological equilibrium. However, the creation of large, state-protected areas currently is politically difficult because of the lack of support by landowners and because thousands of landless people are pressing the government for appropriation of "sterile" areas for agricultural use.

The second alternative, the integration between land utilization and

conservation, is a potentially valid option. Except for animal traders and few caiman breeders, the pecuniary value of the fauna perceived by most of the locals is low. And, few cattle ranchers have realized the potential income associated with ecotourism. However, there is no substantial information on the specific requirements of native large mammals. The effects that cattle ranching, and its modern practices, impose on wildlife are poorly understood. Because ecotourism represents only a small source of additional income, cattle ranchers may abandon interest in participating in the ecotourism industry if local fauna begins to decrease. The long-term goal of managing wild fauna and their habitats will depend ultimately on scientific studies of resource use among the






3


cattle and native large herbivores although the consortium between cattle and wildlife is currently feasible.

South American's large herbivore guild is poor in contrast to Africa and Asia. This guild is usually represented by cervids, a rodent (capybara: (Hydrochaeris hydrochaeris), and a tapir ( Tapirus terrestris). Common species inhabiting the Pantanal floodplain are marsh deer (Blastocerus dichotomus), pampas deer (Ozotoceros bezoarticus), brown brocket deer (Mazama gouazoubira), and capybara. The tapir and red brocket deer (Mazama americana) are rare. These two latter species are more commonly found in the semi-deciduous forests bordering the floodplain. Currently, only a few ecological studies of the cervids have been conducted. Capybaras have been studied more extensively due to their economical importance in some areas (Escobar & Jimenez, 1976; Jorgenson, 1986; Alho et al. 1987a,b; Herrera & MacDonald, 1989).

Published studies on cervids include aspects of feeding habits and habitat use of pampas deer by Bianchini and Perez (1972), Jackson and Giulietti (1988), Heinonen et al, (1989), and Merino (1993) in Argentina; by Jackson and Langguth (1987) in Uruguay; and by Resende and Leeuwenberg (1992) in the Brazilian Cerrado. And, despite a larger geographic distribution, brown brocket deer have been studied in only a few localities. Stallings (1984, 1986) studied feeding habits and reproduction of the brown brocket deer in the Paraguayan Chaco, and Bodmer (1989) and Resende and Leuwenberg (1992) investigated







4


habitat use and diet in the Peruvian Amazon forest and in the Brazilian Cerrado, respectively. Habitat use and diet of the marsh deer, the least studied of these three species, has been researched by Schaller (1976), Beccaceci (1994), Beccaceci and Merino (1994) in Argentina, and by Voss et al. (1981), Mauro (1993), and Pinder (1994, 1996) in Brazil. No investigation has been conducted on niche partitioning among the three cervids, or on competition between cervids and livestock in the Pantanal.

No competition between capybaras and cattle, or other native ungulates, has been documented in the South American savannas although this rodent constitutes a large biomass in the Pantanal (Schaller, 1983; Lourival & Fonseca, 1997). Capybaras are semi-aquatic grazers; thus, most of their forage is found in or near water bodies (Escobar & Gonzalez Jimenez, 1976; De Azcerate, 1981; MacDonald, 1981; Schaller & Crawshaw, 1981; Jorgenson, 1986; Herrera & MacDonald, 1989, Quintana et al., 1994). Cattle, as well as pampas and brown brocket deer in contrast to capybaras, avoid areas permanently flooded (Bodmer, 1989; Leeuwenberg & Resende, 1994). Furthermore, Pott et al. (1986) have observed that in the Brazilian Pantanal, the capybaras consumed medium to high proportions of sedges (Cyperaceae) instead of grasses, which are preferred by cattle. Similarly, marsh deer would have little niche overlap with cattle or other cervids due to their preference for permanently flooded areas (Schaller, 1976; Mauro, 1993; Beccaceci, 1994; Pinder, 1994).






5


This dissertation presents the results from research conducted in the Pantanal of Brazil between January 1990 and December 1992. The research consisted of two phases: a classification of the habitats existing in the seasonally flooded plains in the study area; and a documentation of habitat preferences and diet of pampas deer, brown brocket deer, and cattle. The first phase was important as it provided data of habitat and plant species availability that were fundamental for the second phase.

General Obiectives of the Study

This study sought to answer questions that related indirectly to the

consequences of introducing livestock into the habitat of two species of deer. My major goal was to obtain the base-line information needed to conserve deer populations on the private lands of the Pantanal although South American deer have been little studied, and, hence, are of considerable theoretical scientific interest. A secondary goal was to explore theoretical questions related to the classification of feeding habits of South American deer (Bodmer, 1989). Pampas deer occupying the Argentinean grasslands have been classified as grazers (Jackson & Giuletti, 1988), whereas brocket deer occupying tropical forests and the Paraguayan Chaco have been classified as frugivores (Bodmer, 1989). The Pantanal, with its mosaic of grasslands, scrub, and forest patches, provided an ecosystem to test these previous classifications.

To address these goals the following questions were asked:






6


1. Which plant formations were available to the three large herbivores

(brown brocket deer, pampas deer, and cattle) in the seasonal floodplain of the Pantanal?

2. What constituted the diet of the brown brocket deer, cattle, and pampas deer during the three major phenological seasons in the region: rainy season, flood season, and dry season?

3. Was there significant dietary and habitat overlap among the three herbivores?

4. Did the dietary classification of grazer and frugivore hold for pampas and brocket deer in a habitat where there were multiple habitats available?

5. How did the answers to these questions contribute to the future management and conservation of these wild ungulates in the Pantanal?

Organization of the Dissertation

Chapter 2 describes the climate, hydrography, vegetation, fauna, and economy of the Pantanal. Chapter 3 characterizes the study site. Chapter 4 presents the results of classification and ordination of the different plant formations observed in the seasonal floodplain. Chapter 5 analyzes the feeding habits of cattle, foraging value of each of the major plant species consumed, and foraging value of the habitats available to cattle. Chapters 6 and 7, respectively, describe brown brocket deer and pampas deer habitat preferences and their diets, and discusses their dietary classification. Chapter 8 discusses the niche overlap among the three ungulates.














CHAPTER 2
THE PANTANAL


Introduction

Pantanal means a large continuous swamp in Portuguese. It is considered one of the largest interior wetlands in the world (Silvestre Filho & Romeu, 1974; IBGE, 1977; Rizzini, 1979). However, most of the area is not permanently inundated, but rather is flooded only seasonally. The elevation of the Pantanal ranges from 83 to 165 m above mean sea level. In contrast, the surrounding plateaus of the Brazilian Shield exceed 600 m elevation (Silvestre Filho & Romeu, 1974; Rizzini, 1979). The landscape of the Pantanal remains little modified by human activities although completely occupied by ranches (Por, 1995).

Location

The Pantanal is located in southwestern South America, between parallels 160 and 220 S, and 550 and 580 W (Fig. 2-1). It extends approximately 770 km North-South and 250 km East-West within Brazilian borders, occupying an area of 160,000 km2 of seasonally flooded plains.

Geomorphology

The Pantanal floodplain consists of Pleistocene alluvial sediments: sand and silt. One of these two sediment types dominate the floodplain, depending on region.


7










8


6O"








15'




SOUTH _16

AMERICA






BO VIA PACIFIC ATLANTIC

OCEAN OCEAN



20'







*PAF AGUAY 22. - -Nr !


SW 57* 5W


5F*


EM8ALAA(E (0umBA













MATO GROSSO






9' 8 .7 5 5y 5


Figure 2-1. Location of the Pantanal showing main rivers (thin lines) and limits of the floodplain (thick line).


CAANDE


DO SUL


54 53'


MAT GROSS


- 15* 17















-2D




-21.


4" 53'






9


In general, sandy soils are poor in nutrients and silty soils are fertile. There is only local redistribution (vertical and horizontal) of nutrients in areas not influenced by alluvial nutrients. Most of the nutrients of the Pantanal's forest patches are in the arboreal biomass rather than in the soil, similarly to what happens in the Amazon forest. There is some degree of nutrient loss in areas of open drainage, which is accentuated by the annual burning promoted by ranchers at the beginning of the rainy season (Pott, 1988).

Climate

The climate is tropical hot and humid with 3 months of dry season (Jun.Aug.), and 3 months of rainy season (Nov.-Jan.). Annual rainfall ranges from 1000 to 1500 mm, with a mean annual temperature of 240 C. More than 80% of the annual rainfall occurs between November and March, which causes flooding between January and April. Absolute temperatures range from 0* during the winter to 420 C in the spring (IBGE, 1977, Silvestre Filho & Romeu, 1974).

Hydroloqv

The entire region is drained through the Paraguay River, which, with the

Paran6 River, constitutes the second largest riverine basin in South America after the Amazon. The drainage basin of the Pantanal is fan-shaped with the center in the Paraguay River on the Bolivian border, and the wings toward the east (Fig. 2-1). There are permanent and temporary streams between the main tributaries, locally called "corixos" and "vazantes", that link a network of lakes or "baias". River waters






10


flow slowly and form countless ponds in some areas because there is little relief in the region (2-5 cm/km N-S, and 30-50 cm/km E-W) (Brasil, 1974; Pott, 1988).

Vegetation

The flora of the Pantanal, which contains approximately 1500 plant species, has not been studied extensively. Due to the large areal size of the Pantanal, its flora is influenced by different, adjacent biomes (Pott, 1988). The gallery forests of the Pantanal are influenced by the Amazonian tropical rainforest in the north (Eiten, 1985). The Chacoan vegetation dominates the landscape in the south. The largest influence however, consists of species characteristic of the Brazilian savanna, the Cerrado (SEMATEC, 1990), which originally covered the interior of the country below the Amazon forest (Pott, 1988).

Many of the plant species in the Pantanal are of general geographic

distribution. Some are present in other South American floodplains, such as the Colombian Llanos, Amazonian Maraj6 Island, Bananal Island, and the Paraguayan Humedales. Others are common to the dry forests of northeastern Brazil and Argentina (Pott, 1988).

General descriptions of the Pantanal's vegetation types have been provided by Veloso (1948), Prance & Schaller (1982), Eiten (1985), and Adamoli (1986). Prance and Schaller (1982) provided information on phytosociology and phenological patterns for forest habitats, but similar studies on grassland habitats are lacking.






11


Small differences in relief (1-4 m) influence plant associations due to seasonal flooding. Areas above the flood line are covered by different forest vegetation types: cerrado (scrub savanna), cerrad~o (scrub forest), and semideciduous forest (Eiten, 1972). The arboreal cover of a given area is dependent on the micro-climate and the type of soil. For example, Cerrado vegetation is found on sandy soils, whereas Chaco vegetation is common on clay or alkaline soils (Pott, 1988). Finally, there is the xeric vegetation type (steppe savanna), which is characteristic of slopes of dry and calcareous soils.

According to Veloso (1948), seasonally flooded areas can be divided into three categories: the Aquatic Zone, the Hygrophile Zone, and the Mesophile Zone. The Aquatic Zone is characterized by floating and rooted aquatic species (e.g., Eichhomia crassipes, Salvia spp., and Pontederia spp.). The Hygrophile Zone is sub-divided into two categories: 1) Vazantes or shallow drainage channels, characterized by aquatic plants, sedges, grasses, and herbs growing on seasonally flooded ground, which do not completely dry during the winter (dry season); and 2) periodically flooded terrains bordering the rivers, which are dominated by woody plants and trees. The Mesophile Zone occupies higher terrain, and is characterized by species adapted to aquatic and terrestrial environments. Woody plants occupy islands of non-flooded soils, grasses and sedges occupy soils subject to moderate flooding, and aquatic and semi-aquatic plants occupy depressions that remain moist year-round. In areas where the flooding occurs for only a few days during the peak






12


of the flood season, a scrub savanna dominates the plains (Prance and Schaller, 1982, Ratter et al., 1989).

The Mesophile Zone constitutes the native pasture of the Pantanal. The landscape is punctuated by circular forest patches or "cap6es", and narrow and parallel strips of forest or "cordilheiras" never reached by water. Invading woody shrubs may become abundant on the grassland during dry years and in areas overgrazed by cattle (Pott, 1988). On the other hand, the aerial parts of grasses may die and be replaced by aquatic and semi-aquatic plants when the flooding is prolonged for several months.

Water levels affect the abundance and distribution of forest patches. Thus, "cordilheiras" and "cap6es" are more abundant on higher elevations close to the borders of the Pantanal (Brasil, 1982). In general, "cap6es" are less than 20 m wide, and "cordilheiras" are 50-150 m wide and 2 km in length. These two types of forest patches stand just above the water line during the flood season, and are important refuges for wildlife and cattle. Trees have superficial roots because of the shallow water table (0.2-2.5 m), and are characteristic of the forests bordering the floodplain (Pott, 1988).

Plant Phenology

Due to the marked seasonality of rainfall, many of the arboreal species lose their leaves during the dry season (June-August). The peak of defoliation occurs in September, before the onset of the rainy season. This pattern is more accentuated in the semi-deciduous forests of the slopes bordering the floodplain (Schaller, 1983).






13


Unlike leaf-fall, the fruit production and flowering of trees are not synchronized. Schaller (1983) registered peak flowering of Caesalpinia sp. in March, Pseudobombax sp. in May-June, Bowdichia sp. and Pouteria sp. in July-August, Tabebuia caraiba in August, Magonia sp. and Simarouba versicolor in September, and Acosmium sp. in October. Curatella americana and Tabebuia impetiginosa may have more than one flowering peak between May and September, an event probably related to variations in the water table. Schaller (1983) also reported that fruits were available in any month. Herbs and vines bloom from April to June, after the flood waters recede (Schaller, 1983).

Seeds and fruits are more abundant in semi-deciduous forest than in the

floodplain. Schaller (1983) found that 52% of trees bore fruits at some time between March and October in the semi-deciduous forest of the slopes, whereas only 20% of trees in the floodplain bore fruits during the same period.

Fauna

The influence of different biomes and a great variety of habitats allow for a

large faunal diversity, especially waterbirds, within the Pantanal. Psittacidae are very diverse, including five species of macaws and 13 species of parrots (Alho et al., 1988). The Pantanal lacks large raptors, with the two largest species being the great black hawk (Buteogallus urubitinga) and the great homed owl (Bubo virginianus). The king vulture (Sarcoramphus papa) and three other species of vultures form the feathered scavengers guild of the Pantanal. Besides the local






14


avifauna, the Pantanal hosts migrating birds of three major migratory routes: Central Brazil, Rio Negro, and Cis-Andean (Sick, 1983).

The largest reptilian predators include the Pantanal caiman (Caiman crocodilus yacare), which may weigh 110 Kg, and the anaconda (Eunectes murinus), the world's largest snake (Almeida, 1976). Other snakes are rare in the floodplain presumably because the combined effects of flooding and grass fires during the dry season maintain the area voided of their prey.

Many large mammals still abound, including a few vulnerable species. A survey of mammals in the Pantanal yielded 64 species of mammals, one third of which were bats (Schaller, 1983). Other families included the Marsupialia (three species), Primates (five species), Edentata (six species), Lagomorpha (one species), Rodentia (eleven species), Camivora (ten species), Perissodactyla (one species), and Artiodactyla (six species). The total biomass of these species was 380 kg/km2. Ungulates formed the largest biomass (62%) of mammals in the Pantanal. Rodents ranked second in biomass (12%), principally due to the large number of capybaras (Hydrochaeris hydrochaeris) (Schaller, 1983). Livestock had a biomass of 3,750 kg/kM2, a value 10-fold greater than native mammals.

The Pantanal includes most of the Neotropical genera of hoofed mammals, including tapir ( Tapirus terrestris), marsh deer (Blastocerus dichotomus), pampas deer (Ozotceros bezoarticus), red brocket deer (Mazama americana), brown brocket deer (Mazama gouazoubira), peccaries ( Tayassu pecan and T tajacu), and feral






15


hogs (Sus scrofa), the latter introduced with the first European colonizers. Capybaras complete the guild of large herbivores.

Jaguars (Panthera onca) and pumas (Puma concolor) are still present along with other smaller predators such as the crab-eating fox (Cerdocyon thous), ocelot (Felis pardalis), and the rare maned wolf (Chrysocyon brachyurus), even though all of them are regularly hunted (Lourival & Fonseca, 1997).

Economy and Conservation

The Pantanal now supports a population of more than 340,000 people,

including natives and descendants of colonizers. Cattle ranching has been the main economic activity in the region in the 200 years post-settlement. There are more than 12 million head of cattle within 13 million ha of native pastures and 550,000 ha of introduced pastures (IBGE, 1977, Alho et al., 1988). Among the several native species of grasses that occur in the Pantanal, the preferred forages are Paratheria prostata, Setaria geniculata, and Reimarochloa brasiliensis.

Historically, ranchers moved their herds to accommodate the rise and decline of the water on the native pastures. Because of the near lack of fences, cattle ranged free within ranches, often larger than 2,000 km2. Subsequently, as the ranches were divided among descendants, once-extensive lands became smaller and smaller. Today, there are 2,000 ranches, with an average size of 13,000 ha each (Allen & Valls, 1987).

Currently, ranchers have substituted the traditional cattle variety, which was adapted to flood regimes, for zebu cattle that require more productive pastures (Por,






16


1995). Consequently, forest patches within the Pantanal region began to be replaced by exotic, and presumably more productive, grass varieties.

The introduction of new livestock to the region created yet another problem. Although cattle are regularly vaccinated, several domestic livestock diseases were initially introduced to the fauna of the Pantanal. Equine trypanosomiasis (Trypanosoma spp.) infected capybaras and peccaries. Brucellosis and foot-andmouth diseases infected wild ungulates. Periodical outbreaks of these diseases reduced wild faunal populations, and may have been responsible for the decline of marsh deer (Schaller, 1983; Schaller & Vasconcelos, 1978).

Furthermore, rudimentary land management practices have caused the impoverishment of the soil and the invasion of weed species. Ranchers annually bum the native pastures before the rainy season to reduce invader plant species. However, this continuous process increases the dominance of less palatable grass species, such as Elyonurus muticus (IBGE, 1977). Furthermore, the annual frequency and the timing of the burning may have impacts on soil nutrient cycling, vegetation, and fauna although the ecosystem is adapted to fire (IBGE, 1977; Coutinho, 1990).

Since the 1980s, cattle production has increased due to the deforestation of cordilheiras and forests on the slopes surrounding the floodplain. However, this deforestation has caused severe erosion, and, consequently, the silting of rivers (Alho et al., 1988).






17


Agriculture enterprises beyond cattle production are hindered by the

environmental conditions, and are practiced only at subsistence level (Allen & Valls, 1987; Por, 1995). The conversion of land to crop production has not been significant in terms of changing the landscape because the human population in the region is still low.

Another economic activity in the Pantanal, which has environmental

implications, is illegal professional hunting and the associated smuggling of animal products. In the early 20th century, skins and other products were harvested legally and shipped from Brazil to the markets in Uruguay, from which they were exported to Europe (Miller, 1930). In 1967, the Brazilian government passed a decree outlawing professional and sport hunting; however, the lack of law enforcement allowed for the continuous hunting. In 1970 only, 15,511 skins, 3,140 furs, 2 tons of rhea feathers, and 1,903 tons of fish were taken illegally in the Pantanal. These numbers have increased steadily ever since (Alho et al., 1988). Most of the animals or their products are smuggled by land or small airplanes across the borders of Bolivia, Paraguay, and Argentina. Forged documents are then prepared in these countries, and the products are exported to Europe and Japan.

Predator control and subsistence hunting may also contribute to the local decrease of wildlife numbers. Many of the traditional ranchers keep dogs for the hunting of large cats, and other predators are killed when casually encountered by cowboys (Lourival & Fonseca, 1997). Additionally, capybara, armadillos, and all






18


ungulates are hunted for food by ranch personnel year-round (Schaller, 1983; Lourival & Fonseca, 1997).

The tourism industry was very incipient economically until the mid-1 970s; however, tourism has gained popularity as a source of income to ranchers (Por, 1995). The few tourists that visited the region prior to the 1970s were in most of the cases wealthy trophy hunters or sport-fishermen (Almeida, 1976). Tourism has boomed in large cities around the Pantanal since the mid-1 980s, and, accordingly, a few ranchers have tried to capitalize on tourists by building facilities to accommodate them (Por, 1995). Some of these tourism operations, however, are conducted by poorly informed tour groups, and have caused disturbance to waterbird rookeries and nesting areas, in addition to exacerbating widespread littering (Alho et al., 1988).














CHAPTER 3
STUDY AREA


Rationale

Caiman Ranch was chosen as the study site in the Pantanal for the following reasons:

1. The site contained natural densities of brown brocket deer (Mazama gouazoubira) and pampas deer (Ozotoceros bezoarticus).

2. The brown brocket deer, pampas deer, and cattle occupy the same range of habitats within the floodplain.

3. The entire area has been protected from hunting for over 30 years, which has allowed cervids to become habituated to the presence of humans.

4. The original assemblage of fauna on the Caiman Ranch is still extant.

Location and Climate

The study area is a cattle ranch located between the Aquidauna and the

Miranda rivers in the southeastern Pantanal. It lies 207 km west of the State capital, Campo Grande, and 36 km north of Miranda, the nearest town (19057'S, 56025'W) (Figure 2-1). The climate at Caiman Ranch is seasonal and characterized by a wet and dry season, which is ranked between Aw and Am using Koepen's classification (Sudo, 1974) (Fig. 3-1). Mean monthly temperatures are relatively constant


19






20


throughout the year. In 1991, the annual mean monthly temperature was 240 C, with annual mean maximum and minimum temperatures of 300 C and 170 C. The monthly mean temperature was lowest in August (190 C), and highest in February (260 C) (Fig. 3-2). The mean annual rainfall for the 20 years, 1972-1991, was 1,7733.40 mm, but exhibited a declining trend (Fig. 3-3). Crawshaw and Quigley (1984) reported a dry period between 1937 and 1950, when annual rainfall averaged 1,213 mm. In recent years, the highest rainfall at the ranch occurred in 1974 (2,298.70 mm), and the lowest in 1988 (1,257.40 mm). However, the weather is unpredictable from year to year with regard to amount of monthly rainfall (Fig. 3-3).

Forty-nine percent of the annual rainfall occurs during the 3-month rainy

season (Nov.-Jan.), with January being the wettest month (x =320.4 mm). The dry season occurs from June to August, with July being the driest month (x =26.7 mm). A secondary peak of rainfall occurs in May, when sudden lower temperatures associated with the first polar air masses of the winter season reach the Pantanal (Fig. 3-1).

Flood Regime

Rainfall has a significant effect on the water level of lakes and streams., the lake in front of the ranch headquarters showed a variation of 0.74 m over the year in 1991, reaching its highest level in April, and its lowest in August. Crawshaw and Quigley (1984) reported a difference of 1.66 m between the highest and lowest levels in 1983.







21


-j
-j
U
z

w

w


E


350 300 250
200 150 100 50
0


50
40 w
30 )
F
20 W
w
0
w 0i


JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

-1- RAINFALL -TEMPERATURE




Figure 3-1. Annual pattern of rainfall (1972-1991) and temperature (1991-1992), Pantanal Study Area (Caiman Ranch). Both rainfall and temperature statistics were collected on-site.









& 50 - -- - - --


w
40 <30


w 0 z 0
w


JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC


---Min. A -Max



Figure 3-2. Annual mean monthly minimum and maximum temperatures (0C), Pantanal Study Area (Caiman Ranch), 1991.






22


2500
E 2000
1500 1000
Q 500
z z
< 0
72 74 76 78 80 82 84 86 88 90


Figure 3-3. Annual rainfall (mm), Pantanal Study Area (Caiman Ranch), 19721991.


The effects of heavy rainfall can be observed in the water level of streams within 3 days after the rain. However, the flood season begins only weeks after the onset of heavy rains (Nov.) because of the cumulative effects of the raising watertable and the river overflows. During 1991 and 1992, the flood season began in late February and the lowlands remained covered by 50 cm of water until the end of April. However, it is possible that flood conditions may extend from January to June in extremely wet years.

Habitats

The floodplain, which occupies 98% of the ranch, is bounded on the north by the Aquidauana River. The floodplain provides the principal native habitat for pampas deer and brown brocket deer, and pasture for cattle. The hills bordering the floodplain, which comprise 2% of the area, are located on the southern limits of the ranch.






23


A vegetation map, produced at the request of the ranch owner by a private company using satellite imagery, revealed 7 habitat types. Semi-deciduous forest (1,071.40 ha), dense scrub forest (4,055.78 ha), and introduced pastures (834.73 ha) are located in the borders of the floodplain. Scrub savanna (7,964.60 ha), grasslands (30,208.01 ha), shallow drainage channels (4,735.27 ha), and lakes (634.51 ha) lie in the seasonally flooded plains. The remaining area is occupied by roads and facilities.

However, the satellite imagery (definition 30 m) classified only the major

vegetation patterns. Thus, it was necessary to conduct a phytosociological study to discriminate differences among the mosaic of small plant formations that constitute the habitat of the species studied (Chapter 4).

Land Use

Development in the region is relative recent. The access to the town of Miranda until the early 1970s was mostly performed by train, boat, or plane (Silvestre Filho & Romeu, 1974). Most of the ranches, however, have their own airports because many of the owners do not live on their ranches, but rather commute to them periodically (Por, 1995).

The main economic activity at the Pantanal Study Area (Caiman Ranch)

always has been cattle reproduction. Additionally, in 1989, a tourist lodge was built on the old ranch headquarters to accommodate tourists. New facilities were constructed in 1993, which has approximately doubled the local lodge capacity.






24


The awareness and concern with the conservation of the fauna also have

evolved on the ranch. Several sportsmen, including international celebrities, hunted jaguars and other game in the ranch until the mid 1960s, when the owners decided to protect the fauna (Almeida, 1976). Recently, the Caiman Ranch's owner has stimulated scientific studies in the area to better protect and manage its natural resources. As a result, research on rare species, such as jaguar (Panthera onca palustris), hyacinth macaw (Anodorhynchus hyacinthinus), and jabiru stork (Jabiru mycteria), has been conducted on the ranch (Crawshaw and Quigley, 1984; Banks, 1991).














CHAPTER 4
CLASSIFICATION AND ORDINATION OF THE PLANT FORMATIONS IN THE FLOODPLAIN

Introduction

The analysis and understanding of the habitat use by any faunal species depends on a knowledge of the environment where the animal lives. Frequently, wildlife biologists rely on previous descriptions of the vegetation types. However, in the present study, I had neither available references on the habitat selection and use by ungulates nor detailed reports of the vegetation types on the Pantanal Study Area.

The local vegetation, particularly within grassland and scrub, has distinct physiognomies depending on the height of the grasses and abundance of shrubs and small trees. All of these plant formations were previously assembled by botanists and other researchers in only one or two phytophysiognomic categories (Hoehne, 1923, 1936; Veloso, 1948; Brasil, 1982; Eiten, 1985; Allen & Valls, 1987). These gross groupings were of little use for the level of definition necessary to understand habitat selection by the three ungulates.

Objectives

The first objective of this study was to classify and ordinate the different

plant formations, based on numerical criteria obtained through a phytosociological study. The seasonal availability of plant species could also be determined as a sub-


25






26


product of the phytosociological study. The following hypothesis was formulated to achieve the main objective:

H.1. Visually identified floristic formations constitute numerically distinct plant associations in the floodplain.

Methods

Rationale and Sampling Methodology

The study was conducted in three steps to describe the vegetation types of the floodplain: 1) subjective identification of the vegetation types most used by brown brocket deer (Mazama gouazoubira) and pampas deer (Ozotoceros bezoarticus), based on structure and species dominance; 2) sampling within homogeneous vegetation types described in Step 1, with determination of species composition and cover; and 3) treatment of the phytosociological data collected in Step 2 to verify whether or not the vegetation types described in Step 1 represented significantly different plant formations.

Preliminary observations on pampas deer and brown brocket deer were

conducted from August (dry season) of 1989 to November (rainy season) of 1990 to accomplish Step 1. Ten distinct vegetation types were classified, based on the vegetation structure and species dominance: 1) Marsh Ponds, 2) Moist Basins, 3) Short Grass, 4) Medium Grass (Elyonurus muticus (Spr.) Kunth), 5) Medium Grass/Tall Grass (Elyonurus muticus /Schizachyrium microstachyum Desvaux), 6) Short Grass/Tall Grass, 7) Tall Grass, 8) Scrub, and 9) Forest Edge ("capao" or tree






27


island), and 10) Mixed Association (a tall grass formation that had been plowed and seeded with Brachiaria humidicola (Rendle) six months before the sampling effort).

Representative tracts of each of the 10 plant formations were selected

arbitrarily for sampling with the point-quadrat method in November of 1991 (Levy, 1927). Use of this method allowed the estimation of the cover of grasses, forbs, aquatic plants, and shrubs in each plant formation. Mantovani and Martins (1990) summarized and discussed the limitations of the method, which has been recommended by several plant ecology researchers (Crocker & Tiver, 1948; Du Reitz, 1930 in Levy & Madden, 1933; Whitmann & Siggeirson, 1954; Thomas, 1960).

Twenty sampling points were taken from each of 25 1 -m2 quadrats, which

were distributed within five 25-M2 quadrats. The 25-M2 sampling units containing five

1 -M2 quadrats were placed randomly along a transect within each plant formation, except Forest Edge. Only 1-rM2 quadrats were established along the interface between the forest and the grassland of different forest patches selected at random due to their circular shape and small size. The corners of quadrats were marked permanently with wooden poles for repeated vegetation samplings in January (peak of rainy season), March (flood season), September (dry season), and November (beginning of the rainy season). A graph of the cumulative number of species versus number of plots sampled was made to determine the total number of sampling units required for seasonal vegetation analysis. As a result, a total of 30 1 -M2 quadrats






28


were sampled for Forest Edge and 50 1-rM2 quadrats for Marsh Pond, Marsh Basin, and Mixed Association.

The sampling points were determined with the aid of a metal rod, 20 cm in length x 7 mm diameter, suspended by a nylon string from a 1-iM2 grate of 1 0-cm 2 mesh. This method allowed for unbiased sampling of the vegetation below the frame because gravity determined the exact sampling point. The string was released from the grate until it touched the vegetation, water, or soil. The sampled plants were identified immediately to the best knowledge of the observer (family, genus, or specific epithet).

Specimens of all species sampled were collected in triplicate. One collection was sent for identification to the Brazilian Center for Agricultural Research of the Pantanal (CPAP/EMBRAPA); another was kept for comparisons in the field; and a third was used for developing a microhistological reference collection, which was prepared at the laboratory of plant anatomy in the Botany Department of The University of S~o Paulo (USP).

A board with horizontal strips of alternate colors at 1 0-cm intervals was used to measure the height of vegetation in each of the plant formations described. The board was placed at the center of each sample plot and the vegetation height was visually determined by an observer from a 20-m distance. The same procedure was employed to measure the water depth during the flood.






29


Classification and Ordination of Plant Formations

The process of classification was performed in 2 steps. First, the initial matrix was transformed to express dissimilarities among all sample pairs (Williams et al., 1966; Williams et al., 1973). In the second step, the dissimilarities, expressed in terms of geometric distances, were fused into clusters (Goldsmith et al., 1986). To classify the initial matrix of stands-by-species into clusters of closely related units, I used the computer program PHYTOPAK developed at UNICAMP University, Sao Paulo, Brazil.

Initially, a qualitative Q-classification analysis was performed by calculating the importance of each plant species within each of the plant formations (Index of Specific Value modified from McCloskey's IBV, 1970). The one-tailed significance limits (a = 0.05) for Jaccard Similarity between formations (0.25 and 0.35, corresponding to 0.05 and 0.95 percentiles of the frequency distribution of similarities obtained by chance) were established after 40 permutations of all species occurrences. To examine affinities between species, a qualitative R-Classification analysis also was performed (UPGMA, Jaccard Similarity). To minimize noise, species with ISV lower than 0.2 were excluded from further analyses.

To explore and test quantitative relations between the sampled

formations, Q-Classification (Bray Curtis Distance) and Correspondence Analyses (Ter Braak, 1989) were performed using the complete data set. Correspondence Analyses was executed subsequently with reduced data. The






30


significance limits (0.57 and 0.76) for distance values were established as explained above.

Richness and Diversity of Plant Formations

Two procedures were used to measure richness of each of the plant

formations discriminated by the Quantitative Q-Classification technique. In one procedure, the index of richness was measured by dividing the total number of species of a given plant formation by the area sampled. All plant formations had equal sample size (1,000 sampling points) except Forest Edge, which had a smaller sample (600 points). I employed the rarefraction method (Hurlbert, 1971), which estimates a richness index regardless of sample size, as a comparative method.

I employed Shannon's diversity index to obtain a measure of diversity within each plant formation (Shannon & Weaner, 1949, in Ludwig & Reynolds, 1988). Shannon's index (H') has two properties: 1) H' = 0, when there is only one species in the formation; 2) H' is maximum when the number of species equals the number of specimens sampled. Here, I refer to diversity as an index of cover diversity instead of species diversity because large plants (e.g., grass tussocks) can be recorded more than once per sampling point.

Results

Cumulative Number of Species

In general, the grassland formations had a low number of species, especially those dominated by carpet grass Axonopus purpusii or wire grass Elyonurus muticus. The number of plant species was greater in moist shallow patches of






31


vegetation, where hydrophytic forbs and shrubs grow, and in the least flooded soils of the Scrub and Forest Edge (Table 4-1). The smaller number of species in the grasslands is probably a consequence of the more drastic hydrological alterations suffered by plants on grassland soils. Only few resilient species can survive total submersion and subsequent drought.

Species Associations

A total of 155 species of 56 families of plants were identified during the four months of sampling. The majority (103) of the species were associated with savanna, whereas fewer (52) were associated with forest. The most represented families were Poaceae, Cyperaceae, and Fabaceae with 22, 12, and 11 species, respectively. Ten plant species were unidentified.

The following species were found consistently co-occurring (Jaccard Similarity Index [JSI] > 0.7):

a) Eleochais elegans (H.B.K.) Roem. & Schi. (Cyperaceae), Aeschynomene fluminensis Vell. (Fabaceae), and Hydrolea spinosa L. (Hydrophyllaceae) which were common in Marsh Pond formation; b) Nymphoides indica (L.) 0. Kuntze (Menyanthoideae) and Ludwigia inclinata (L.F.) Gomez (Onagraceae) which were common in Marsh Pond formation; c) Richardia grandflora (Cham. & Schl.) Steud. (Rubiaceae) and Schizachynium microstachyum (Poaceae) which were common in Tall Grass and Mixed Association.







32


Table 4-1. Cumulative number of plant species identified per sampling unit (five 1m2 X 20 sampling points) in the Pantanal Study Area (Caiman Ranch), November, 1991.


SUa MPb MBc SGd SG/TGe TG' MA MG' MG/TG' SC FE'
01 11 12 08 04 07 12 03 09 07 31
02 18 16 10 06 09 16 06 12 17 50
03 25 21 11 08 13 18 11 15 27 58
04 27 22 14 12 16 22 14 16 29 69
05 30 25 14 17 16 27 16 16 30 71
06 30 26 31 72
07 32 27 31
08 34 28 35
09 34 28 36
10 34 28 36
a Sampling unit.
b Marsh Pond.
Moist Basin.
d Short Grass.
'Short Grass/Tall Grass. Tall Grass.
g Mixed Association.
'h Medium Grass.
ii Medium Grass/Tall Grass. Scrub.
Forest Edge.






33


The following groups were recognized by reducing the JSI to > 0.6,: a) Eleochanis elegans, Aeschynomene fluminensis, Hydrolea spinosa, and Hymenachne amplexicaulis (Rudge) Nees (Poaceae) which were common to Marsh Pond;

b) Echinodorus longiscapus Arech (Alismataceae) and Justicia laevilinguis (Nees) Lind. (Acanthaceae) which were common to Marsh Basin; c) Caperonia castaneifolia (L.) St. Hil.(Euphorbiaceae), Diodia kuntzei Schum.(Rubiaceae), and Eleochais acutangula (Roxb.) Steud. which were common to Marsh Basin;

d) Richardia grandiflora, Schizachyrium microstachyum, Axonopus purpusii (Mez) Chase (Poaceae), Vernonia scabra Pers. (V. brasiliensis) (Compositae), Elyonurus muticus (Spr.) Kunth (Poaceae), and Trachypogon sp. (Poaceae) which occurred together in formations including Medium Grass, Tall Grass, Mixed Association, and Scrub.

Classification of Plant Formations

After eliminating rare species from the data base, 73 species were ordered under R-Classification analyses. Plant species abundance within formations is presented in Fig. 4-1.

Qualitatively, two groups of plant formations appeared significantly

distinct (Similarity Index > 0.70): a) Marsh Ponds, Moist Basins, and Short Grass (plants adapted to flood; lowest microrelief ); and b) all other formations (Fig. 42). Group b was constituted by three distinct sub-groups including: Medium/Tall






34


Grass, Short/Tall Grass, and Tall Grass; Medium Grass and Scrub; and Mixed Association and Forest Edge. These sub-groups are possibly related due to similar flood regimes and soil characteristics.

On the other hand, quantitative Q-Classification revealed 4 to 6 distinct groups within the confidence interval for the Bray Curtis distances. The most conservative classification included 4 groups of plant formations: a) Marsh Pond group; b) Short Grass group; c) Scrub group; d) Forest Edge (Fig. 4-3). The lower limit for the confidence interval encompassed 6 groups of plant formations: a) Marsh Ponds; b) Moist Basins; c) Short Grass group; d) Tall grass group; e) Scrub group; and f) Forest Edge.

Only the upper limit of the confidence interval was selected to segregate distinct plant formations and to minimize possible errors of identification during the habitat use observations for ungulates. The plant formations included the following characteristics:

a) low areas that were permanently inundated or strongly affected by periodic flood, i.e. Moist Depressions;

b) areas of low relief that were covered mainly by short grasses; c) areas of intermediate relief that were covered by short and tall grasses, with shrubs and small trees interspersed in different combinations of densities, i.e. Scrub.

d) areas of high relief in the borders of tree islands, i.e. Forest Edge.









35


Medium/Tallgrass

Tall grass Mixed assoc.Mid-grass

ScrubShort grassShort/Tall
grass

Forest edge Marsh pond Moist basin


00 . 0 0 000000 - -ojc o.-....

00 - oo ( 0000).000 O - - 000 oooc(D CO 0-


O OD

0 9 ..00 0.0

0-. 0.00 Oo1.00-- oCXODo....0.




0 0 . - -. .


S - 0 - 0 .


- .0.


... .00000 0 *(flOO 0 - 0. 0 .. - C)000- 00. . - OOO000.O0. 000 0

0 Q-o O--D cM)


5 10 15 20


25 30 35 40 45 50 55 60 65'


Figure 4-1. Quantitative nodal diagram for Indexes of Species Values according to the sampled plant formations at the Pantanal Study Area, (Caiman Ranch), 1991-1992. Species names and index values are presented in Appendix A.


70


1 - Annocom 16 - Vemscab 31 - Angi 46 - Stylacum (2) 61 - Echipani
2 - Ichnproc 17 - Annodioi 32 - Heliguaz 47 - Mimoaden 62 - Sagiguia 3 - Byttdent 18 - Curaamer 33 - Forspube 48 - Hymeampl 63 - Eleoacut 4 - Sebahisp 19 - Wedebrac 34 - Arb.gros 49 - Justlaev 64 - Aescflum 5 - Crotcoru 20 - Brachumi 35 - Arb.BC 50 - Leerhexa 65 - Aeschist 6 - Andrsell 21 - Andrhypo 36 - Ruta 51 - Eleonodu 66 - Eleoeleg
7 - Erytdeci 22 - Sporjacq 37 - Bauhmoll 52 - Nympindi 67 - Hyptmicr 8 - Byrsorbi 23 - Mirtglab 38 - Attaphal 53 - Phyllind 68 - Hidrspin 9 - Psidguin 24 - Cypebrev 39 - Sidalini 54 - Nympsp 69 - Diodkunt 10 - Melovill 25 - Hiptlapp 40 - Convamar 55 - Echilong 70 - Reimbras 11 - Axonpurp 26 - Euphthym 41 - Baccmedu 56 - Capecast 71 - Panilaxu 12 - Elyomuti 27 - Coccsp 42 - Plandesc 57 - Senntor 72 - Chomobtu 13 - Schimicr 28 - Caseacul 43 - Crotsp 58 - Rhyntenu 73 - Melosimp 14 - Richgran 29 - Paulpinn 44 - Stylacum (1) 59 - Eleomin 15 - Tracsp 30 - Adenflor 45 - Ruelgemm 60 - Staccaya I


0








36


01


0.2
0.25

0.3


04


0.9


S06
U U
U 07


08


0.91


10
(ci~ ~ C)) -O(i - c 0 ( A (














Figure 4-2. Classification of plant formations at the Pantanal Study Area (Caiman Ranch) according to the presence of species (Qualitative QClassification). The two horizontal lines represent higher and lower significance limits (a = 0.05). Solid circles indicate significantly distinct groups of plant formations. Open circles indicate significantly related groups of plant formations.












37


0.76


I


I


fl ______._ _______57____I____ ___


U) U) (U
0,
(U I-


ca (U 0)
x


(


E

CD M)


.0

ci U)


U) U) (U 0)

0 U)


cc,


U)
(U 0) ~0 (U

0)
0 U-


*0 C0


cc (U


U)

U) (U
-o
(0 .5


Figure 4-3. Classification of the plant formations at the Pantanal Study Area (Caiman Ranch) according to the relative frequency of species in sample units (Quantitative Q-Classification). The 2 horizontal lines represent higher and lower significance limits (P=0.05). Solid circles indicate significantly distinct groups of plant formations. Open circles indicate significantly related groups of plant formations.


0 'U


1.0



0.9



0.8



0.7



0.6







0.4



03



0.2



0.1



0.0-


W)





E


()


()


()






38


Ordination of Plant Formations

The site ordination diagram, all species included, revealed that points corresponding to Marsh Pond, Moist Basin, Short Grass, and Forest Edge constituted distinct clusters. The remaining formations were less clearly separated (Fig. 4-4). In addition, the ordination diagram demonstrated a gradient (x-axis) ranging from the wettest formation (Marsh Pond) to the driest (Scrub) in the floodplain. Pedological studies in other regions of the Pantanal indicated that the y-axis could be related to the levels of micronutrients in the soil, although measurements of soil characteristics were not conducted in the Pantanal Study Area (Brum et al., 1987, Pott et al., 1987, Pott et al., 1989). Thus, the y-axis would range from the least fertile soils (bottom) to the most fertile soils (top, Fig 4-4).

The species ordination revealed essentially the same pattern as site

ordination, but formations were not distributed as distinctly due to the presence of transition species (Fig. 4-5). However, it suggested that Forest Edge species were related more to the grassland species than to species associated with moist soils.

Richness and Diversity of Plant Formations

An analysis of the richness and diversity was performed for the six groups segregated by the lower limit of the Bray Curtis distances to further explore the differences and similarities among the distinct plant formations (Fig. 4-3). Mixed












39


*


MW
TG3 W


7-7-~: 0




MrG6 0 M


MTG2 S13 MIl - SO W,2R~ SM3A 16


kwa W65 _we-~~
0 134


T 1


-332


-10


Wi


1dX3 W* I4 &Pe --- ** L2

W70
weO

W10 LP6


Figure 4-4. Site ordination diagram showing clusters of sampling units, Pantanal Study Area (Caiman Ranch). MP = Marsh Pond; MB =Moist Basin; SG = Short Grass; STG = Short-Tall Grass; MG = Mid-Grass; MTG = Mid/Tall Grass; TG = Tall Grass; MA = Mixed Association; SCR = Scrub; FE = Forest Edge. The numbers refer to the sample unit.


kG3
SYA )0-1 kG4
SCR2- SM6


I











40


CI


-tcm .*


af * *



A~r1a


H5



~ I~ "*, W


IT
ft La*




-t~ I ~t


TnJPd Ac NE

AtQ'z






037.bmlU


Be >


B ~ po Bud- *' Ti

0 01P'_


Figure 4-5. Species ordination diagram showing clusters of closely associated species Pantanal Study Area (Caiman Ranch). MP = Marsh Pond; MB =Moist Basin; SG = Short Grass; STG = Short-Tall Grass; MG = Mid-Grass MTG = Mid/Tall Grass; TG = Tall Grass; MA = Mixed Association; SCR = Scrub; FE = Forest Edge. The numbers refer to the sample unit.


i


I







41


Association was analyzed independently from the other plant formations to determine how soil disturbance influenced the richness and diversity of Tall Grass.

Species richness indices indicated a trend of increasing richness from Marsh Pond to Forest Edge (Table 4-2, Fig. 4-6). The increase in species richness coincided with decreasing levels of inundation, or increasing elevation of formations. The most flooded formations exhibited the least species richness, and the driest, the greatest richness. Grassland formations exhibited intermediary indices of species richness, mostly due to the presence of rare and ephemeral species amidst the few perennials that could tolerate the extremes of flood and drought. Thus, diversity indices were lower in grassland formations than in other formations, except for Mixed Association. The greater richness and diversity exhibited by Mixed Association was presumably caused by the plowing of a Tall Grass soil, which allowed the colonization of species characteristic of Forest Edge (Figs. 4-2 and 4-3).

Characteristics of Plant Formations

Marsh Pond

Marsh Pond was characterized by the constant presence of water. Water

depth in sampling units ranged from 0 to 40 cm in the dry season and 40 to 80 cm in the flood season. Vegetation height averaged 40 cm, ranging from 0 to 80 cm. Dominant species for this formation were the hydrophytic forbs Hydrolea spinosa and Hyptis microphylla and sedges (Eleocharis spp.); leguminous shrubs (Aeschynomene spp.) also were common (Table 4-3),







42


Table 4-2. Richness and diversity indices for plant formations at the Pantanal Study Area (Caiman Ranch), 1991-1992. Data for all sampling seasons were combined.


MP MBb SGc TGd SCe MA! FE Nh 39 36 52 59 63 66 114
R' 0.78 0.72 1.04 1.18 1.26 1.32 3.80 H'j 2.34 2.48 2.29 2.24 2.30 2.83 3.61 n' 4,000 4,000 4,000 4,000 4,000 4,000 2,400
a Marsh Pond.
b Moist Basin.
Short Grass (Short Grass + Short/Tall Grass). d Tall Grass (Tall Grass + Medium/Tall Grass). 'Scrub (Medium Grass + Scrub). ' Mixed Association.
g Forest Edge.
h Cumulative number of species. Richness Index.
Shannon-Weaner Index of diversity. k Total number of sampling points (20 points/m2).







43


120 100

ci 80 0 60

E 40
z
20

0
0 500 1000 1500 2000 2500 3000
Number of Sample Points + Marsh Pond -x- Marsh Basin - Short Grass + Tall Grass -+- Scrub -u- Forest Edge







Figure 4-6. Rarefraction curves showing cumulative number of species versus sample points for distinct plant formations, Pantanal Study Area (Caiman Ranch), 1991-1992.






44


Moist Basin

Moist Basin was characterized by small and shallow depressions in the

terrain, and thus, exhibited higher moisture levels than the adjacent grassland even during the dry season. These basins accumulated 10 to 20 cm of water during the peak of the rainy season, and 40 to 50 cm during the flood season. This formation was characterized by a transition between Marsh Pond and Short Grass formations. Height of the vegetation varied between 20 and 40 cm. At the beginning of the rainy season (November), ground cover consisted of short grasses (40%), medium grasses (3%), and several forbs (35%); 22% of the surface was bare soil due to cattle trampling. Moist Basin was identified by the presence of its dominant forb species, Melochia simplex (Sterculiaceae), and a common grass, Panicum laxum (Table 4-3).

Short Grass

This formation was characterized by a carpet of short grasses, with

vegetation height averaging 10 cm, but also contained isolated tussocks of tall grass that attained 120 cm in height. Axonopus purpusii (short grass) was the dominant species, but small forbs such as Diodia kuntzei and sedges, especially Eleocharis acutangula, were prominent (Table 4-4). Tussocks of tall grasses (e.g., Schyzachyium microstachyum, Trachypogon sp., Sporobolusjacquemonti) were present, but in very low abundance. Shrubs also were present in low densities, e.g. Vernonia scabra and Annona cornifolia. This formation accumulated 20 to 30 cm of water in the flood season, although it remained dry during most of the year. Short






45


Table 4-3. Relative cover percentages of the six common plant species in Marsh Pond and Moist Basin plant formations in the Pantanal Study Area (Caiman Ranch) 1991-1992. Water regime: Marsh Pond = permanently submersed, with 40 - 80 cm of water in the flood season; Marsh Basin = submersed during the rainy and flood seasons only, with 40 - 50 cm of water in the flood season.


Marsh Pond Moist Basin
Species % Cover Species % Cover
Hydrolea spinosa 28.4 Melochia simplex 32.9
Hyptis microphilla 27.5 Panicum laxum 12.1
Eleocharis elegans 6.9 A. fluminensis 8.4
Eleochanis acutangula 6.0 Eleochanis nodulosa 7.2
A. fluminensis 5.3 Hyptis microphilla 5.1
Aeschynomene histix 4.8 Echinodorus longiscapus 4.8


Table 4-4. Relative cover percentages of the six commonest species for seasonally flooded plant formation types in the Pantanal Study Area (Caiman Ranch), 19911992. Surface water accumulated only during the flood season: SG = 20 - 30 cm; TG = 20 cm; MG < 10 cm; SC < 5 cm.


Short Grass Tall Grass
Species % Cover Species % Cover
Axonopus purpusii 44.6 S. microstachyum 38.5
Diodia kuntzei 9.9 Elyonurus muticus 14.0
Reimarochloa brasiliensis 8.4 Axonopus purpusii 10.8
Vernonia scabra 5.9 Annona dioica 8.4
S. microstachyum 4.6 Richardia grandflora 7.8
E. acutangula 3.4 Helicteres guazumaefolia 3.9
Scrub Forest Edge
Species % Cover Species % Cover
Elyonurus muticus 44.7 Chomelia obtusa 14.1
Axonopus purpusii 8.1 Panicum laxum 7.9
Vernonia scabra 7.3 Psidium kennedyannum 5.9
Annona dioica 6.2 S. microstachyum 5.4
Trachypogon sp. 4.6 Paspalum virgatum 5.2
Curatella americana 4.4 Hyptis spp. 5.1






46


grasses represented 17-53% of the cover, tall grasses 0-28%, forbs 37-44%, and shrubs 8-11%.

Tall Grass

The characteristic of this formation was the dominance of tall grasses, chiefly Schyzachyrium microstachyum (Table 4-4). The lower stratum consisted of short grasses, such as Axonopus purpusi, and forbs such as the very abundant Richardia grandiflora. Short grasses constituted 11-16% of total cover, tall grasses 32-44%, forbs 19-22%, and shrubs 22-35%. Average vegetation height was 120 cm, and flood reached 20 cm depth.

Scrub

This formation was identified by the dominant presence of a tussock grass of 30-40 cm height, Elyonurus muticus, and the abundance of shrubs and small trees. Common shrubs included Vernonia scabra (Compositae) and Annona dioica (Annonaceae) (Table 4-3). The most common tree was Curatella americana (Dillenaceae). Water depths rarely exceeded 2 cm during the flood. Short grasses covered 8-11 % of the terrain, tall and medium grasses 28-33%, forbs 18-25%, and shrubs 25-46%.

Forest Edge

Forest Edge constituted the interface between any of the above formations and semi-deciduous forest patches. Forest Edge consisted of a diverse array of species including vines, fruit trees (Psidium kennedyanum, Myrtaceae), and shrubs (Chomelia obtusa, Rubiaceae) in addition to the species that occurred in the






47


floodplain. The core of the forest patches was covered most commonly by bromeliads and trees, such as Tabebuia heptaphylla (Bignoniaceae), Genipa americana (Rubiaceae), and Vitex cymosa (Verbenaceae). Two palm trees were common, Acrocomia totai and Attalea phalerata.

Discussion

Visual identification of vegetation types was a valid method to discriminate different plant formation types in the Pantanal Study Area. Visual identification correlated with numerically distinct plant formations for moist soils (Marsh Pond and Moist Basin) and Forest Edge. Furthermore, this empirical method, based solely on vegetation structure, allowed descrimination of grassland formations beyond actual differences in species composition and density. Therefore, habitat-use frequencies obtained from visual observations of cattle and cervids could be correctly assigned to the smallest possible variations of habitat within the floodplain of the Pantanal Study Area.

This classification of vegetation represents a level of refinement beyond previous classifications of the floodplain vegetation (Veloso, 1948). General categories of vegetation were partitioned into smaller physiognomic units, which better represented the complex mosaic of marshes-grassland-tree islands, that constituted the floodplains in the study area. More specifically, the Mesophillous Zone (seasonally flooded grasslands) decribed by Veloso was sub-divided into Short Grass, Tall Grass, and Scrub, and the Hygrophillous Zone was segregated into Marsh Pond and Moist Basin.






48


The proposed classification is not exhaustive for the Pantanal, rather it is

limited to the floodplain of the study area. Similar phytosociological studies need to be expanded locally, and especially to other regions of the Pantanal, due to the great diversity of habitats in the region. Traditionally, flooded grasslands have been classified broadly according to the dominance of conspicuous species e.g., "canjiqueiral" dominated by a shrub (Byrsonima intermedia); "caronal" dominated by a species of grass Elyonurus muticus, etc. (Allen & Valls, 1987). However, there has been little phytosociological research in the Pantanal (Veloso, 1948; Eiten, 1985; Prance & Schaller, 1982), and therefore, several distinct plant formations within the floodplains may have been overlooked.

The ordination of the plant formations within the floodplain correlated with the water depth during the flood, wich is ultimately related to the duration of flooding and, probably, to the availability of moisture in the soil during the dry season. The gradient from wet to dry soils, observed for Moist Depression, Short Grass, Scrub, and Forest Edge, indicated that the availbility of water, associated with the microrelief, was possibly the main factor that determined the pattern of distribution of plant formations within the floodplain. Corroboratively, Prance & Schaller (1982) suggested that the degree of humidity, soil type, and duration of standing water were probably responsible for the differences in plant dominance observed in their Pantanal study site. Multivariate analysis (Robertson et al., 1984) indicated that the environmental factors, which were related strongly to the vegetation gradient in a North American swamp, were depth of flooding, soil texture, and drainage.






49


The general reduction of diversity in the plant formations from the Forest Edge to the Marsh Pond probably reflected the stress imposed by increased flooding and poorly drained soils (Robertson et al., 1978, 1984). Diversity was highest in Forest Edge because reproduction, growth, and survival of both Forest Edge and floodplain species were not strongly limited by extremely dry or wet conditions.

The second variable implied in the ordination of plant formations was soil

fertility. Grassland soils have greater levels of Fe and Al than the soils in the forest patches (Pott et al., 1989). High levels of Fe can be toxic (Nores, 1984), and, together with Al, increase deficiencies of P (Conrad et al., 1985). Brum et al. (1987) found that soils in the forest patches had greater levels of P compared to the soils of the scrub and swamps.

Variables such as soil texture, ph, salinity, and frequency of fire may also influence the distribution of plant species within an ecosystem. The multifactor control of plant species distributions has been implicated in numerous studies (Robertson et al., 1978, 1984; Westman, 1980; Muller, 1982). Likewise, there is evidence that other environmental variables than flood and soil fertility are influencing grass species abundance and distribution in the Pantanal Study Area. Axonopus purpusii and Elyonurus muticus, for example, dominated sandy soils, which were characterized by high acidity and low salinity, whereas soils characterized by high levels of salinity were dominated by Paratheria spp., which require neutral or alkaline soils (Allen & Valls, 1987).














CHAPTER 5
CATTLE DIET AND HABITAT SELECTION


Introduction

The future success of wildlife conservation in the Pantanal ultimately

depends on wise use of natural resources. Until today traditional cattle ranching and low human densities have enabled wildlife to thrive in the region. However, there are no studies to verify if modern techniques of cattle management have modified the habitat and disrupted wildlife populations.

Currently, there is an urgent need for knowledge of how cattle and wildlife use the natural resources within their environment, because cattle ranching is the prioritized source of income to entrepreneurs exploiting the combination of cattle ranching and tourism. Therefore, improved forage management techniques that promote both wildlife conservation and cattle production could be derived from such information.

Cattle coexist with two large native herbivores, the brown brocket deer (Mazama gouazoubira) and the pampas deer (Ozotoceros bezoarticus ) in the native pastures of the Pantanal Study Area. Marsh deer (Blastocerus dichotomus) and capybaras (Hydrochaeris hydrochaeris) are restricted to areas adjacent to permanent sources of water, which represent a small portion of the


50






51

Pantanal Study Area. Therefore, the selected study site provides satisfactory conditions for a study of native herbivore habitat use and diet in the presence of cattle.

A complete quantitative analysis of the diet of cattle in the Pantanal has not been reported, although there have been a few studies describing it qualitatively (Allen & Valls, 1987; Pott, 1988). Consequently, the potential competition for palatable forages between cattle and both pampas and brown brocket deer has not been assessed.

Objectives

The objective of this chapter was to investigate the diet of cattle and their habitat preferences. Specifically, the following hypotheses were addressed: H.1. The proportions of distinct plant species (families) in the diet of cattle do not differ among seasons.

H,2. Cattle use the different plant formations according to availability.

Methods

Assessment of the diet of free-ranging animals is a difficult task. Field biologists have employed several techniques to assess diets of ruminants: a) direct observation of what the animal is eating (Pott, 1986; Rajasekaran, 1988; Rodrigues, 1996); b) fecal analysis (Elliott Ill & Barrett, 1985; Green, 1987; Quintana et al., 1994; Martinez et al., 1997); and c) collection of digestive tracts (Pott, 1982; Branan et al., 1985; Bodmer, 1989). The method selected in the present study was required to fulfill the following conditions: 1) be applicable to






52

cattle as well as to free-ranging deer; 2) be humane and harmless to individuals; 3) be able to provide a significant sample of the local population; and 4) not interfere with the natural habits of the individuals sampled. Accordingly, fecal analysis was selected as the method of choice.

Fecal analysis has been used in a number of studies of diet overlap

between cattle and other ungulates (Elliot Ill & Barrett, 1985 ; Jackson & Giuletti, 1988 ; Martinez et al., 1997). It also allows comparisons for the same species in different areas or seasons (Bonino & Shriller, 1991; Quintana et al., 1994; Martinez et al., 1997). Furthermore, fecal analysis is considered as accurate and precise as other existing methods used for the determination of dietary composition in herbivores (Homolka & Heroldove, 1992). Finally, the analysis of the nutrient status of feces may offer a guide to dietary quality of ingesta (Erasmus et al., 1978; Putman, 1984; Green, 1987). Fecal Sampling and Analysis

Five 200-mg samples of fresh cattle dung were collected in the first week of each month and preserved in 75% ethanol in individual plastic vials. Samples were pooled by season: rainy (November, December, and January); flood (March, April, and May); dry (July, August, and September). The adequacy of number of samples was tested as described by Hanson & Graybill (1956). Samples were washed, clarified in NaOH (10%), and passed through a net of 80 BTN to segregate the undigested fragments. Samples were equally diluted to reach a density of 3-6 identifiable fragments per microscopic field. Four slides of






53

each fecal sample were mounted in silicone gel. Fragments from 25 fields (100 x magnification) on each slide were identified by comparison with the reference collection. The resulting frequency data for each identified species were then transformed into the percentage composition of the diet by season (Johnson, 1982).

Diet composition was compared among seasons using Spearman Rank Correlation Index (Sokal & Rohlf, 1969). Differences between the percentage contribution of plant species, families, and food categories were tested with the STP Mann-Whitney pairwise tests, which control the comparisonwise Type I error rate (Sokal & Rohlf, 1969). Trophic diversity was expressed by the Shannon-Weaner Index, and niche breadth by Levins method as standardized by Hurlbert (Krebs, 1989).

Reference Collection

Leaves and thin stems were washed and cleared as described by Sparks and Malecheck (1968). Grasses were prepared by rasping the mesophilum from the leaves so as to leave only fragments of the epithelium. Microhistological slides were mounted in silicone gel and labeled. Drawings of each species' histological structures such as stomata, epithelial cells, inclusions, and trichomes were made, from the slides in the collection, to save time and facilitate identification of microscopic structures in the fecal samples. A key to segregate species was created to further aid in the identification of fragments by using a






54

group of diagnostic characteristics such as stomata, trichomes, and cell inclusions as basic discriminatory elements. Habitat Use

Habitat-use observations of cattle and native mammals were obtained during daylight hours in all plant formations by excursions on foot and by vehicle. A minimum of one survey of 64 km of dirt roads was conducted monthly by vehicle at an approximate rate of 15 km/h during early morning or late afternoon. Incursions on foot were performed once per month year-round across a 2-km trail located in a strip of forest within the floodplain. Ad libitum observations also were recorded when traveling the study area.

The availability of the different types of plant formations (Chapter 4) was obtained by a random sample of 500 points distributed within the floodplain. This method was selected over the planimetry of the different plant formations based on satellite imagery, because of small size and discontinuous pattern of distribution of the plant formation patches. Forest patches and Moist Basin, for example, were frequently smaller than 200 m2

Actual proportions of plant formations are recommended for parametric statistical analyses in use/availability studies (Alldredge & Ratti, 1986). However, estimates are acceptable if large sample sizes are obtained so that the margin of error is sufficiently small (Thomas & Taylor, 1990). Thomas & Taylor (1990) compared different methods for determining the number of random points necessary for studies with three or four distinct resources (95% confidence






55

limits), and found numbers ranging from 385 points (Marcum & Loftsgaarden, 1980) to 510 (Thompson, 1987). Therefore, the sample size of 500 points used to estimate plant formation availability in the Pantanal Study Area was considered adequate, given the low number of different categories treated by the analysis (Thomas & Taylor, 1990).

The procedure started by superimposing a grid of 1 ha on a topographic map of the Pantanal Study Area. Ten cartesian coordinates were then randomly selected, and used as starting point for 10, 2.5 km length transects; directions for conducting the transects also were selected randomly. The vegetation structure and plant species dominance were recorded on the transect at 50-m intervals to avoid dependence between sampling points. Later, the frequency data were assigned to the distinct plant formation types (Chapter 4) by using the information collected on structure and plant species dominance.

Significance of difference in use versus availability of plant formations was tested with the Chi-square goodness-of-fit-test, and the related multiple comparisons of the Bonferroni test (Neu et al., 1974; Byers et al., 1984). This method requires two assumptions: 1) availability is measured correctly; and 2) observations are independent events. The fact that cattle form herds might violate the second assumption. However, the large number of observations and the small size of plant formation patches relative to the size of a herd were assumed to minimize the chance of error.







56

Results

Diet Composition

A total of 78 plant species from 25 families was recorded in the annual diet of cattle. Most of the diet consisted of grasses (x = 80.1 %; sd =7.2%; 20 species) and sedges (x = 8.3%; sd = 5.1%; 11 species). The families Sterculiaceae (x = 4.4%; sd = 3.0%; 7 species) and Lamiaceae (x = 2.1%; sd =

1.5%; 3 species) were the most represented among non-graminoid families. No other family comprised more than 1 % of the annual diet (Table 5-1).

Fewer species were consumed by cattle during the rainy and flood seasons, 62 and 63 species respectively, than during the dry season, 74 species. However, there was no significant difference in diet between seasons because cattle foraged mostly on the same grasses and sedges throughout the year (Table 5-2).

The percentage use of the 11 most consumed species varied little among seasons, except for Mesosetum chaseae, which was consumed more in the flood season than in the rainy or dry seasons (Table 5-3) . On the other hand, the dominance of species varied by season (Kruskal-Wallis, P > 0.05). Axonopus purpusii (Short Grass), Hymenachne amplexicaulis (Marsh Pond), and M. chaseae (Scrub) were the dominant species in the diet during the rainy season (STP, P 0.03). M. chaseae was the most important species during the flood season. M. chaseae grows in areas less subjected to flooding than H. amplexicaulis and A. purpusii (Chapter 4), which might explain its greater







57


Table 5-1. Number of species consumed and percentage consumption of major plant families (> 5%) in the diet of cattle, Pantanal Study Area (Caiman Ranch), 1991-1992.


Percentage Composition of Diet
Rain Flood Dry Year
Family n % n % n % Tb b A
Poaceae 19 80.0 20 86.0 20 74.2 20 80.1
Cyperaceae 10 7.0 9 4.5 11 13.5 11 8.3 Sterculiaceae 6 7.8 2 3.5 7 2.0 7 4.4 Lamiaceae 2 2.1 2 2.1 3 2.0 3 2.1
Others 25 3.1 30 3.9 33 8.3 37 5.1
a Number of species consumed by season. b Total number of species consumed. Average percentage of use.


Table 5-2. Comparison of the cattle diet between seasons, Pantanal Study Area (Caiman Ranch), 1991-1992.


Comparison rsa nb 95% Cl
Dry/rainy 0.76 29 0.546 to 0.881 Dry/flood 0.70 29 0.451 to 0.850 Rainy/flood 0.86 29 0.724 to 0.934
a Spearman's Rank Correlation Index. b Number of food items in comparisons.


P value
<0.001 <0.001 <0.001


Result
Similar diets Similar diets Similar diets






58

utilization in flood season. A. purpusii was consumed more than the other species during the dry season (STP, P < 0.03). Habitat Use

Cattle foraged predominantly in Short Grass during the rainy season (Table 5-4). Short Grass offered an almost homogeneous source of palatable growing grasses, which allowed the cattle to travel less in selecting food items. Cattle shifted their foraging activities to Scrub when Short Grass was inundated by flood waters and foraged on the new-growth of Short Grass after the flood receded. However, the forage was depleted rapidly because of the drought that followed the flood. Thus, cattle shifted foraging to Scrub after depleting the forage supply in Short Grass, which was indicated by the appearance of bare soil within the carpet of grass.

Seasonal Diet Diversity and Niche Breadth

The analysis of trophic diversity and niche breadth suggested that cattle adopted different feeding strategies according to the limitations imposed by seasonal climatic variations on food availability. New-growth of grasses provided cattle with abundant diversity and quantity of forage throughout the range in the rainy season. Cattle concentrated foraging activities in the Scrub during the flood season when a greater proportion of their diet consisted of fewer grass species, such as Mesosetum chaseae, presumably in response to a limitation in the availability of selected species (Tables 5-3 and 5-4). Consequently, trophic diversity and niche breadth were reduced from the rainy season to the flood









Table 5-3. Percentage consumption (> 5%) by season of graminoid species to the diet of cattle as determined by fecal analysis, Pantanal Study Area (Caiman Ranch), 1991-1992.


Percentage Composition


Family
Poaceae










Cyperaceae


59


Species
Axonopus purpusili Brachiaria sp. Cynodon dactylon Hymenachne amplexicaulis Leersia hexandra Mesosetum chaseae Panicum laxum Paspalum plicatum P. pontanalis Reimarochloa brasiliensis Cyperus sesquiflorus
Other Species


H'a


Rainy
10.10
2.40 2.15 6.96 5.07 15.20
4.62 5.64 5.16 5.23 3.05
34.42 3.15


Bab 0
a Shannon-Weaner Index of diversity. b Niche Breadth Index (standardized Levins).


.20


Flood
8.06 1.26
5.41 5.43 2.51
33.40 2.82 2.79 5.73 1.99 1.05 29.55 2.79 0.09


Dry
20.60 5.16 2.70 2.86 2.79
4.84 5.53
2.01 6.24 3.20 5.67
38.40 3.31 0.18


Average
12.92
2.94 3.42 5.08
3.46
17.81
4.32 3.48 5.71
3.47 3.26
34.13


Table 5-4. Habitat selection of cattle, Pantanal Study Area (Caiman Ranch), 1991-1992.


Habitat
Season X2 P value MDb SGc SCd FEe
Rainy 318.84 <0.001 Avoided Selected NS Avoided
Flood 272.72 <0.001 Avoided Avoided Selected Avoided Dry 465.23 <0.001 Avoided Avoided Selected Avoided
a Avoided = habitat used less than expected based on its availability. Selected = habitat used more than expected based on its availability. NS = no selection, habitat used in proportion to its availability (Bonferroni tests,
a =0.05).
b Moist Depression.
Short Grass.
d Scrub.
e Forest Edge.







60

season (Table 5-4). Then, during the dry season, cattle consumed more of less selected food types, such as sedges, forbs, and bowse, in consequence of a reduction in biomass and palatability of grasses. The relative increase of these items in the diet caused a concomitant compensation in the trophic diversity and niche breadth.

Discussion

The percentage consumption of different food types as well as the selection among different plant formations by cattle varied according to the season. It is presumed that this variation was caused by differential availability and quality of selected food items in time and space. The distribution of biomass in fiber categories is spatially and temporarily dynamic, and herbivores can be expected to manipulate their food intake not only by dietary selection, but also by habitat choice and seasonal movement (Jarman & Sinclair, 1979; Hansen et al., 1985; Gordon 1989a, b; Murray & Brown, 1993).

Although feeding preferentially on Short Grass and Scrub, cattle in the Pantanal showed marked seasonal changes between these two formations. During the rainy season they selected Short Grass, and in the flood and dry seasons they selected Scrub. The plant formation use and selection by cattle in the Pantanal showed similar trends to that of cattle on the island of Rhum and to the African buffalo in the Serengeti (Jarman & Sinclair, 1979; Gordon, 1989a, b). Cattle and buffalo are primarily grazers because their large gut capacity and relative low energetic requirements per unit of weight allow them enough time to






61

digest the carbohydrates of the plant cell-walls (Hofmann, 1973; Van Soest, 1982). On the other hand, these large bovids are forced to feed on a lower quality diet to meet their total daily energetic requirements, because highly digestible plants are much less abundant (Parra, 1978; Demment & Van Soest, 1985). Thus, buffalo occur everywhere in the Serengeti except on the short grass plains, which might have too low a biomass of forage to maintain such a large ruminant (Illius & Gordon, 1991). Similarly, cattle in the Isle of Rhum were able to use short grasslands only during the spring and summer, because of the large amount of vegetation growth. As the abundance of live material declined in the fall season, cattle left the short grasslands to feed on the tall grassland areas (Gordon, 1989a, b).

The seasonal characteristic of the climate determined presumably the variation of availability and quality of forage in time and space in the Pantanal. During the rainy season cattle found optimal conditions of foraging in the Short Grass due to the higher proportion of new-growth of grass leaves to structural tissues in this plant formation than in the Scrub, which have a larger proportion of tall grasses. Demment and Van Soest (1985) compared the percentage biomass in cell wall for different grasslands and demonstrated that a greater percentage of the biomass tended to be concentrated in the high-fiber categories as the standing crop increased. This occurred because taller grasses possess a large proportion of stems and also because maturity and late-season temperatures increase the cell-wall content of grasses (Deinum & Dirven, 1971).






62

Thus, it is valid to assume that Short Grass represents the best foraging ground for cattle during the growing-season.

Short Grass was completely covered by water during the flood season in the Pantanal forcing cattle into the Scrub. Some species, such as A. purpusi, that were common to this habitat, became less available and cattle increased the percentage of other species in the diet. The decrease of growth and advancing maturity of grasses in the dry season made cattle to include a greater proportion of other food categories in the diet because selected species were less available and presumably less digestible. The exhaustion of selected forage in the Short Grass induced cattle to forage preferentially on the Scrub during that season as also Gordon (1989a) observed in the Isle of Rhum.

Studies of cattle feeding habits in other regions of the Pantanal confirm the results of dietary preferences of the main species consumed in this study, especially A. purpusii and M. chaseae (Allen & Valls, 1987; Pott, 1988). In addition to these two species, Pott (1982, 1986) reported that Panicum laxum was an important species for cattle in the Paiaguas and Nhecolandia regions of the Pantanal. However, P. laxum was not consumed in a greater proportion than other grass species in the Pantanal Study Area, even though it was abundantly available. In fact, its consumption was lower than Hymenachne amplexicaulis and Paspalum spp.

Other grasses of secondary importance to cattle in the Pantanal Study Area, Cynodon dactylon, Leersia hexandra, Reimarochloa brasiliensis, and






63

Setaria geniculata, also were considered less important to cattle by Pott (1982, 1986). On the other hand, Pott (1982) suggested that Andropogon spp, Elionurus muticus, and Trachypogon sp. were consumed regularly by cattle. The present study, however, indicated that the 3 latter species were avoided, as cattle did not consume E. muticus, and ingested little of Andropogon spp. and Trachypogon sp..

Avoidance of certain plants may be related to the amount of silica and/or large percentage of cell wall in plant tissues, because the micronutrient content of plants cannot explain the full spectrum of dietary preferences exhibited by cattle. For example, Pott et al. (1987) found that A. purpusii was richer in Ca and Mg than M. chaseae, but similar in K and P contents. Also, only A. purpusii reached the Ca:P ratio (1.9:1.0) considered satisfactory for ruminants by the National Research Council (1976). Similar results were obtained by Brum et al. (1987) for the Pantanal's sub-region of Paiagues, where A. purpusii showed higher contents of Cu and Zn. However, indices of preference (Chap. 08), for the top species consumed by cattle, indicated that A. purpusii was consumed less than expected in relation to its availability in the floodplain, whereas M. chaseae was consumed in proportion to its availability. Therefore, differences in silica and digestibility between the two species may explain the observed differences in consumption.

Another important aspect for cattle and range management in the

Pantanal is that cattle use native forbs and browse to complement their diet,






64

which can not be satisfied nutritionally by grasses only (Brum et al., 1987). Nongraminoid species averaged aproximately 12 % of the seasonal and annual diets of cattle, a value higher than reported previously (1%) (Pott, 1982; 1986). Forbs and shrubs of the Sterculiaceae family were especially significant during the rainy season. In another study, in addition to Helicteres guazumaefolia (Sterculiaceea), Pott and Pott (1987) also reported that Attalea phalerata, Cecropia pachystachya, Cordia glabrata, Costus sp., Smilax sp., Vitex cymosa,

and few other plants not sampled in this study, were the most important forage based on nutritional value, acceptability by cattle and frequency of occurrence.

Native grasses and sedges of the Pantanal did not contain the minimum levels recommended for Ca, Mg, P, Cu, and Zn during most of the year (Brum et al., 1987; Pott et al., 1987). For example, Grace (1983) recommended 10 ppm of Cu for lactating cows, but none of the grass or sedge species studied by Pott et al. (1989) achieved this levels. Deficiencies of plant micronutrients are more severe during the rainy season (Pott et al., 1989). However, Pott et al. (1989) found no significant seasonal differences of limiting micronutrients, such as Cu, in the liver of cows. Pott & Pott (1987) suggested that cattle complemented their need for Cu by consuming other plant species rich in this element, such as Cordia glabrata (Boraginaceae), Byrsonima orbigniana (Malpighiaceae), Vernonia scabra (Compositae), Helicteres guazumaefolia (Sterculiaceae), and Chomelia obtusa (Rubiaceae).






65

Collectively, these studies indicate that cattle in the Pantanal require the plant diversity provided by native pastures in order to meet their year-round nutritional needs. Furthermore, the survival of wildlife depends on the current diversified ecosystem, especially native forages. The transformation of the floodplain into mono-specific introduced pastures would create two negative effects: a decrease in sources of nutrients to cattle and a reduction of the wildlife diversity.

Currently, the Pantanal has a carrying capacity three times less than the improved pastures outside the floodplain (Allen & Valls, 1987). Productivity of forage for cattle could be increased by controlling the two most important weed species in the study area: Elyonurus muticus and Schizachyrium microstachyum. Allen & Valls (1987) recommended plowing to control Elyonurus muticus so as to allow fast-growing species such as Axonopus purpusii to colonize these areas; however, such a measure is unlikely to be practiced currently due to its probable high cost-benefit ratio. Cattle grazing should be maintained within carrying capacity to avoid proliferation of weed species, and pastures should be grazed in a rotational scheme both within and between seasons.

Currently, cattle are kept in the same pastures year round. Movement of cattle among pastures is exercised only when seasonal flooding reaches intolerable levels. However, cattle density should be reduced in the floodplains during the flood to reduce grazing intensity to that dictated by the available carrying capacity. Cattle uproot tussocks of grasses and trample the soft soil






66

during the flood causing large areas of bare soil when the water recedes. As a result, not only is carrying capacity reduced, but also, the invasion of weed species is encouraged.

Presently, EMBRAPA (the Brazilian Institute for Agricultural Research)

recommends a grazing quota of one cattle-unit per 3 hectares of native pastures for an extensive cattle management regime in the Pantanal (Dr. Ara6, pers. comm.). However, this figure represents an average, that does not consider individual carrying capacity for the different types of plant formations that form the floodplain mosaic. Therefore, the knowledge of the carrying capacity for each of the identified plant formations (Chapter 4) will help to improve cattle management and productivity.

The rational and efficient use of the native pastures of the Pantanal, plus the development of new breeds of cattle more adapted to the ecosystem, could guarantee cattle productivity (Behnke & Abel, 1996) better aligned to the maintenance of biodiversity. Ultimately, the preservation of biodiversity is important not only to the conservation of natural resources, but also to the development of ecotourism in the region.














CHAPTER 6
BROWN BROCKET DEER DIET AND HABITAT USE Introduction

The brown brocket deer, Mazama gouazoubira, is a small Neotropical cervid that occupies a broad variety of habitats from Mexico to northern Argentina (Avila-Pires, 1959; Czernay, 1987). Its shoulder height ranges from 350 to 610 mm, and its weight from 13 to 25 kg depending on the geographic region and ecosystem (Eisenberg, 1981; 1989; Czernay, 1987; Emmons, 1990; Redford & Eisenberg, 1992; Townsend, 1996). The smallest and darkest forms occur in tropical rainforests, whereas the largest and palest forms inhabit savanna habitats.

Hunted for food, sport, and pelts trade (Ojeda & Mares, 1982; Bodmer,

1989, Townsend, 1996), brown brocket deer occur in estimated densities of 0.83 individuals/km2 in the Amazon (Bodmer, 1989), 1.4-2.1/km2 in the Brazilian Cerrado (Leeuwenberg & Resende, 1994), and 0.5-2.75/km2 in the Pantanal (Schaller, 1983; Lourival & Fonseca, 1997). Brown brocket deer are usually solitary, but may group for reproduction or for feeding when food is limited and patchy. Fawning may occur in any month of the year (Stallings, 1986; Redford & Eisenberg, 1992, Bisbal, 1994; Townsend, 1996).


67






68

Brown brocket deer are found from the Atlantic coast to the Andes, and occupy a variety of ecosystems, including the forests of the northern South America and the eastern Brazilian Atlantic Forest, and the bushlands of the Venezuelan Llanos, Brazilian Cerrado, Pantanal, Paraguayan Chaco, Argentina, and Uruguay (Czernay, 1987). Brown brocket deer also can survive in cultivated areas, provided forested patches are available and hunting pressure is not intense. The spectrum of ecosystems occupied demonstrates the degree of adaptation that is characteristic of this species.

Brown brocket deer are considered frugivores (Bodmer, 1989; 1991),

although they also consume browse and flowers (Branan et al., 1985; Stallings, 1984). On the other hand, brown brocket deer would be expected to exhibit diets reflective of the local availability of food items as suggested from the wide geographic distribution and range of habitats occupied by this species. For example, most of the plant biomass in the rainforest is located in the canopy, and, thus, is unavailable to cervids. Fallen fruits, due to their overt abundance, are perhaps the most significant source of food for brown brocket deer and to a number of other terrestrial herbivores in this ecosystem. In contrast, one may expect to find a larger proportion of browse in the brown brocket diet in open habitats such as the Pantanal, even though fruits also are present in the floodplain. Thus, the Pantanal is an ideal ecosystem to test the hypothesis that brown brocket deer are essentially frugivores.






69

Objectives

The following assumptions should be met if brown brockets are frugivores: H.1. Fruits are available throughout the year. H.2 The largest proportion of the diet consists of fruits. H,3. The selected habitat is the forest patches where fruit trees are found.

Methods

Fecal Sampling and Direct Observations

Five samples of brown brocket deer pellets were collected in the first week of each month, and preserved in 75% ethanol in individual plastic vials. Samples were pooled by season: rainy (November, December, and January); flood (March, April, and May); and dry (July, August, and September). Fresh samples were collected in early morning from "latrines" or known bed-sites of individual deer at the edge of forest patches. In a few occasions, pellets were collected after observed individuals had defecated. Additionally, defecation within specific home ranges of individuals was induced by depositing pellets from one brown brocket deer into the home range of another. This method was employed after a radio-collared male was observed defecating beside the pellets of both sexes at the edge of its home range (Pinder, 1992). Monthly samples were assumed to be from adult males and females because samples were collected from sites known to be used by identified individuals (five females and six males).






70

Direct observations of foraging events were used to help in identifying

plant species consumed by brown brocket deer, and as a comparative method to fecal analysis. Observations were conducted daily in hourly blocks from 0600 to 1800 hrs, which coincided with the time period when data on habitat use and location of the radio-collared male were collected. Ad libitum observations of feeding events for at least 10 additional different individuals were conducted in diverse portions of the Pantanal Study Area. Microhistologic Analyses

Fecal samples were washed, clarified in NaOH (10%), and mounted on four semi-permanent microhistologic slides. Distinct fecal samples were equally diluted to reach a density of three to six identifiable fragments per microscopic field. Twenty-five fields were located systematically on each slide, and were viewed at 1 00x magnification for identifiable fragments. The resulting frequency data for each identified species were then transformed into the percentage composition of the diet by season (Johnson, 1982).

The presence of loose trichomes (epidermal hairs) of a given plant

species in the fecal samples may result in a bias for quantification of the relative contribution of each plant species in the diet. Trichomes are very abundant in some plant species and the simple quantification of the frequency of loose trichomes for every sampled microscopic field can overestimate the relative presence of that given species in the diet. Johnson (1982) recommended that only fragments of epithelium should be considered for the quantification to avoid






71

this problem. However, the counting of epithelium fragments only would result in an underestimation of those species in the diet of brocket and pampas deer because of the great digestibility of these particular species.

I created a correction index for species of the families Malvaceae and Sterculiaceae, which possess abundant trichomes, to minimize this problem. Random samples of fecal material collected during the rainy season were quantified for 100 microscopic fields containing trichomes of species of the two families. A correction factor was then derived by dividing the frequencies of fragments with trichomes by the frequency of loose trichomes:

Cl= E freqspi
E tricspi

where:

CI = correction index;

freq = frequency of epithelium fragments with trichomes of species i; tric = frequency of loose trichomes of species i.

The analysis revealed that 37% of the microscopic fields contained

fragments of trichomes that were attached to the epithelium. This percentage was used as standard for adjusting the frequencies of Malvaceae and Sterculiaceae species in a given fecal sample.

Flowers, fruits, and seeds of grasses were difficult to identify after

digestion, and, therefore, their quantification was more subject to accuracy bias than the remains of grasses and woody browse. Nonetheless, the microhistologic method was used to assess the seasonal consumption of these






72

food categories because it allowed for better precision in contrast to quantification by direct observations.

Differences of percentage of use of species, families, and food categories were tested for significance with Kruskal-Wallis test (Sokal & Rohlf, 1969). Hierarchy of consumption among food items (species and categories) was verified with STP Mann-Whitney pairwise tests, which control the comparisonwise Type I error rate (Sokal & Rohlf, 1969). Correlations of diets between seasons was performed with the Spearman Rank of Correlation Index (Sokal & Rohlf, 1969). Trophic diversity was expressed by the Shannon-Weaner Index, and niche breadth by the Levins method as standardized by Hurlbert (Krebs, 1989).

Phenoloqy of Trees

The phenological stage of the most common species producing edible fruits in the floodplain was recorded from November 1991 to October 1992. A sample of 86 trees of seven species were selected randomly for determining fruiting phenology. Individual trees were marked with colored vinyl tape and visited at 30-day intervals. Data recorded on the monthly visits to each tree included the frequency of trees with flowers, the number of fruits on the ground beneath each marked tree, and the presence of brown brocket deer or peccary sign. Additonally, specimens of Tabebuia caraiba and Tabebuia heptaphylla (Bignoniaceae) also were monitored, after a brocket deer was observed eating their flowers.






73

The selected species for fruiting analyses consisted of three palm trees (Acrocomia totai, Attalea phalerata, and Copernicia australis), one Rubiaceae (Genipa americana), one Myrtaceae (Psidium kennedyanum), one Cecropiaceae (Cecropia pachystachia), and one Fabaceae (Enterolobium contortisiliquum). A. phalerata and E. contortisiliquum grow inside tree islands. Copernicia australis inhabits the flooded plains, especially those that suffer cycles of flood and fire. The remaining species were found principally along the edge of forest patches.

The data on fruiting were pooled by month for each species. To

standardize the relative availability of fruits, an index of fruit availability was created:

FAI = T * F * 10
N

where:

FAI = fruit availability index;

F = number of ripe fruits of each species found on the ground each month; T = number of trees of each species with ripe fruits on the ground each month; N = total number of ripe fruits on the ground during the study. Habitat Use

Observations of brown brocket deer were obtained during daylight hours in all plant formations by excursions on foot and by vehicle (Chapter 5). Additionally, a radio-collared adult male was monitored (April 1991-October 1992) to provide an equally proportioned number of observations of habitat use






74

during all daylight hours. Ad libitum observations of habitat use were also recorded.

To ensure independence of data, 30-minute intervals were considered for the quantification of plant formation use (Swihart & Slade, 1985). Observations spaced temporally in this manner allowed brocket deer to have access to their entire area of use, which was approximately < 100 ha monthly (this study). Only locations performed during foraging or traveling events were used for habitat preference analysis. Availability of distinct plant formations in the Pantanal Study Area was obtained by transect sampling (Chapter 5).

Results

Diet Composition

Microhistologic analyses of feces revealed that brown brocket deer

consumed at least 139 species of 39 plant families in the Pantanal Study Area. The families that contributed most to the diet were Compositae, Euphorbiaceae, Malvaceae, Sterculiaceae, and Rubiaceae. However, there was no significant difference in consumption among these five families within seasons (KruskalWallis, P > 0.05), except for the rainy season, when Sterculiaceae was selected (STP, P = 0.009) (Table 6-1).

Forbs and browse dominated the diet of brown brocket deer (STP, P < 0.05), comprising > 80% of their diet (x = 84.6%, sd = 8.4%) during the rainy, flood, and dry seasons (Table 6-2). There were no significant differences in the consumption of forbs and browse within each season (STP, P > 0.05). Browse







75


Table 6-1. Percentage composition (>5%) of plant families in the diet of brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 1991-1992.


Percentage Composition of Dietb
Rainy Flood Dry Year
Plant Family na % n % n % n %
Compositae 5 15.20 4 18.60 5 15.02 6 16.27
Euphorbiaceae 5 10.58 6 11.02 5 13.08 7 11.56 Malvaceae 1 10.62 2 11.62 2 13.56 2 11.93
Rubiaceae 6 13.56 6 18.68 3 17.96 6 16.73
Sterculiaceae 7 27.14 7 16.90 6 15.50 7 19.85 Others 71 22.90 72 23.18 63 24.88 139 23.65
a Number of species in the family.
b See appendices for statistical tests comparing percentages.


Table 6-2. Percentage composition of different food categories in the diet of brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 19911992.


Rainy Flood Dry Year'
Food Category FAc DOd FA DO FA DO FA (398)e (119) (372)
Graminoids 1.32 4.77 3.04 27.73 3.55 0.27 2.64
Small Forbs 22.67 28.14 33.39 13.45 47.01 54.84 34.36 Browsea 58.17 35.93 54.11 50.42 37.59 37.63 49.96
Broad Leaf Forbsb 0.04 0.00 0.73 0.00 0.12 0.00 0.30 Lianas 2.15 0.25 1.60 5.88 2.82 4.84 2.19
Grass Seeds 0.00 11.31 0.00 0.01 0.00 0.00 0.00
Flowers 0.00 0.00 0.00 0.00 0.00 2.42 0.00
Fruits 8.42 19.20 2.76 1.68 1.71 0.00 4.30
Unidentified 7.23 0.65 4.37 0.83 7.20 3.72 6.26
a Shrubs and tree seedlings.
b Alismataceae and Pontederiaceae.
c Fecal analysis.
d Direct observations.
Number of observations.
Mean values among rainy, flood, and dry seasons.







76

was consumed more than forbs only during the rainy season (STP, P = 0.009). Both fecal analyses and direct observations confirmed that other food categories were of minor importance, at least quantitatively, in the diet of the brown brocket deer. Direct observations indicated also that fruits and seeds of grasses were consumed more frequently in the rainy season, whereas tree flowers were consumed more frequently in the dry season. Flowers and grass seeds were not identified in fecal samples, possibly because of a higher digestibility of these tissues (Table 6-2).

The percentage consumption of graminoids and browse remained constant between seasons (STP, P > 0.05), but forbs were consumed less during the rainy season than during the flood or dry seasons (STP, P = 0.028) . Direct observations indicated that the reduced consumption of forbs during the rainy season was replaced by the increased consumption of fruits.

Overall, 10 species of plants constituted the main source of food in the diet of brown brocket deer: Vernonia scabra, Wedelia brachycarpa, Caperonia castaneifolia, Euphorbia thymifolia, Sida santamarensis, Chomelia obtuse, Richardia grandiflora, Bytneria dentata, Melochia pyramidata, and Melochia villosa (Table 6-3). Percentage consumption differed among these 10 plant species for rainy and dry seasons (Table 6-3) (Kruskal-Wallis, P < 0.02). In the flood season, however, the great variance in the diet of different deer did not allow for a significant difference of consumption among the top 10 plant species (Table 6-3) (Kruskal-Wallis, P = 0.126).






77

Only a few species of fruits were observed being consumed by brown brocket deer: Soroceia spruce, Vitex cimosa, Chomelia obtusa, Genipa americana, and Acrocomia totai. Seeds of Psidium guinense were found in the stomach of a female found dead during the dry season. Finally, fragments of Eryngium ebracteatum fruits were identified through microhistologic analysis. Large seeds (> 1 cm) of Acrocomia totai and Vitex cymosa fruits were discarded by the animals after mastication of the fleshy pericarp. Only fallen flowers of Tabebuia spp., especially Tabebuia caraiba (Bignoniaceae), were observed being eaten in abundance by brown brocket deer.

Diet composition did not vary much from season to season, but a few plant species were consumed in greater proportions in a specific season. Two browse species characteristic of the Forest Edge, Bytneria dentata and Melochia pyramidata (Sterculiaceae), were consumed in greater proportions during the rainy season than during the flood and dry seasons (Kruskal-Wallis, P 0.01). Two other browse species characteristic of the grasslands, Chomelia obtusa and Vernonia scabra, were consumed in greater proportions during the flood season, and a forb associated with the Scrub, Wedelia brachycarpa, was consumed the most during the dry season (Kruskal-Wallis, P 0.07).

Overall, a comparison of the total diet between seasons indicated a significant similarity among the seasons (Rs > 0.48, P < 0.001; Table 6-4). However, the diet composition for the dry season differed from the rainy season(R, = 0.34, P = 0.065), when only browse and forb species were







78


Table 6-3. Percentage composition (> 5%) of species by season to the diet of brown brocket deer as determined by fecal analysis, Pantanal Study Area (Caiman Ranch), 1991-1992.


Species
Vernonia scabra Wedelia brachycarpa Caperonia castaneifolia Euphorbia thymifolia Sida santamarensis Chomelia obtusa Richardia grandiflora Bytneria dentata Melochia pyramidata Melochia villosa,


Ba'
a Plant species not identified. b Shannon-Weaner index of diversity.


Rainy
6.81 0.08 1.00
0.04 12.21 3.57 3.16
14.17 19.29 2.87 32.09
14.23 3.24 0.11


Flood
16.16 2.05 0.09 7.21 11.52
10.49 0.82 2.35 1.57 8.33 37.18 4.88 3.20 0.10


Compositiond
Dry Average
1.02 8.00 13.58 5.24 8.37 3.15 0.46 2.57 8.48 10.74 0.42 4.83 17.46 7.15 0.67 5.73 5.00 8.62 5.09 5.43 31.77 33.68 8.93 9.93
3.20 0.09


Niche breadth index (standardized Levins). d See appendices for statistical tests comparing percentages.







Table 6-4. Comparison of brown brocket deer diets (browse and forbs) between successive seasons, Pantanal Study Area (Caiman Ranch), 1991-1992.


Comparison rsa nb 95% Cl P value Result
Dry/rainy 0.34 30 0.000 to 0.625 0.065 Different diets Dry/flood 0.50 39 0.224 to 0.707 0.001 Similar diets Rainy/flood 0.50 46 0.244 to 0.690 < 0.001 Similar diets
a Spearman's rank correlation.
b Number of food items in comparisons.


Percentage


Family
Compositae Euphorbiaceae Malvaceae Rubiaceae Sterculiaceae


Other Species niH
H' i






79

considered due to the increased contribution of forbs to the diet during the dry season (STP, P = 0.028; Table 6-2). Nevertheless, the values of the Shannon Diversity Index and Trophic Niche Width obtained for the diet were similar in the three seasons, suggesting that the dry season was not limiting for brown brocket deer in the Pantanal Study Area (Table 6-3). Availability of Fruits and Flowers

Data from the phenological study suggested that fruits consumed by brown brocket deer (A. totai, G. americana, and P. kennedyanum) were relatively unavailable on the ground, except during the rainy season, although no attempt was made to quantify the biomass of fruits in the Pantanal Study Area (Figs. 6-1, 6-2). Large flowers of Tabebuia spp., which were consumed by cattle, pampas deer, and brown brocket deer, were available during the dry season (July to September). Vitex cimosa (Verbenaceae) trees produced a great number of olive-like fruits from October to December, which were also observed being consumed by brown brocket deer.

P. kennedyanum (berry), E. contortisiliquum (dry fruit), and C. australis (lignified palm fruits) had the highest individual production per individual tree, but fruits of A. totai (pulpy mesocarp under a thin hard pericarp) were the most frequently available overtime (Table 6-5). The largest combined availability of fruits on the ground, except for dry and lignified fruits, occurred from October to February, which coincided with the rainy season. The peak of fleshy fruits abundance occurred in November, when P. kennedyanum also produced fruits







80

a)


100

80
4
o 60
a)
40
a) -)
a) 20

0
Nov Jan Mar May Jul Sep
-A. totai -x-A. phalerata -u- C. australis


b)


U 100
a)
4- 80
0
a) 60
0)40
CD
CU 408 20 a ) 0
Nov Jan Mar May Jul Sep
-A. totai -x-A. phalerata -+-C. australis






Figure 6-1. a) Monthly percentages of palm trees with fruits; b) monthly percentages of palm trees with fruits on the ground, Pantanal Study Area (Caiman Ranch), 1991-1992.






81


a)


Jan Mar May


Jul Sep


- C. pachystachia -x- G. americana
-4- E. contorticiliquum + P. kennedyanum


b)


Nov Jan Mar May


Jul Sep


- C. pachystachia -x- G. americana
-- E. contorticiliquum -A- P. kennedyanum


Figure 6-2. a) Monthly percentages of trees with fruits; b) monthly percentage of trees with fruits on the ground, Pantanal Study Area (Caiman Ranch), 19911992.


CO (D (D
4-
0 a) 0)
C


100 80 60
40 20
0


Nov


Ch)
a)
0
a) CY)
C
a) L..
a)
a.


100 80 60
40 20
0


,A /*







82


Table 6-5. Index of fruit availability for tree species, Pantanal Study Area (Caiman Ranch), 1991-1992.


Spp/month
November December January February March April May June July
August September October Total


Cpa
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.18 0.18


EC
1.24 0.00 0.03 0.00 0.00 0.00 0.00 0.00 0.82 0.39 12.33 16.79 31.6


GAc
0.04 0.00 0.00 0.01 0.00 0.00 0.00 0.01 0.01 0.00
0.04 0.00 0.11


PKd
1.06 0.11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.17


Ape 0.25 0.00
0.34 1.52 0.00 0.26 0.00 0.05 0.00 0.00 0.00
0.02 2.44


CA 0.00 0.00 0.36
2.40 15.17 19.65 0.00 0.00 0.00 0.00 0.00 0.00 37.58


ATg
1.70 0.00 0.26 0.25 0.08
0.04 0.00 0.07 0.11
0.02 0.23 0.52 3.28


Total
4.29 0.11 0.99
4.18 15.25 19.95 0.00 0.13
0.94 0.41 12.60 17.51


a Cecropia pachystachia b Enterolobium contortisiliquum Genipa americana d Psidium kennedyanum Attalea phalerata Copernicia australis
* Acrocomia totai.


Table 6-6. Habitat selection of brown brocket deer in the Pantanal Study Area (Caiman Ranch), 1991-1992.

Habitat
Season x2 P value MDb SGc SCd FEe
Rainy 31.61 <0.0001 Avoided Avoided Selected NS Flood 47.76 <0.0001 Avoided Avoided Selected NS
Dry 26.04 <0.0001 Avoided NS NS Selected
'Avoided = habitat used less than expected based on its availability. Selected = habitat used more than expected based on its availability. NS = no selection, habitat used in proportion to its availability (Bonferroni tests, P = 0.05).
b Moist Depression. Short Grass. d Scrub.
e Forest Edge.






83

(Table 6-5). A large production of dry fruits of E. contortisiliquum and lignified fruits of C. australis occurred at the onset of the rains (September/October), and in the flood season (March/April), respectively, but there was no evidence that brown brocket deer consumed these fruits. Fruits of A. phalerata were important during the flood, when a number of species of birds and mammals, such as parrots (Amazona aestiva), macaws (Ara chloropterus and Anodorhynchus hyacinthinus), coatis (Nasua nasua), howler (Ailouata caraya) and capuchin (Cebus appela) monkeys, foxes (Cerdocyon thous), armadillos (Dasypus spp. and Euphractus sexcintus), feral pigs, and peccaries (Tayassu pecari and Tayassu tajacu) fed abundantly on these fruits. Habitat Use

Brown brocket deer avoided open vegetation, such as Short Grass and Moist Depression, in all seasons (Table 6-6). They sought protection by concealment in tree islands and thickets for rumination and sleeping. However, brown brocket deer selected Scrub in wet seasons, and Forest Edge in dry seasons for foraging activities (Bonferroni, P = 0.05). During the rainy season, 51 % of 170 brown brocket deer sightings were in Scrub in contrast to 17% in Short Grass and 9% in Moist Depression. During the flood, 58% of 165 sightings were in Scrub in comparison to 17% in Short Grass and 5% in Moist Depression. In the dry season, 355 sightings indicated that brown brocket deer decreased their use of Scrub (41 %), and increased their use of Forest Edge (24%) during those months.







84

Discussion

Browse and forbs dominated the diet of brown brocket deer in the

Pantanal. The 3 assumptions tested in this study were not consistent with the hypothesis that brown brocket deer is essentially frugivorous in the Pantanal: 1) fruits were not abundantly available throughout the year; 2) less than 20% of the diet consisted of fruits, even during the peak of fruiting; and 3) Forest Edge was not selected over the other plant formations during the rainy season, when fleshy fruits were most available. Furthermore, brown brocket deer in the Pantanal Study Area were never observed eating dry fruits as they did in Paraguay (Stallings, 1984).

The diet of brown brocket deer in the Pantanal was compatible with what should be predicted by the digestive capacity of their gut (Parra, 1978; Demment & Van Soest, 1985). Because gut capacity decreases linearly with body size, but metabolic rate is proportional to the three-fourth power of body weight, smaller herbivores require relatively higher energy per unit of body weight to maintain their daily metabolic needs (Parra, 1978). This means that smaller herbivores need higher rates of turnover fermentation compared to large herbivores, to be able to produce enough volatile fatty acids used to generate energy. This higher rate of turnover can be obtained when the animal selects high-quality foods, that is, plant parts with low lignin content and a high ratio of cell content to cell-wall. These conditions are met by the ingestion of young browse and forb leaves,






85

which have a smaller percentage of cell-wall compared to grasses, and a smaller percentage of lignin than seeds (Demment & Van Soest, 1985).

One condition for brown brocket deer to be frugivorous was that fruits occurred in abundance, and were reliably available throughout the year. However, most commonly, only a few trees, all species combined, would have ripe fruits available on the ground at any specific time in the Pantanal Study Area. Furthermore, only A. totai, G. americana, and few other rainy season fleshy fruits, such as V. cymosa, were consumed by brown brocket deer during the rainy season. The lignified and small fruits of C. australis, which were never observed being consumed by any animal, presumably evolved to be dispersed by water during the flood, because they have no soft mesocarp to attract mammals. The other palm fruit, A. phalerata, is too large to be swallowed by brocket deer, and has a thick pericarp that needs to be handled and peeled-off (monkeys, macaws, coatis, etc.) or crunched by powerful jaws (peccaries and tapir) before the pulp can be eaten. Guanacaste tree fruits (E. contorticiliquum) are large, covered with a hard exocarp, and rich in saponin, and, therefore, are probably consumed only by large ungulates, such as tapir and cattle, or rodents (Rodentia). Cecropia fruits are probably consumed on the trees by birds and frugivorous bats, because very few were found on the ground. Finally, the characteristics of A. totai and G. Americana fruits indicate that they evolved to be dispersed by monkeys, large birds, or extinct members of the megafauna, but not by current terrestrial mammals (Abrahamson, 1989). These fruits remain on






86

the trees for long periods, and sometimes decay while still attached to the branches (pers. observ.). The few ones that fall to the ground are, in its most part, those discarded by frugivorous or seed-predator birds, because most of the forest patches in the floodplain are too small to maintain monkeys. Therefore, the Pantanal seems to do not offer the necessary conditions for exclusive frugivory for a herbivore of the size of the brown brocket deer.

Previous food-habits studies on brown brocket deer, based on the

percentage of fruit remains encountered in their digestive system, suggested that the species prefers fruits and palm seeds (Stallings, 1984; Branan et al. 1985; Bodmer, 1989; 1990, 1991; Bisbal 1994). However, the results from the diet analyses in this study indicated that fruits/seeds did not constitute an important item year-round in the Pantanal. Although medium-sized or large seeds of fruits are compact food packets, the time spent in digesting their covering, and possibly detoxifying the seeds, may not be cost-effective in their foraging time/energy budgets, when other foods are readily available in the savanna (MacArthur, 1972; Pyke et aL., 1977; Demment & Van Soest, 1985). Seeds may contain a number of palatability inhibitors, toxins, and indigestible matter to prevent seed predation (Abrahamson, 1989). In contrast, the pulp of some fruits may contain high percentages of lipids, carbohydrates, and mineral nutrients (Abrahamson, 1989), which might explain why cervids in the Pantanal Study Area were observed eating the pulp of fruits, but discarding large seeds.






87

Frugivory in brown brocket deer seems to be only a contingency or

adaptation to the environment, rather than a characteristic of the species. The fact that the sub-species of brown brocket deer studied by Stallings (1984) in the Paraguayan Chaco, which ate large proportions of fruits, is the same as in the neighbouring Pantanal, supports this hypothesis. In fact, brown brocket deer seem to be as well adapted to a diet of browse and forbs as to a diet of fruits. Brown brocket deer in the savanna seem to have similar size and densities as their counterparts in the forests (Stallings, 1986; Bodmer, 1989; Leeuwenber & Resende, 1994; Townsend, 1996; this study).

Unfortunately, a direct comparison between the results of this study and other studies is difficult because different methodologies were employed. Neither microhistological analysis nor stomach content research can accurately measure the percentage amount of fruits ingested. For example, it could be argued that because seeds are less digestible than flowers and leaves, the contribution of fruits in the diet could be overestimated by simply comparing the volume or weight of fruits with other items in the rumen or stomach. In the Peruvian Amazon, for example, palm fruits were ingested entirely (Bodmer, 1989; 1991). On the other hand, in this study the large seeds of palm fruits and V. cymosa were discarded in the mastication process. Therefore, if the pulp of fruits is digested quickly, then the percentage contribution of fruits to the diet could be underestimated by the fecal analysis method. Direct observations of feeding by brown brocket deer in the Pantanal Study Area suggested that fecal analyses






88

underestimated the contribution of fruits to the diet, but demonstrated that fruits were not the principal food source for the diet.

Furthermore, if fruits were a selected item in the diet, then Forest Edge should be the selected habitat during the period of fruit abundance. However, during the rainy season, Scrub was selected over Forest Edge, which was used only in proportion to its availability. Forest Edge was the selected plant formation only during the dry season, when fruit production was low.

Observations on habitat use by brown brocket deer in the Pantanal

paralleled the findings of Leeuwenberg and Resende (1994) for the Cerrado, where most of animal locations were obtained in the Scrub. The study conducted by Schaller (1983), in a different area of the Pantanal, also indicated the avoidance of dense forests by brown brocket deer. Most of Schaller's sightings were at forest edge, and secondarily, in what he called Cerrado Forest, the equivalent to Scrub in this study.

Brown brocket deer required shelter, such as forest patches, bushes or tall grass, for rumination or resting in the Pantanal, although they selected to forage outside or at the edge of the forest patches. Brown brocket deer are solitary animals that maintain concealment from predators by using vegetation as a visual barrier, and as a hiding place when being pursued (this study). Thus, two factors seem to contribute to the use of intermediary habitats between the forest and open habitats by brown brocket deer: food choice and predator avoidance. Niche separation with pampas deer in the grasslands, and red






89

brocket deer (Mazama americana) in the deciduous and gallery forests (Schaller, 1983), could also explain such habitat preferences.

In conclusion, brown brocket deer in the Pantanal are not frugivorous. The comparison between the studies in the Amazon and the present study suggested that the brown brocket deer is a generalist with respect to feeding habits. The hypothesis is that brown brocket deer adapts to a diet according to the characteristics of the ecosystem. Their digestive system enables them to survive on a browse diet or a frugivorous diet (Bodmer, 1989) depending on the type of habitat occupied within their large geographic distribution (Czernay, 1987; Pinder, 1997).

The greater availability of browse and forbs in the Pantanal, in contrast to more heavily forested areas, allows the brown brocket deer a large diversity of food items from which to select their diet. The only viable option for them in forested areas may be to increase the percentage of fruits in the diet, because browse is presumably sparse and rich in secondary compounds (Golley, 1978; Leigh & Windsor, 1982). The xeric leaves and stems of the shrubs in the Paraguayan Chaco may not be sufficiently digestible to allow a balanced diet without the support of protein from seeds. Stallings (1984) found that brown brocket deer even consumed dry fruits during the dry season in the Chaco. The digestive system of brocket deer species allows them to exploit fruits, as pointed-out by Bodmer (1989), and, therefore, permits them to occupy habitats where exclusive browsers or grazers would be excluded.




Full Text

PAGE 1

NICHE OVERLAP AMONG BROWN BROCKET DEER, PAMPAS DEER, AND CATTLE IN THE PANTANAL OF BRAZIL BY LAURENZ PINDER 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 1997

PAGE 2

This dissertation is dedicated to my parents, Angelina Di Giuseppe Pinder and Laurenz Heinrich Julius Pinder; to my wife, Euzelita Almeida Sousa Pinder; and to my grandmother, Olivia, who is very much alive in the heart of all of our family.

PAGE 3

ACKNOWLEDGMENTS I am very grateful to the institutions that have supported this study, and the many people v^o have encouraged me throughout the different phases of my doctoral study program. This research could not have been possible without the financial support of Conservation International, National Geographic Society, Tropical Conservation and Development Program (University of Florida), Wildlife Conservation Society, and World Wildlife Fund-US. The first 2.5 years of the study at the University of Florida were sponsored by the Conselho Nacional de Ciencia e Tecnologia (CNPq) of the Brazilian Ministry for Education. Authorization to capture and handle pampas deer and brocket deer in the field were provided by the Institute Brasileiro do Meio Ambiente (Agriculture Ministry). I especially thank Mr. Roberto Klabin, owner of Caiman Ranch, for allowing the research to be conducted on his lands and for providing logistical support during the field work. In the Pantanal, I owe my gratitude to the people who contributed in one v\/ay or another to the data collection. Seu Celestino, a former professional hunter of big cats and hunting guide, provided valuable information and showed me the farthest corners of the ranch on horseback. His son, Cordeiro, and another cowboy kindly helped me in several occasions in the capture of brown iii

PAGE 4

brocket deer. Mauricio, a botany student, helped with the phytosociological and habitat survey data cxjilection. In special, I would like to thank my wife Nininha for her companionship during the fieldwork, and later for helping in the preparation of the microhistological reference collection and in the drawings of the several histologic structures. The immense difficulties in conducting a microhistologic analysis were overcome through the assistance offered by several key people. Drs. Arnildo Pott and Valle Pott of the Empresa Brasileira de Pesquisa Agropecuaria (Ministry of Agriculture, Brazil) identified the plants collected in the floodplain. Dr. Jane Kraus facilitated the use of the plant anatomy laboratories at the Universidade de Sao Paulo, and contributed strong technical support for the preparation of the microhistological reference collection. Statistical analyses of the large data-base generated in the field were facilitated greatly by the advice of Dr. Sergio Rosso of the Department of Ecology at the Universidade de Sao Paulo, and Jay Harrison at the University of Florida. I am grateful to my supervisory committee for their advice, understanding, and support. I thank Dr. Ronald F. Labisky (Chair) for his extreme patience in revising and editing earlier versions of this dissertation, and for inspiring discussions. I am grateful to Dr. John F. Eisenberg for his words of encouragement when needed most. I am indebted to Dr. George Tanner for helpful comments on this dissertation. I thank Dr. Richard Bodmer for valuable discussions and comments that improved the quality of this dissertation. Dr. Kent iv

PAGE 5

H. Redford offered great support during the preparation of the proposal for the dissertation and early field work. Finally, I thank Dr. David Webb for stimulating conversation on the evolution of the cervids in South America. I wish to thank my parents and great friends, Laurenz Heinrich and Angelina, for their love and support throughout all the difficulties I have faced. I am also deeply grateful to my aunts, Marie Anne and Ingeborg, who always have given me incentive and support when most needed. Nor could I forget my Aunt Marlene, who constantly had a word of advice and comfort. I would like to thank my great friend Sofia, and her many friends, for pointing me in the best direction. Finally, I am extremely grateful to all those people who, although not mentioned, have demonstrated faith in me and in my work. V

PAGE 6

TABLE OF CONTENTS ACKNOWLEDGMENTS jjj ABSTRACT x CHAPTERS 1 INTRODUCTION 1 Background 1 General Objectives of the Study 5 Organization of the Dissertation 6 2 THEPANTANAL 7 Introduction 7 Location 7 Geomorphology 7 Climate 9 Hydrology 9 Vegetation 10 Plant Phenology 12 Fauna 13 Economy and Conservation 15 3 STUDY AREA 19 Rationale 19 Location and Climate 19 Flood Regime 20 Habitats 22 Land Use 23 4 CLASSIFICATION AND ORDINATION OF THE PLANT ORMATIONS IN THE FLOODPLAIN 25 Introduction 25 Objectives 25 vi

PAGE 7

Methods 26 Rationale and Sampling Methodology 26 Classification and Ordination of Plant Formations 29 Richness and Diversity of Plant Formations 30 Results 30 Cumulative Number of Species 30 Species Associations 31 Classification of Plant Formations 33 Ordination of Plant Formations 38 Richness and Diversity of Plant Formations 38 Characteristics of Plant Formations 41 Marsh Pond 41 Moist Basin 44 Short Grass 44 Tall Grass 46 Scrub 46 Forest Edge 46 Discussion 47 5 CATTLE DIET AND HABITAT SELECTION 50 Introduction 50 Objectives 51 Methods 51 Fecal Sampling and Analysis 52 Reference Collection 53 Plant Species Availability 54 Habitat Use 54 Results 56 Diet Composition 56 Habitat Use 58 Seasonal Diet Diversity and Niche Breadth 58 Discussion 60 6 BROWN BROCKET DEER DIET AND HABITAT USE 67 Introduction 67 Objectives 69 Methods 69 Fecal Sampling and Direct Observations 69 Microhistologic Analyses 70 Phenology of Trees 72 Habitat Use 73 Results 74 Diet Composition 74 Availability of Fruits and Flowers 79 vii

PAGE 8

Habitat Use 83 Discussion 84 7 PAMPAS DEER DIET AND HABITAT USE 91 Introduction 91 Objectives 94 Methods 94 Fecal Sampling and Analyses 94 Habitat Use 96 Results 96 Diet Composition 96 Habitat Use 103 Discussion 105 Fecal Analysis vs. Direct Observations 105 Pampas Deer Diet in Relation to the Habitat Characteristics 106 8 ECOLOGICAL SEPARATION AMONG CATTLE, PAMPAS AND BROWN BROCKET DEER 114 Introduction 114 Methods 117 Density Estimates 117 Similarity and Niche Overlap Indices 117 Plant Species Availability and Selection 118 Results 119 Density and Biomass of Cattle and Cervids 119 Dietary Similarities 121 Habitat Use 124 Niche Overlap and Partitioning 127 Cattle and Deer Responses to Environmental Changes 133 Discussion 136 Resource Partitioning 136 Classification of the Cervid Feeding Strategies 140 Conservation Implications 141 9 SUMMARY AND CONCLUSIONS 145 APPENDICES A LIST OF COMMON PLANT SPECIES IN THE FLOODPLAIN 152 B UNGULATE DIET COMPOSITION 160 viii

PAGE 9

C ANTLER CYCLE AND REPRODUCTION 169 Brown Brocket Deer and Pampas Deer Antler Cycles 169 Reproduction 169 D HOME RANGE SIZE 177 Brown Brocket Deer Home Range 1 77 Pampas Deer Home Range 178 E BONFERRONI TESTS FOR HABITAT USE AND AVAILABILITY 183 F KRUSKAL-WALLIS TESTS FOR DIET COMPOSITION 187 G STP AND SPEARMAN RANK CORRELATION TESTS 191 REFERENCES 202 BIOGRAPHICAL SKETCH 220 ix

PAGE 10

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Philosophy NICHE OVERLAP AMONG BROWN BROCKET DEER, PAMPAS DEER, AND CATTLE IN THE PANTANAL OF BRAZIL By Laurenz Pinder December, 1997 Chairman: Dr. Ronald F. Labisky Major Department: Wildlife Ecology and Conservation Niche partitioning among brown brocket deer {Mazama gouazoubira gouazoubira), pampas deer {Ozotoceros bezoarticus leucogaster), and cattle was studied in the floodplains of the Pantanal (19° 57' S, 56° 25' W), Brazil. The climate was seasonal, with wet, flood, and dry seasons. The vegetation in the study area consisted of six plant formations that were determined through the classification and ordination of random sampling plots: Marsh Ponds, Moist Basins, Short Grass, Tall Grass, Scrub, and Forest Edge. Pampas deer used all plant formations except forest patches, but used Short Grass most frequently. Moist Basin was an important source of food for pampas deer during the dry season, when 44% of their diet consisted of a forb {Melochia simplex), which grov^ in this habitat. The progressive reduction of X

PAGE 11

dietary diversity and niche breadth from rainy to dry seasons suggested that availability of new-growth was important for pampas deer. Graminoids, forbs, and browse dominated their diet during the rainy season; forbs and browse were prominent during the flood season; and forbs alone constituted 79% of the diet in the dry season. Brown brocket deer preferred Scrub and Forest Edge habitats, and were affected little by seasonal changes in food availability. Brown brocket deer were mostly browsers throughout the year and consumed the greatest proportion of forbs (55%) in the dry season, when they foraged more frequently in Moist Basin and Short Grass, as compared to other seasons. Brown brocket deer managed to maintain dietary diversity and niche breadth throughout the year by increasing their use of different types of habitat when food was limited in the selected habitat. Neither fruits and flowers of trees contributed significantly to the diet of brown brocket deer although they were consumed in the rainy season and in the dry season respectively. Eighty percent of the cattle's diet consisted of graminoids, reducing probabilities of competition between cattle and cervids. Niche partitioning among the three ungulates was achieved principally by the selection of different plant species and by the type of plant formation selected. xi

PAGE 12

CHAPTER 1 INTRODUCTION Background Wet areas are important to global biodiversity preservation besides providing substantial resources of food and water for people (UNEP, 1991/1992). The Pantanal, the world's largest wetland, constitutes a substantial component of South American biodiversity which has floristic and faunal elements from the Amazonian, Cerrado, and Chaco biogeographical regions (Adamoli, 1984; Pott, 1988). The region has been relatively free from effects caused by economic activities, such as skin trade, animal traffic, gold mining, deforestation, and cattle production until recent years. Currently, the decline of large ranches, and a future waterway conversion of the Paraguay River, are threatening the ecological balance in the region (Gomes, 1997). As a result, several faunal species are in danger of a future extinction in the wild (Alho et al., 1988). Additionally, there is social pressure for land utilization and for optimization of current agricultural methods; therefore, large mammals, as a group, will become difficult to conserve. Collectively, the vastness of the region, low density of people, and large ranches (hundreds of square kilometers each) ensured the survival of the entire fauna in natural condition until the1970s. 1

PAGE 13

2 However, South American large mammals occur in low numbers naturally (Eisenberg, 1989; Redford & Eisenberg, 1992), which potentially exposes them to greater chances of local extinction when subjected to environmental changes or excessive hunting pressure (Redford, 1997). Therefore, conservation of local fauna will depend either on large reserves containing several habitat types, which do not exist currently, or on adequate land management practices that are compatible with the local ecological equilibrium. However, the creation of large, state-protected areas currently is politically difficult because of the lack of support by landowners and because thousands of landless people are pressing the government for appropriation of "sterile" areas for agricultural use. The second alternative, the integration between land utilization and conservation, is a potentially valid option. Except for animal traders and few caiman breeders, the pecuniary value of the fauna perceived by most of the locals is low. And, few cattle ranchers have realized the potential income associated with ecotourism. However, there is no substantial information on the specific requirements of native large mammals. The effects that cattle ranching, and its modern practices, impose on wildlife are poorly understood. Because ecotourism represents only a small source of additional income, cattle ranchers may abandon interest in participating in the ecotourism industry if local fauna begins to decrease. The long-term goal of managing wild fauna and their habitats will depend ultimately on scientific studies of resource use among the

PAGE 14

3 cattle and native large herbivores although the consortium between cattle and wildlife is currently feasible. South American's large herbivore guild is poor in contrast to Africa and Asia. This guild is usually represented by cervids, a rodent (capybara: {Hydrochaeris hydrochaeris), and a tapir {Tapirus terrestris). Common species inhabiting the Pantanal floodplain are marsh deer {Blastocerus dichotomus), pampas deer {Ozotoceros bezoarticus), brown brocket deer {Mazama gouazoubira), and capybara. The tapir and red brocket deer {Mazama americana) are rare. These two latter species are more commonly found in the semi-deciduous forests bordering the floodplain. Currently, only a few ecological studies of the cervids have been conducted. Capybaras have been studied more extensively due to their economical importance in some areas (Escobar & Jimenez, 1976; Jorgenson, 1986; Alho etal. 1987a,b; Herrera & MacDonald, 1989). Published studies on cervids include aspects of feeding habits and habitat use of pampas deer by Bianchini and Perez (1972), Jackson and Giulietti (1988), Heinonen etal, (1989), and Merino (1993) in Argentina; by Jackson and Langguth (1987) in Uruguay; and by Resende and Leeuwenberg (1992) in the Brazilian Cerrado. And, despite a larger geographic distribution, brown brocket deer have been studied in only a few localities. Staliings (1984, 1986) studied feeding habits and reproduction of the brown brocket deer in the Paraguayan Chaco, and Bodmer (1989) and Resende and Leuwenberg (1992) investigated

PAGE 15

4 habitat use and diet in the Peruvian Amazon forest and in the Brazilian Cerrado, respectively. Habitat use and diet of the marsh deer, the least studied of these three species, has been researched by Schaller (1976), Beccaceci (1994), Beccaceci and Merino (1994) in Argentina, and by Voss etal. (1981), Mauro (1993), and Pinder (1994, 1996) in Brazil. No investigation has been conducted on niche partitioning among the three cervids, or on competition between cervids and livestock in the Pantanal. No competition between capybaras and cattle, or other native ungulates, has been documented in the South American savannas although this rodent constitutes a large biomass in the Pantanal (Schaller, 1983; Lourival & Fonseca, 1997). Capybaras are semi-aquatic grazers; thus, most of their forage is found in or near water bodies (Escobar & Gonzalez Jimenez, 1976; De Azcarate, 1981 ; MacDonald, 1981; Schaller & Crawshaw, 1981; Jorgenson, 1986; Herrera & MacDonaid, 1989, Quintana et a/., 1994). Cattle, as well as pampas and brown brocket deer in contrast to capybaras, avoid areas permanently flooded (Bodmer, 1989; Leeuwenberg & Resende, 1994). Furthermore, Pon etal. (1986) have observed that in the Brazilian Pantanal, the capybaras consumed medium to high proportions of sedges (Cyperaceae) Instead of grasses, which are preferred by cattle. Similarly, marsh deer would have little niche overlap with cattle or other cervids due to their preference for permanently flooded areas (Schaller, 1976; Mauro, 1993; Beccaceci, 1994; Pinder, 1994).

PAGE 16

5 This dissertation presents the results from research conducted in the Pantanal of Brazil between January 1990 and December 1992. The research consisted of two phases: a classification of the habitats existing in the seasonally flooded plains in the study area; and a documentation of habitat preferences and diet of pampas deer, brown brocket deer, and cattle. The first phase was Important as it provided data of habitat and plant species availability that were fundamental for the second phase. General Objectives of the Study This study sought to answer questions that related indirectly to the consequences of introducing livestock into the habitat of two species of deer. My major goal was to obtain the base-line information needed to conserve deer populations on the private lands of the Pantanal although South American deer have been little studied, and, hence, are of considerable theoretical scientific interest. A secondary goal was to explore theoretical questions related to the classification of feeding habits of South American deer (Bodmer, 1989). Pampas deer occupying the Argentinean grasslands have been classified as grazers (Jackson & Giuletti, 1988), whereas brocket deer occupying tropical forests and the Paraguayan Chaco have been classified as frugivores (Bodmer, 1989). The Pantanal, with its mosaic of grasslands, scrub, and forest patches, provided an ecosystem to test these previous classifications. To address these goals the following questions were asked:

PAGE 17

6 1 . Which plant formations were available to the three large herbivores (brown brocket deer, pampas deer, and cattle) in the seasonal floodplain of the Pantanal? 2. What constituted the diet of the brown brocket deer, cattle, and pampas deer during the three major phonological seasons in the region: rainy season, flood season, and dry season? 3. Was there significant dietary and habitat overlap among the three herbivores? 4. Did the dietary classification of grazer and frugivore hold for pampas and brocket deer in a habitat where there were multiple habitats available? 5. How did the answers to these questions contribute to the future management and conservation of these wild ungulates in the Pantanal? Organization of the Dissertation Chapter 2 describes the climate, hydrography, vegetation, fauna, and economy of the Pantanal. Chapter 3 characterizes the study site. Chapter 4 presents the results of classification and ordination of the different plant formations observed in the seasonal floodplain. Chapter 5 analyzes the feeding habits of cattle, foraging value of each of the major plant species consumed, and foraging value of the habitats available to cattle. Chapters 6 and 7, respectively, describe brown brocket deer and pampas deer habitat preferences and their diets, and discusses their dietary classification. Chapter 8 discusses the niche overlap among the three ungulates.

PAGE 18

CHAPTER 2 THE PANTANAL Introduction Pantanal means a large continuous swamp in Portuguese. It is considered one of the largest interior wetlands in the world (Silvestre Filho & Romeu, 1 974; IBGE, 1977; Rizzini, 1979). However, most of the area is not permanently inundated, but rather is flooded only seasonally. The elevation of the Pantanal ranges from 83 to 165 m above mean sea level. In contrast, the surrounding plateaus of the Brazilian Shield exceed 600 m elevation (Silvestre Filho & Romeu, 1974; Rizzini, 1979). The landscape of the Pantanal remains little modified by human activities although completely occupied by ranches (Por, 1995). Location The Pantanal is located in southwestern South America, between parallels 16° and 22° S, and 55° and 58° W (Fig. 2-1 ). It extends approximately 770 km North-South and 250 km East-West within Brazilian borders, occupying an area of 160,000 km^ of seasonally flooded plains. GeomorpholoQv The Pantanal floodplain consists of Pleistocene alluvial sediments: sand and silt. One of these two sediment types dominate the floodplain, depending on region. 7

PAGE 19

Figure 2-1. Location of the Pantanal showing main rivers (thin lines) and limits of the floodplain (thick line).

PAGE 20

In general, sandy soils are poor in nutrients and silty soils are fertile. There is only local redistribution (vertical and horizontal) of nutrients in areas not influenced by alluvial nutrients. Most of the nutrients of the Pantanal's forest patches are in the arboreal biomass rather than in the soil, similarly to what happens in the Amazon forest. There is some degree of nutrient loss in areas of open drainage, which is accentuated by the annual burning promoted by ranchers at the beginning of the rainy season (Pott, 1988). Climate The climate is tropical hot and humid with 3 months of dry season (Jun.Aug.), and 3 months of rainy season (Nov.-Jan.). Annual rainfall ranges from 1000 to 1 500 mm, with a mean annual temperature of 24° C. More than 80% of the annual rainfall occurs between November and March, which causes flooding between January and April. Absolute temperatures range from 0° dunng the winter to 42° C in the spring (IBGE, 1977, Silvestre Filho & Romeu, 1974). Hydrology The entire region is drained through the Paraguay River, which, with the Parana River, constitutes the second largest riverine basin in South America after the Amazon. The drainage basin of the Pantanal is fan-shaped with the center in the Paraguay River on the Bolivian border, and the wings toward the east (Fig. 2-1). There are permanent and temporary streams between the main tributaries, locally called "corixos" and "vazantes", that link a network of lakes or "baias". River waters

PAGE 21

10 flow slowly and form countless ponds In some areas because there is little relief in the region (2-5 cm/km N-S, and 30-50 cm/km E-W) (Brasil, 1974; Pott, 1988). Vegetation The flora of the Pantanal, which contains approximately 1500 plant species, has not been studied extensively. Due to the large areal size of the Pantanal, its flora is influenced by different, adjacent biomes (Pott, 1988). The gallery forests of the Pantanal are influenced by the Amazonian tropical rainforest in the north (Eiten, 1985). The Chacoan vegetation dominates the landscape In the south. The largest influence however, consists of species characteristic of the Brazilian savanna, the Cerrado (SEMATEC, 1990), which originally covered the interior of the country below the Amazon forest (Pott, 1988). Many of the plant species in the Pantanal are of general geographic distribution. Some are present in other South American floodplains, such as the Colombian Llanos, Amazonian Marajo Island, Bananal Island, and the Paraguayan Humedales. Others are common to the dry forests of northeastern Brazil and Argentina (Pott, 1988). General descriptions of the Pantanal's vegetation types have been provided by Veloso (1948), Prance & Schaller (1982), Eiten (1985), and Adamoli (1986). Prance and Schaller (1982) provided information on phytosociology and phenological patterns for forest habitats, but similar studies on grassland habitats are lacking.

PAGE 22

11 Small differences in relief (1-4 m) influence plant associations due to seasonal flooding. Areas above the flood line are covered by different forest vegetation types: cerrado (saub savanna), cerradao (scrub forest), and semideciduous forest (Eiten, 1972). The arboreal cover of a given area is dependent on the miCTO-climate and the type of soil. For example, Cerrado vegetation is found on sandy soils, whereas Chaco vegetation is common on clay or alkaline soils (Pott, 1988). Finally, there is the xeric vegetation type (steppe savanna), which is characteristic of slopes of dry and calcareous soils. According to Veloso (1948), seasonally flooded areas can be divided into three categories: the Aquatic Zone, the Hygrophile Zone, and the Mesophile Zone. The Aquatic Zone is characterized by floating and rooted aquatic species (e.g., Eichhornia crassipes, Salvia spp., and Pontederia spp ). The Hygrophile Zone is sub-divided into two categories: 1 ) Vazantes or shallow drainage channels, characterized by aquatic plants, sedges, grasses, and herbs growing on seasonally flooded ground, which do not completely dry during the winter (dry season); and 2) periodically flooded terrains bordering the rivers, which are dominated by woody plants and trees. The Mesophile Zone occupies higher terrain, and is characterized by species adapted to aquatic and terrestrial environments. Woody plants occupy islands of non-flooded soils, grasses and sedges occupy soils subject to moderate flooding, and aquatic and semi-aquatic plants occupy depressions that remain moist year-round. In areas where the flooding occurs for only a few days during the peak

PAGE 23

12 of the flood season, a saub savanna dominates the plains (Prance and Schaller, 1982, Ratter ef a/., 1989). The Mesophile Zone cx)nstitutes the native pasture of the Pantanal. The landscape is punctuated by circular forest patches or "capoes", and narrow and parallel strips of forest or "cordilheiras" never reached by water. Invading woody shrubs may become abundant on the grassland during dry years and in areas overgrazed by cattle (Pott, 1988). On the other hand, the aerial parts of grasses may die and be replaced by aquatic and semi-aquatic plants when the flooding is prolonged for several months. Water levels affect the abundance and distribution of forest patches. Thus, "cordilheiras" and "capoes" are more abundant on higher elevations close to the borders of the Pantanal (Brasil, 1982). In general, "capoes" are less than 20 m wide, and "cordilheiras" are 50-150 m wide and 2 km in length. These two types of forest patches stand just above the water line during the flood season, and are important refuges for wildlife and cattle. Trees have superficial roots because of the shallow water table (0.2-2.5 m), and are characteristic of the forests bordering the floodplain (Pott, 1988). Plant Phenoloov Due to the marked seasonality of rainfall, many of the arboreal species lose their leaves during the dry season (June-August). The peak of defoliation occurs in September, before the onset of the rainy season. This pattern is more accentuated in the semi-deciduous forests of the slopes bordering the floodplain (Schaller, 1983).

PAGE 24

13 Unlike leaf-fall, the fruit production and flowering of trees are not synchronized. Schaller (1983) registered peak flowering of Caesalpinia sp. in March, Pseudobombax sp. in May-June, Bowdichia sp. and Pouteria sp. in July-August, Tabebuia caraiba in August, Magonia sp. and Simarouba versicolor in September, and Acosmium sp. in October. Curatella amehcana and Tabebuia impetiginosa may have more than one flowering peak between May and September, an event probably related to variations in the water table. Schaller (1 983) also reported that fruits were available in any month. Herbs and vines bloom from April to June, after the flood waters recede (Schaller, 1983). Seeds and fruits are more abundant in semi-deciduous forest than in the floodplain. Schaller (1 983) found that 52% of trees bore fruits at some time between March and October in the semi-deciduous forest of the slopes, whereas only 20% of trees in the floodplain bore fruits during the same period. Fauna The influence of different biomes and a great variety of habitats allow for a large faunal diversity, especially waterbirds, within the Pantanal. Psittacidae are very diverse, including five species of macaws and 1 3 species of parrots (Alho et al., 1988). The Pantanal lacks large raptors, with the two largest species being the great black hawk {Buteogallus urubitinga) and the great homed owl {Bubo virginianus). The king vulture {Sarcoramphus papa) and three other species of vultures form the feathered scavengers guild of the Pantanal. Besides the local

PAGE 25

14 avifauna, the Pantanal hosts migrating birds of three major migratory routes: Central Brazil, Rio Negro, and Cis-Andean (Sick, 1983). The largest reptilian predators include the Pantanal caiman {Caiman crocodilus yacare), which may weigh 1 1 0 Kg, and the anaconda {Eunectes murinus), the world's largest snake (Almeida, 1976). Other snakes are rare in the floodplain presumably because the combined effects of flooding and grass fires during the dry season maintain the area voided of their prey. Many large mammals still abound, including a few vulnerable species. A survey of mammals in the Pantanal yielded 64 species of mammals, one third of which were bats (Schaller, 1983). Other families included the Marsupialia (three species). Primates (five species), Edentata (six species), Lagomorpha (one species), Rodentia (eleven species), Camivora (ten species), Perissodactyla (one species), and Artiodactyla (six species). The total biomass of these species was 380 kg/km^. Ungulates formed the largest biomass (62%) of mammals in the Pantanal. Rodents ranked second in biomass (12%), principally due to the large number of capybaras {Hydrochaeris hydrochaeris) (Schaller, 1983). Livestock had a biomass of 3,750 kg/km^ a value 10-fold greater than native mammals. The Pantanal includes most of the Neotropical genera of hoofed mammals, including tapir (Tapirus terrestris), marsh deer {Blastocerus dichotomus), pampas deer {Ozotceros bezoarticus), red brocket deer {Mazama amehcana), brown brocket deer {Mazama gouazoubira), peccaries {Tayassu pecan and T. tajacu), and feral

PAGE 26

15 hogs {Sus scrofa), the latter introduced with the first European colonizers. Capybaras complete the guild of large herbivores. Jaguars {Panthera onca) and pumas {Puma concolor) are still present along with other smaller predators such as the aab-eating fox {Cerdocyon thous), ocelot {Felis pardalis), and the rare maned wolf {Chrysocyon brachyurus), even though all of them are regularly hunted (Lourival & Fonseca, 1997). Economy and Conservation The Pantanal now supports a population of more than 340,000 people, including natives and descendants of colonizers. Cattle ranching has been the main economic activity in the region in the 200 years post-settlement. There are more than 12 million head of cattle within 13 million ha of native pastures and 550,000 ha of introduced pastures (IBGE, 1977, Alho etal., 1988). Among the several native species of grasses that occur in the Pantanal, the preferred forages are Paratheria prostata, Setaria geniculata, and Reimaroctiloa brasiliensis. Historically, ranchers moved their herds to accommodate the rise and decline of the water on the native pastures. Because of the near lack of fences, cattle ranged free within ranches, often larger than 2,000 km^. Subsequently, as the ranches were divided among descendants, once-extensive lands became smaller and smaller. Today, there are 2,000 ranches, with an average size of 13,000 ha each (Allen & Vails, 1987). Currently, ranchers have substituted the traditional cattle variety, which was adapted to flood regimes, for zebu cattle that require more productive pastures (For,

PAGE 27

16 1995). Consequently, forest patches within the Pantanal region began to be replaced by exotic, and presumably more productive, grass varieties. The introduction of new livestock to the region created yet another problem. Although cattle are regularly vaccinated, several domestic livestock diseases were initially introduced to the fauna of the Pantanal. Equine trypanosomiasis {Trypanosoma spp.) infected capybaras and peccaries. Brucellosis and foot-andmouth diseases infected wild ungulates. Periodical outbreaks of these diseases reduced wild faunal populations, and may have been responsible for the decline of marsh deer (Schaller, 1983; Schaller & Vasconcelos, 1978). Furthermore, rudimentary land management practices have caused the impoverishment of the soil and the invasion of weed species. Ranchers annually bum the native pastures before the rainy season to reduce invader plant species. However, this continuous process inaeases the dominance of less palatable grass species, such as Elyonurus muticus (IBGE, 1977). Furthermore, the annual frequency and the timing of the burning may have impacts on soil nutrient cycling, vegetation, and fauna although the ecosystem is adapted to fire (IBGE, 1977; Coutinho, 1990). Since the 1980s, cattle production has inaeased due to the deforestation of cordilheiras and forests on the slopes surrounding the floodplain. However, this deforestation has caused severe erosion, and, consequently, the silting of rivers (Alhoefa/., 1988).

PAGE 28

17 Agriculture enterprises beyond cattle production are hindered by the environmental conditions, and are practiced only at subsistence level (Allen & Vails, 1987; Por, 1995). The conversion of land to crop production has not been significant in terms of changing the landscape because the human population in the region is still low. Another economic activity in the Pantanal, which has environmental implications, is illegal professional hunting and the associated smuggling of animal products, in the early 20'^ century, skins and other products were harvested legally and shipped from Brazil to the markets in Uruguay, from which they were exported to Europe (Miller, 1930). In 1967, the Brazilian government passed a deaee outlawing professional and sport hunting; however, the lack of law enforcement allowed for the continuous hunting. In 1970 only, 15,51 1 skins, 3,140 furs, 2 tons of rhea feathers, and 1 ,903 tons of fish were taken illegally in the Pantanal. These numbers have increased steadily ever since (Alho et ai, 1 988). Most of the animals or their products are smuggled by land or small airplanes aaoss the borders of Bolivia, Paraguay, and Argentina. Forged documents are then prepared in these countries, and the products are exported to Europe and Japan. Predator control and subsistence hunting may also contribute to the local deaease of wildlife numbers. Many of the traditional ranchers keep dogs for the hunting of large cats, and other predators are killed when casually encountered by cowboys (Lourival & Fonseca, 1997). Additionally, capybara, armadillos, and all

PAGE 29

18 ungulates are hunted for food by ranch personnel year-round (Schaller, 1983; Lourival & Fonseca, 1997). The tourism industry was very incipient economically until the mid-1970s; however, tourism has gained popularity as a source of income to ranchers (Por, 1 995). The few tourists that visited the region prior to the 1 970s were in most of the cases wealthy trophy hunters or sport-fishermen (Almeida, 1976). Tourism has boomed in large cities around the Pantanal since the mid-1980s, and, accordingly, a few ranchers have tried to capitalize on tourists by building facilities to accommodate them (For, 1995). Some of these tourism operations, however, are conducted by poorly informed tour groups, and have caused disturbance to waterbird rookeries and nesting areas, in addition to exacerbating widespread littering (Alho etal., 1988).

PAGE 30

CHAPTER 3 STUDY AREA Rationale Caiman Ranch was chosen as the study site in the Pantanal for the following reasons: 1 . The site contained natural densities of brown brocket deer {Mazama gouazoubira) and pampas deer {Ozotoceros bezoarticus). 2. The brown brocket deer, pampas deer, and cattle occupy the same range of habitats within the floodplain. 3. The entire area has been protected from hunting for over 30 years, which has allowed cervids to become habituated to the presence of humans. 4. The original assemblage of fauna on the Caiman Ranch is still extant. Location and Climate The study area is a cattle ranch located between the Aquidauna and the Miranda rivers in the southeastern Pantanal. It lies 207 km west of the State capital, Campo Grande, and 36 km north of Miranda, the nearest town (19°57'S, 56°25'W) (Figure 2-1). The climate at Caiman Ranch is seasonal and characterized by a wet and dry season, which is ranked between Aw and Am using Koepen's classification (Sudo, 1974) (Fig. 3-1). Mean monthly temperatures are relatively constant 19

PAGE 31

20 throughout the year. In 1991 , the annual mean monthly temperature was 24° C, with annual mean maximum and minimum temperatures of 30° C and 17°C. The monthly mean temperature was lowest in August (19° C), and highest in February (26° C) (Fig. 3-2). The mean annual rainfall for the 20 years, 1972-1991 , was 1 ,7733.40 mm, but exhibited a declining trend (Fig. 3-3). Crawshawand Quigley (1984) reported a dry period between 1937 and 1950, when annual rainfall averaged 1,213 mm. In recent years, the highest rainfall at the ranch occurred in 1974 (2,298.70 mm), and the lowest in 1988 (1,257.40 mm). However, the v\«ather is unpredictable from year to year with regard to amount of monthly rainfall (Fig. 3-3). Forty-nine percent of the annual rainfall occurs during the 3-month rainy season (Nov. -Jan.), with January being the wettest month (x =320.4 mm). The dry season occurs from June to August, with July being the driest month (x =26.7 mm). A secondary peak of rainfall occurs in May, when sudden lower temperatures associated with the first polar air masses of the winter season reach the Pantanal (Fig. 3-1). Flood Regime Rainfall has a significant effect on the water level of lakes and streams., the lake in front of the ranch headquarters showed a variation of 0.74 m over the year in 1991, reaching its highest level in April, and its lowest in August. Crawshawand Quigley (1984) reported a difference of 1.66 m between the highest and lowest levels in 1983.

PAGE 32

21 Figure 3-1. Annual pattern of rainfall (1972-1991) and temperature (1991-1992), Pantanal Study Area (Caiman Ranch). Both rainfall and temperature statistics were collected on-site. JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Min. Max Figure 3-2. Annual mean monthly minimum and maximum temperatures (°C), Pantanal Study Area (Caiman Ranch), 1 991 .

PAGE 33

22 ^ 2500 n E 2000 1500 ^ 1000 < g 500 Z < 0 ^ 72 74 76 78 80 82 84 86 88 90 Figure 3-3. Annual rainfall (mm), Pantanal Study Area (Caiman Ranch), 1972The effects of heavy rainfall can be observed in the water level of streams within 3 days after the rain. However, the flood season begins only weeks after the onset of heavy rains (Nov.) because of the cumulative effects of the raising watertable and the river overflows. During 1991 and 1992, the flood season began in late February and the lowlands remained covered by 50 cm of water until the end of April. However, it is possible that flood conditions may extend from January to June in extremely wet years. The floodplain, which occupies 98% of the ranch, is bounded on the north by the Aquidauana River. The floodplain provides the principal native habitat for pampas deer and brown brocket deer, and pasture for cattle. The hills bordering the floodplain, which comprise 2% of the area, are located on the southern limits of the ranch. 1991. Habitats

PAGE 34

23 A vegetation map, produced at the request of the ranch owner by a private company using satellite imagery, revealed 7 habitat types. Semi-deciduous forest (1 ,071 .40 ha), dense scrub forest (4,055.78 ha), and introduced pastures (834.73 ha) are located in the borders of the floodplain. Scrub savanna (7,964.60 ha), grasslands (30,208.01 ha), shallow drainage channels (4,735.27 ha), and lakes (634.51 ha) lie in the seasonally flooded plains. The remaining area is occupied by roads and facilities. However, the satellite imagery (definition 30 m) classified only the major vegetation patterns. Thus, it was necessary to conduct a phytosociological study to discriminate differences among the mosaic of small plant formations that constitute the habitat of the species studied (Chapter 4). Land Use Development in the region is relative recent. The access to the town of Miranda until the early 1970s was mostly performed by train, boat, or plane (Silvestre Filho & Romeu, 1974). Most of the ranches, however, have their own airports because many of the owners do not live on their ranches, but rather commute to them periodically (Por, 1995). The main economic activity at the Pantanal Study Area (Caiman Ranch) always has been cattle reproduction. Additionally, in 1989, a tourist lodge was built on the old ranch headquarters to accommodate tourists. New facilities were constructed in 1993, which has approximately doubled the local lodge capacity.

PAGE 35

24 The awareness and cx)ncem with the conservation of the fauna also have evolved on the ranch. Several sportsmen, including international celebrities, hunted jaguars and other game in the ranch until the mid 1960s, when the owners decided to protect the fauna (Almeida, 1976). Recently, the Caiman Ranch's owner has stimulated scientific studies in the area to better protect and manage its natural resources. As a result, research on rare species, such as jaguar {Panthera once palustris), hyacinth macaw {Anodorhynchus hyacinthinus), and jabiru stork {Jabiru mycteria), has been conducted on the ranch (Crawshawand Quigley, 1984; Banks, 1991).

PAGE 36

CHAPTER 4 CLASSIFICATION AND ORDINATION OF THE PLANT FORMATIONS IN THE FLOODPLAIN Introduction The analysis and understanding of the habitat use by any faunal species depends on a knowledge of the environment where the animal lives. Frequently, wildlife biologists rely on previous descriptions of the vegetation types. However, in the present study, I had neither available references on the habitat selection and use by ungulates nor detailed reports of the vegetation types on the Pantanal Study Area. The local vegetation, particularly within grassland and scrub, has distinct physiognomies depending on the height of the grasses and abundance of shrubs and small trees. All of these plant formations were previously assembled by botanists and other researchers in only one or two phytophysiognomic categories (Hoehne, 1923, 1936; Veloso, 1948; Brasil, 1982; Eiten, 1985; Allen & Vails, 1987). These gross groupings were of little use for the level of definition necessary to understand habitat selection by the three ungulates. Objectives The first objective of this study was to classify and ordinate the different plant formations, based on numerical aiteria obtained through a phytosociological study. The seasonal availability of plant species could also be determined as a sub25

PAGE 37

26 product of the phytosociological study. The following hypothesis was formulated to achieve the main objective: Ho1 Visually identified floristic formations constitute numerically distinct plant associations in the floodplain. Methods Rationale and Sampling Methodolocv The study was conducted in three steps to describe the vegetation types of the floodplain: 1 ) subjective identification of the vegetation types most used by brown brocket deer {Mazama gouazoubira) and pampas deer {Ozotoceros bezoarticus), based on structure and species dominance; 2) sampling within homogeneous vegetation types described in Step 1 , with determination of species composition and cover; and 3) treatment of the phytosociological data collected in Step 2 to verify whether or not the vegetation types described in Step 1 represented significantly different plant formations. Preliminary observations on pampas deer and brown brocket deer were conducted from August (dry season) of 1989 to November (rainy season) of 1990 to accomplish Step 1 Ten distinct vegetation types were classified, based on the vegetation structure and species dominance: 1) Marsh Ponds, 2) Moist Basins, 3) Short Grass, 4) Medium Grass {Elyonurus muticus (Spr.) Kunth), 5) Medium Grass/Tall Grass {Elyonurus muticus /Schizachyrium microstachyum Desvaux), 6) Short Grass/Tall Grass, 7) Tall Grass, 8) Scrub, and 9) Forest Edge ("capao" or tree

PAGE 38

27 island), and 10) Mixed Association (a tall grass formation that had been plowed and seeded with Brachiaria humidicola (Rendle) six months before the sampling effort). Representative tracts of each of the 1 0 plant formations were selected arbitrarily for sampling with the point-quadrat method in November of 1 991 (Levy, 1927). Use of this method allowed the estimation of the cover of grasses, forbs, aquatic plants, and shrubs in each plant formation. Mantovani and Martins (1990) summarized and discussed the limitations of the method, which has been recommended by several plant ecology researchers (Crocker & Tiver, 1948; Du Reitz, 1930 in Levy & Madden, 1933; Whitmann & Siggeirson, 1954; Thomas, 1960). Twenty sampling points were taken from each of 25 1 -m^ quadrats, which were distributed within five 25-m^ quadrats. The 25-m^ sampling units containing five 1-m^ quadrats were placed randomly along a transect within each plant formation, except Forest Edge. Only 1-m^ quadrats were established along the interface between the forest and the grassland of different forest patches selected at random due to their circular shape and small size. The comers of quadrats were marked permanently with wooden poles for repeated vegetation samplings in January (peak of rainy season), March (flood season), September (dry season), and November (beginning of the rainy season). A graph of the cumulative number of species versus number of plots sampled was made to determine the total number of sampling units required for seasonal vegetation analysis. As a result, a total of 30 1-m^ quadrats

PAGE 39

28 were sampled for Forest Edge and 50 quadrats for Marsh Pond, Marsh Basin, and Mixed Association. The sampling points were determined with the aid of a metal rod, 20 cm in length x 7 mm diameter, suspended by a nylon string from a 1-m^ grate of 10-cm ^ mesh. This method allowed for unbiased sampling of the vegetation below the frame because gravity determined the exact sampling point. The string was released from the grate until it touched the vegetation, water, or soil. The sampled plants were identified immediately to the best knowledge of the observer (family, genus, or specific epithet). Specimens of all species sampled were collected in triplicate. One collection was sent for identification to the Brazilian Center for Agricultural Research of the Pantanal (CPAP/EMBRAPA); another was kept for comparisons in the field; and a third was used for developing a microhistological reference collection, which was prepared at the laboratory of plant anatomy in the Botany Department of The University of Sao Paulo (USP). A board with horizontal strips of alternate colors at 10-cm intervals was used to measure the height of vegetation in each of the plant formations described. The board was placed at the center of each sample plot and the vegetation height was visually detemnined by an obsen/er from a 20-m distance. The same procedure was employed to measure the water depth during the flood.

PAGE 40

29 Classification and Ordination of Plant Formations The process of classification was performed in 2 steps. First, the initial matrix was transformed to express dissimilarities among all sample pairs (Williams etal., 1966; Williams etal., 1973). In the second step, the dissimilarities, expressed in terms of geometric distances, were fused into clusters (Goldsmith et al. , 1 986). To classify the initial matrix of stands-by-species into clusters of closely related units, I used the computer program PHYTOPAK developed at UNICAMP University, Sao Paulo, Brazil. Initially, a qualitative Q-classification analysis was performed by calculating the importance of each plant species within each of the plant formations (Index of Specific Value modified from McCloskey's IBV, 1970). The one-tailed significance limits (a = 0.05) for Jaccard Similarity betv^en formations (0.25 and 0.35, corresponding to 0.05 and 0.95 percentiles of the frequency distribution of similarities obtained by chance) were established after 40 permutations of all species occurrences. To examine affinities between species, a qualitative R-Classification analysis also was performed (UPGMA, Jaccard Similarity). To minimize noise, species with ISV lower than 0.2 were excluded from further analyses. To explore and test quantitative relations between the sampled formations, Q-Classification (Bray Curtis Distance) and Correspondence Analyses (Ter Braak, 1989) were performed using the complete data set. Correspondence Analyses was executed subsequently with reduced data. The

PAGE 41

30 significance limits (0.57 and 0.76) for distance values were established as explained above. Richness and Diversity of Plant Formations Two procedures were used to measure richness of each of the plant formations discriminated by the Quantitative Q-Classification technique. In one procedure, the index of richness v^s measured by dividing the total number of species of a given plant formation by the area sampled. All plant formations had equal sample size (1 ,000 sampling points) except Forest Edge, which had a smaller sample (600 points). I employed the rarefraction method (Hurlbert, 1971), which estimates a richness index regardless of sample size, as a comparative method. I employed Shannon's diversity index to obtain a measure of diversity within each plant formation (Shannon & Weaner, 1949, in Ludwig & Reynolds, 1988). Shannon's index (H') has two properties: 1) H' = 0, when there is only one species in the formation; 2) H' is maximum when the number of species equals the number of specimens sampled. Here, I refer to diversity as an index of cover diversity instead of species diversity because large plants (e.g., grass tussocks) can be recorded more than once per sampling point. Results Cumulative Number of Species In general, the grassland formations had a low number of species, especially those dominated by carpet grass Axonopus purpusii or wire grass Elyonurus muticus. The number of plant species was greater in moist shallow patches of

PAGE 42

31 vegetation, where hydrophytic forbs and shrubs grow, and in the least flooded soils of the Scrub and Forest Edge (Table 4-1 ). The smaller number of species in the grasslands is probably a consequence of the more drastic hydrological alterations suffered by plants on grassland soils. Only few resilient species can survive total submersion and subsequent drought. Species Associations A total of 1 55 species of 56 families of plants were identified dunng the four months of sampling. The majority (1 03) of the species were associated with savanna, whereas fewer (52) were associated with forest. The most represented families were Poaceae, Cyperaceae, and Fabaceae with 22, 12, and 11 species, respectively. Ten plant species were unidentified. The following species were found consistently co-occurring (Jaccard Similarity Index [JSI] > 0.7): a) Eleocharis elegans (H.B.K.) Roem. & Schl. (Cyperaceae), Aeschynomene fluminensisVeW. (Fabaceae), and Hydrolea spinosa L. (Hydrophyllaceae) which were common in Marsh Pond formation; b) Nymphoides indica (L.) 0. Kuntze (Menyanthoideae) and Ludwigia inclinata (L.F.) Gomez (Onagraceae) which were common in Marsh Pond formation; c) Richardia grandiflora (Cham. & Schl.) Steud. (Rubiaceae) and Schizachyrium microstachyum (Poaceae) which were common in Tall Grass and Mixed Association.

PAGE 43

32 Table 4-1 . Cumulative number of plant species identified per sampling unit (five 1m^ X 20 sampling points) in the Pantanal Study Area (Caiman Ranch), November, 1991. MP** IVID oo Our 1 U 1 o [VIM Mo Mo/ 1 o rt 01 11 12 08 04 07 12 03 09 07 31 02 18 16 10 06 09 16 06 12 17 50 03 25 21 11 08 13 18 11 15 27 58 04 27 22 14 12 16 22 14 16 29 69 05 30 25 14 17 16 27 16 16 30 71 06 30 26 31 72 07 32 27 31 08 34 28 35 09 34 28 36 10 34 28 36 ^ Sampling unit. Marsh Pond. Moist Basin. Short Grass. ^ Short Grassrr all Grass. 'Tall Grass. ^ Mixed Association, 'h Medium Grass, 'i Medium Grass/Tall Grass. 'Scrub. ^ Forest Edge.

PAGE 44

33 The following groups were recognized by reducing the JSI to > 0.6,: a) Eleocharis elegans, Aeschynomene fluminensis, Hydrolea spinosa, and Hymenachne amplexicaulis (Rudge) Nees (Poaceae) which were common to Marsh Pond; b) Echinodorus longiscapus Arech (Alismataceae) and Justicia laevilinguis (Nees) Lind. (Acanthaceae) which were common to Marsh Basin; c) Caperonia castaneifolia (L.) St. Hil.(Euphorbiaceae), Diodia kuntzei Schum.(Rubiaceae), and Eleocharis acutangula (Roxb.) Steud. which were common to Marsh Basin; d) Richardia grandiflora, Schizachyhum microstachyum, Axonopus purpusii (Mez) Chase (Poaceae), Vernonia scabra Pers. {V. brasiliensis) (Compositae), Elyonurus muticus (Spr.) Kunth (Poaceae), and Trachypogon sp. (Poaceae) which occurred together in formations including Medium Grass, Tall Grass, Mixed Association, and Scrub. Classification of Plant Formations After eliminating rare species from the data base, 73 species were ordered under R-Classification analyses. Plant species abundance within formations is presented in Fig. 4-1 . Qualitatively, two groups of plant formations appeared significantly distinct (Similarity Index > 0.70): a) Marsh Ponds, Moist Basins, and Short Grass (plants adapted to flood; lowest microrelief ); and b) all other formations (Fig. 42). Group b was constituted by three distinct sub-groups including: Medium/Tall

PAGE 45

34 Grass, Short/Tall Grass, and Tall Grass; Medium Grass and Scrub; and Mixed Association and Forest Edge. These sub-groups are possibly related due to similar flood regimes and soil characteristics. On the other hand, quantitative Q-Classification revealed 4 to 6 distinct groups within the confidence interval for the Bray Curtis distances. The most conservative classification included 4 groups of plant formations: a) Marsh Pond group; b) Short Grass group; c) Scrub group; d) Forest Edge (Fig. 4-3). The lower limit for the confidence interval encompassed 6 groups of plant formations: a) Marsh Ponds; b) Moist Basins; c) Short Grass group; d) Tall grass group; e) Scrub group; and f) Forest Edge. Only the upper limit of the confidence interval was selected to segregate distinct plant formations and to minimize possible errors of identification during the habitat use observations for ungulates. The plant formations included the following characteristics: a) low areas that were permanently inundated or strongly affected by periodic flood, i.e. Moist Depressions; b) areas of low relief that were covered mainly by short grasses; c) areas of intermediate relief that were covered by short and tall grasses, with shrubs and small trees interspersed in different combinations of densities, i.e. Scrub. d) areas of high relief in the borders of tree islands, i.e. Forest Edge.

PAGE 46

35 MediumH'air grass -1 o o • O O Tall grass Mixed assoc. Mid -grass Scrub Short grass ShorVTall grass Forest edge Marsh pond Moist basin Oo • > o « OCD <@^0° o ooo ^^^siy" • • • ooooooQOD^^C^^ oo ^ O'OQ" 0>''»* -00° • OCOxO' I* • • • O O ' O o e o • O-o •OOOoo O • 0^)00 • o • oO • • • o o OoooooOOoo. ooo<^(J5J)() 0yo o COO' • • • CD oOOOOQ) ' I ' ' ' ' I ' ' ' I ' ' ' ' I ' ' ' I I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' 5 10 15 20 25 30 35 40 45 50 55 60 65 70 1 Annocom 16 Vemscab 31 Angi 46 Stylacum (2) 61 Echipani 2 Ichnproc 17 Annodioi 32 Hellguaz 47 Mimoaden 62 Sagiguia 3 Byttdent 18 Curaamer 33 Forspube 48 HymeampI 63 Eleoacut 4 Sebahisp 19 Wedebrac 34 Arb.gros 49 Justiaev 64 Aescflum 5 Crotcoru 20 Brachumi 35 Arb.BC 50 Leerhexa 65 Aeschist 6 Andrsell 21 Andrhypo 36 Ruta 51 Eleonodu 66 Eleoeleg 7 Erytdeci 22 Sporjacq 37 Bauhmoll 52 Nympindi 67 Hyptmicr 8 Byrsorbi 23 MIrtglab 38 Attaphal 53 Phyllind 68 Hidrspin 9 Psidguin 24 Cypebrev 39 Sidalini 54 Nympsp 69 Diodkunt 10 Melovlll 25 HIptlapp 40 Convamar 55 Echilong 70 Reimbras 11 Axonputp 26 Euphthym 41 Baccmedu 56 Capecast 71 Panilaxu 12 Elyomuti 27 Coccsp 42 Plandesc 57 Senntor 72 Chomobtu 13 Schimicr 28 Caseacul 43 Crotsp 58 Rhyntenu 73 Melosimp 14 Richgran 29 Paulptnn 44 Stylacum (1) 59 Eleomini 15 Tracsp 30 Adenflor 45 Ruelgemm 60 Staccaya Figure 4-1 . Quantitative nodal diagram for Indexes of Species Values according to the sampled plant formations at the Pantanal Study Area, (Caiman Ranch), 1991-1992. Species names and index values are presented in Appendix A.

PAGE 47

36 I" M 1 CO o « u u M 0) (0 (0 u> CO «> CO CO m Scr o> o> To IE £ o 0) Sh CO Figure 4-2. Classification of plant formations at the Pantanal Study Area (Caiman Ranch) according to the presence of species (Qualitative QClassification). The two horizontal lines represent higher and lower significance limits (a = 0.05). Solid circles indicate significantly distinct groups of plant formations. Open circles indicate significantly related groups of plant formations.

PAGE 48

37 I.On o (0 (0 (ft to 0) OS OS u re o D) GO ZJ)
PAGE 49

38 Ordination of Plant Formations The site ordination diagram, all species included, revealed that points corresponding to Marsh Pond, Moist Basin, Short Grass, and Forest Edge constituted distinct clusters. The remaining formations were less clearly separated (Fig. 4-4). In addition, the ordination diagram demonstrated a gradient (x-axis) ranging from the wettest formation (Marsh Pond) to the driest (Scrub) in the floodplain. Pedological studies in other regions of the Pantanal indicated that the y-axis could be related to the levels of micronutrients in the soil, although measurements of soil characteristics were not conducted in the Pantanal Study Area (Brum etal., 1987, Pott etal., 1987, Pott etal., 1989). Thus, the y-axis would range from the least fertile soils (bottom) to the most fertile soils (top. Fig 4-4). The species ordination revealed essentially the same pattern as site ordination, but formations were not distributed as distinctly due to the presence of transition species (Fig. 4-5). However, it suggested that Forest Edge species were related more to the grassland species than to species associated with moist soils. Richness and Diversitv of Plant Formations An analysis of the richness and diversity was performed for the six groups segregated by the lower limit of the Bray Curtis distances to further explore the differences and similarities among the distinct plant formations (Fig. 4-3). Mixed

PAGE 50

39 FEA3 Figure 4-4. Site ordination diagram showing clusters of sampling units, Pantanal Study Area (Caiman Ranch). MP = Marsh Pond; MB =Moist Basin; SG = Short Grass; STG = Short-Tall Grass; MG = Mid-Grass; MTG = Mid/Tall Grass; TG = Tall Grass; MA = Mixed Association; SCR = Scrub; FE = Forest Edge. The numbers refer to the sample unit.

PAGE 51

40 Figure 4-5. Species ordination diagram showing clusters of closely associated species Pantanal Study Area (Caiman Ranch). MP = Marsh Pond; MB =Moist Basin; SG = Short Grass; STG = Short-Tall Grass; MG = Mid-Grass; MTG = Mid/Tall Grass; TG = Tall Grass; MA = Mixed Association; SCR = Scrub; FE = Forest Edge. The numbers refer to the sample unit.

PAGE 52

41 Association was analyzed independently from the other plant formations to determine how soil disturbance influenced the richness and diversity of Tall Grass. Species richness indices indicated a trend of inaeasing richness from Marsh Pond to Forest Edge (Table 4-2, Fig. 4-6). The increase in species richness coincided with deaeasing levels of inundation, or increasing elevation of formations. The most flooded formations exhibited the least species richness, and the driest, the greatest richness. Grassland formations exhibited intermediary indices of species richness, mostly due to the presence of rare and ephemeral species amidst the few perennials that could tolerate the extremes of flood and drought. Thus, diversity indices were lower In grassland formations than in other formations, except for Mixed Association. The greater richness and diversity exhibited by Mixed Association was presumably caused by the plowing of a Tall Grass soil, which allowed the colonization of species characteristic of Forest Edge (Figs. 4-2 and 4-3). Characteristics of Plant Formations Marsh Pond Marsh Pond was characterized by the constant presence of water. Water depth in sampling units ranged from 0 to 40 cm in the dry season and 40 to 80 cm in the flood season. Vegetation height averaged 40 cm, ranging from 0 to 80 cm. Dominant species for this formation were the hydrophytic forbs Hydrolea spinosa and Hyptis microphylla and sedges {Eleochahs spp.); leguminous shrubs {Aeschynomene spp.) also were common fTable 4-%

PAGE 53

42 Table 4-2. Richness and diversity indices for plant formations at the Pantanal Study Area (Caiman Ranch), 1991-1992. Data for all sampling seasons v^re combined. MP^ MB" SG" TG' SC" MA' FE^ 39 36 52 59 63 66 114 R' 0.78 0.72 1.04 1.18 1.26 1.32 3.80 H"' 2.34 2.48 2.29 2.24 2.30 2.83 3.61 n'^ 4,000 4,000 4,000 4,000 4,000 4.000 2,400 ^ Marsh Pond. Moist Basin. " Short Grass (Short Grass + Short/Tall Grass). " Tall Grass (Tall Grass + Medium/Tall Grass). * Scrub (Medium Grass + Scrub). ' Mixed Association. ° Forest Edge. ^ Cumulative number of species. ' Richness Index. ' Shannon-Weaner Index of diversity. ^ Total number of sampling points (20 points/m^).

PAGE 54

43 120 T 0 500 1000 1500 2000 2500 3000 Number of Sample Points -AMarsh Pond -xMarsh Basin Short Grass -•Tall Grass -•Scrub -Forest Edge Figure 4-6. Rarefraction curves showing cumulative number of species versus sample points for distinct plant formations, Pantanal Study Area (Caiman Ranch), 1991-1992.

PAGE 55

44 Moist Basin Moist Basin was characterized by small and shallow depressions in the terrain, and thus, exhibited higher moisture levels than the adjacent grassland even during the dry season. These basins accumulated 1 0 to 20 cm of water during the peak of the rainy season, and 40 to 50 cm during the flood season. This formation was characterized by a transition between Marsh Pond and Short Grass formations. Height of the vegetation varied between 20 and 40 cm. At the beginning of the rainy season (November), ground cover consisted of short grasses (40%), medium grasses (3%), and several forbs (35%); 22% of the surface was bare soil due to cattle trampling. Moist Basin was identified by the presence of its dominant forb species, Melochia simplex (Sterculiaceae), and a common grass, Panicum laxum (Table 4-3). Short Grass This formation was characterized by a carpet of short grasses, with vegetation height averaging 10 cm, but also contained isolated tussocks of tall grass that attained 120 cm in height. Axonopus purpusii (short grass) was the dominant species, but small forbs such as Diodia kuntzei and sedges, especially Eleochahs acutangula, were prominent (Table 4-4). Tussocks of tall grasses (e.g., Schyzachyhum microstachyum, Trachypogon sp., Sporobolus jacquemontii) were present, but in very low abundance. Shrubs also were present in low densities, e.g. Vernonia scabra and Annona cornifolia. This formation accumulated 20 to 30 cm of water in the flood season, although it remained dry during most of the year. Short

PAGE 56

45 Table 4-3. Relative cover percentages of the six common plant species in Marsh Pond and Moist Basin plant formations in the Pantanal Study Area (Caiman Ranch) 1991-1992. Water regime: Marsh Pond = permanently submersed, with 40 80 cm of water in the flood season; Marsh Basin = submersed during the rainy and flood seasons only, with 40 50 cm of water in the flood season. Marsh Pond Moist Basin Species % Cover Species % Cover Hydrolea spinosa 28.4 Melochia simplex 32.9 Hyptis microphilla 27.5 Panicum laxum 12.1 Eleocharis elegans 6.9 A. fluminensis 8.4 Eleochahs acutangula 6.0 Eleocharis nodulosa 7.2 A. fluminensis 5.3 Hyptis microphilla 5.1 Aeschynomene histrix 4.8 Echinodorus longiscapus 4.8 Table 4-4. Relative cover percentages of the six commonest species for seasonally flooded plant formation types in the Pantanal Study Area (Caiman Ranch), 19911992. Surface water accumulated only during the flood season: SG = 20 30 cm; TG = 20 cm; MG < 10 cm; SC < 5 cm. Short Grass Tall Grass Species % Cover Species % Cover Axonopus purpusii 44.6 S. microstachyum 38.5 Diodia kuntzei 9.9 Elyonurus muticus 14.0 Reimarochloa brasiliensis 8.4 Axonopus purpusii 10.8 Vemonia scabra 5.9 Annona dioica 8.4 S. microstachyum 4.6 Richardia grandiflora 7.8 £ acutangula 3.4 Helicteres guazumaefolia 3.9 Scrub Forest Edge Species % Cover Species % Cover Elyonurus muticus 44.7 Chomelia obtusa 14.1 Axonopus purpusii 8.1 Panicum laxum 7.9 Vernonia scabra 7.3 Psidium kennedyannum 5.9 Annona dioica 6.2 S. microstachyum 5.4 Trachypogon sp. 4.6 Paspalum virgatum 5.2 Curatella amencana 4.4 Hyptis spp. 5.1

PAGE 57

46 grasses represented 17-53% of the cover, tall grasses 0-28%, forbs 37-44%, and shrubs 8-11%. Tall Grass The characteristic of this formation was the dominance of tall grasses, chiefly Schyzachyhum microstachyum (Table 4-4). The lower stratum consisted of short grasses, such as Axonopus purpusii, and forbs such as the very abundant Richardia grandiflora. Short grasses constituted 11-16% of total cover, tall grasses 32-44%, forbs 19-22%, and shrubs 22-35%. Average vegetation height was 120 cm, and flood reached 20 cm depth. Scrub This formation was identified by the dominant presence of a tussock grass of 30-40 cm height, Elyonurus muticus, and the abundance of shrubs and small trees. Common shrubs included Vernonia scabra (Compositae) and Annona dioica (Annonaceae) (Table 4-3). The most common tree was Curatella amehcana (Dillenaceae). Water depths rarely exceeded 2 cm during the flood. Short grasses covered 8-1 1 % of the terrain, tall and medium grasses 28-33%, forbs 1 8-25%, and shrubs 25-46%. Forest Edoe Forest Edge constituted the interface between any of the above formations and semi-deciduous forest patches. Forest Edge consisted of a diverse array of species including vines, fruit trees {Psidium kennedyanum, Myrtaceae), and shrubs {Chomelia obtusa, Rubiaceae) in addition to the species that occurred in the J

PAGE 58

47 floodplain. The core of the forest patches was covered most commonly by bromeliads and trees, such as Tabebuia heptaphylla (Bignoniaceae), Genipa americana (Rubiaceae), and Vitex cymosa (Verbenaceae). Two palm trees were common, Acrocomia totai and Attalea phalerata. Discussion Visual identification of vegetation types was a valid method to discriminate different plant formation types in the Pantanal Study Area. Visual identification correlated with numerically distinct plant formations for moist soils (Marsh Pond and Moist Basin) and Forest Edge. Furthermore, this empirical method, based solely on vegetation structure, allowed desaimination of grassland formations beyond actual differences in species composition and density. Therefore, habitat-use frequencies obtained from visual observations of cattle and cervids could be correctly assigned to the smallest possible variations of habitat within the floodplain of the Pantanal Study Area. This classification of vegetation represents a level of refinement beyond previous classifications of the floodplain vegetation (Veloso, 1948). General categories of vegetation were partitioned into smaller physiognomic units, which better represented the complex mosaic of marshes-grassland-tree islands, that constituted the floodplains in the study area. More specifically, the Mesophillous Zone (seasonally flooded grasslands) deaibed by Veloso was sub-divided into Short Grass, Tall Grass, and Scrub, and the Hygrophillous Zone was segregated into Marsh Pond and Moist Basin.

PAGE 59

48 The proposed classification is not exhaustive for the Pantanal, rather it is limited to the floodplain of the study area. Similar phytosociological studies need to be expanded locally, and especially to other regions of the Pantanal, due to the great diversity of habitats in the region. Traditionally, flooded grasslands have been classified broadly according to the dominance of conspicuous species e.g., "canjiqueiral" dominated by a shrub {Byrsonima intermedia); "caronal" dominated by a species of grass Elyonurus muticus, etc. (Allen & Vails, 1987). Ho\A/ever, there has been little phytosociological research in the Pantanal (Veloso,1948; Eiten, 1985; Prance & Schaller, 1982), and therefore, several distinct plant formations within the floodplains may have been overlooked. The ordination of the plant formations within the floodplain correlated with the water depth during the flood, which is ultimately related to the duration of flooding and, probably, to the availability of moisture in the soil during the dry season. The gradient from wet to dry soils, observed for Moist Depression, Short Grass, Scrub, and Forest Edge, indicated that the availbility of water, associated with the microrelief, was possibly the main factor that determined the pattern of distribution of plant formations within the floodplain. Corroboratively, Prance & Schaller (1982) suggested that the degree of humidity, soil type, and duration of standing water were probably responsible for the differences in plant dominance observed in their Pantanal study site. Multivariate analysis (Robertson etal., 1984) indicated that the environmental factors, which were related strongly to the vegetation gradient in a North American swamp, were depth of flooding, soil texture, and drainage.

PAGE 60

49 The general reduction of diversity in the plant formations from the Forest Edge to the Marsh Pond probably reflected the stress imposed by inaeased flooding and poorly drained soils (Robertson etal., 1978, 1984). Diversity was highest in Forest Edge because reproduction, growth, and survival of both Forest Edge and floodplain species were not strongly limited by extremely dry or wet conditions. The second variable implied in the ordination of plant formations was soil fertility. Grassland soils have greater levels of Fe and Al than the soils in the forest patches (Pott etal., 1989). High levels of Fe can be toxic (Nores, 1984), and, together with Al, increase deficiencies of P (Conrad ef a/., 1985). Brumefa/. (1987) found that soils in the forest patches had greater levels of P compared to the soils of the scrub and swamps. Variables such as soil texture, ph, salinity, and frequency of fire may also influence the distribution of plant species within an ecosystem. The multifactor control of plant species distributions has been implicated in numerous studies (Robertson etal., 1978, 1984; Westman, 1980; Muller, 1982). Likewise, there is evidence that other environmental variables than flood and soil fertility are influencing grass species abundance and distribution in the Pantanal Study Area. Axonopus purpusii and Elyonurus muticus, for example, dominated sandy soils, which were characterized by high acidity and low salinity, whereas soils characterized by high levels of salinity were dominated by Paratheria spp., which require neutral or alkaline soils (Allen & Vails, 1987).

PAGE 61

CHAPTER 5 CATTLE DIET AND HABITAT SELECTION Introduction The future success of wildlife conservation in the Pantanal ultimately depends on wise use of natural resources. Until today traditional cattle ranching and low human densities have enabled wildlife to thrive in the region. However, there are no studies to verify if modern techniques of cattle management have modified the habitat and disrupted wildlife populations. Currently, there is an urgent need for knowledge of how cattle and wildlife use the natural resources within their environment, because cattle ranching is the prioritized source of income to entrepreneurs exploiting the combination of cattle ranching and tourism. Therefore, improved forage management techniques that promote both wildlife conservation and cattle production could be derived from such information. Cattle coexist with two large native herbivores, the brown brocket deer {Mazama gouazoubira) and the pampas deer {Ozotoceros bezoarticus ) in the native pastures of the Pantanal Study Area. Marsh deer {Blastocerus dichotomus) and capybaras {Hydrochaeris hydrochaeris) are restricted to areas adjacent to permanent sources of water, which represent a small portion of the 50

PAGE 62

51 Pantanal Study Area. Therefore, the selected study site provides satisfactory conditions for a study of native herbivore habitat use and diet in the presence of cattle. A complete quantitative analysis of the diet of cattle in the Pantanal has not been reported, although there have been a few studies describing it qualitatively (Allen & Vails, 1987; Pott, 1988). Consequently, the potential competition for palatable forages between cattle and both pampas and brown brocket deer has not been assessed. Objectives The objective of this chapter was to investigate the diet of cattle and their habitat preferences. Specifically, the following hypotheses were addressed: Hoi . The proportions of distinct plant species (families) in the diet of cattle do not differ among seasons. Ho2. Cattle use the different plant formations according to availability. Methods Assessment of the diet of free-ranging animals is a difficult task. Field biologists have employed several techniques to assess diets of ruminants: a) direct observation of what the animal is eating (Pott, 1986; Rajasekaran, 1988; Rodrigues, 1996); b) fecal analysis (Elliott III & Barrett, 1985; Green, 1987; Quintana etal., 1994; Martinez etal., 1997); and c) collection of digestive tracts (Pott, 1982; Branan etal., 1985; Bodmer, 1989). The method selected in the present study was required to fulfill the following conditions: 1) be applicable to

PAGE 63

52 cattle as well as to free-ranging deer; 2) be humane and harmless to individuals; 3) be able to provide a significant sample of the local population; and 4) not interfere with the natural habits of the individuals sampled. Accordingly, fecal analysis was selected as the method of choice. Fecal analysis has been used in a number of studies of diet overlap between cattle and other ungulates (Elliot III & Barrett, 1985 ; Jackson & Giuletti, 1988 ; Martinez et al., 1997). It also allows comparisons for the same species in different areas or seasons (Bonino & Shriller, 1991; Quintana etal., 1994; Martinez etal., 1997). Furthermore, fecal analysis is considered as accurate and precise as other existing methods used for the determination of dietary composition in herbivores (Homolka & Heroldova, 1992). Finally, the analysis of the nutrient status of feces may offer a guide to dietary quality of ingesta (Erasmus etal., 1978; Putman, 1984; Green, 1987). Fecal Samplino and Analysis Five 200-mg samples of fresh cattle dung were collected in the first week of each month and preserved in 75% ethanol in individual plastic vials. Samples were pooled by season: rainy (November, December, and January); flood (March, April, and May); dry (July, August, and September). The adequacy of number of samples was tested as described by Hanson & Graybill (1956). Samples were washed, clarified in NaOH (10%), and passed through a net of 80 BTN to segregate the undigested fragments. Samples were equally diluted to reach a density of 3-6 identifiable fragments per microscopic field. Four slides of

PAGE 64

53 each fecal sample \A/ere mounted in silicone gel. Fragments from 25 fields (100 x magnification) on each slide were identified by comparison with the reference collection. The resulting frequency data for each identified species were then transformed into the percentage composition of the diet by season (Johnson, 1982). Diet composition was compared among seasons using Spearman Rank Correlation Index (Sokal & Rohlf, 1969). Differences between the percentage contribution of plant species, families, and food categories were tested with the STP Mann-Whitney pairwise tests, which control the comparisonwise Type I error rate (Sokal & Rohlf, 1969). Trophic diversity was expressed by the Shannon-Weaner Index, and niche breadth by Levins method as standardized by Hurlbert (Krebs, 1989). Reference Collection Leaves and thin stems were washed and cleared as described by Sparks and Malecheck (1968). Grasses v^re prepared by rasping the mesophilum from the leaves so as to leave only fragments of the epithelium. Microhistoiogical slides were mounted in silicone gel and labeled. Drawings of each species' histological structures such as stomata, epithelial cells, inclusions, and trichomes were made, from the slides in the collection, to save time and facilitate identification of microscopic structures in the fecal samples. A key to segregate species was created to further aid in the identification of fragments by using a

PAGE 65

54 group of diagnostic characteristics such as stomata, trichomes, and cell inclusions as basic discriminatory elements. Habitat Use Habitat-use observations of cattle and native mammals were obtained during daylight hours in all plant formations by excursions on foot and by vehicle. A minimum of one survey of 64 km of dirt roads was conducted monthly by vehicle at an approximate rate of 15 km/h duhng early morning or late afternoon. Incursions on foot were performed once per month year-round across a 2-km trail located in a strip of forest within the floodplain. Ad libitum observations also were recorded when traveling the study area. The availability of the different types of plant formations (Chapter 4) was obtained by a random sample of 500 points distributed within the floodplain. This method was selected over the planimetry of the different plant formations based on satellite imagery, because of small size and discontinuous pattern of distribution of the plant formation patches. Forest patches and Moist Basin, for example, were frequently smaller than 200 m^ . Actual proportions of plant formations are recommended for parametric statistical analyses in use/availability studies (Alldredge & Ratti, 1986). However, estimates are acceptable if large sample sizes are obtained so that the margin of error is sufficiently small (Thomas & Taylor, 1990). Thomas & Taylor (1990) compared different methods for determining the number of random points necessary for studies with three or four distinct resources (95% confidence

PAGE 66

55 limits), and found numbers ranging from 385 points (Marcum & Loftsgaarden, 1980) to 510 (Thompson, 1987). Therefore, the sample size of 500 points used to estimate plant formation availability in the Pantanal Study Area was considered adequate, given the low number of different categories treated by the analysis (Thomas & Taylor, 1990). The procedure started by superimposing a grid of 1 ha on a topographic map of the Pantanal Study Area. Ten cartesian coordinates were then randomly selected, and used as starting point for 10, 2.5 km length transects; directions for conducting the transects also were selected randomly. The vegetation structure and plant species dominance were recorded on the transect at 50-m intervals to avoid dependence between sampling points. Later, the frequency data were assigned to the distinct plant formation types (Chapter 4) by using the information collected on structure and plant species dominance. Significance of difference in use versus availability of plant formations was tested mth the Chi-square goodness-of-fit-test, and the related multiple comparisons of the Bonferroni test (Neu etal., 1974; Byers etal., 1984). This method requires two assumptions: 1) availability is measured correctly; and 2) observations are independent events. The fact that cattle form herds might violate the second assumption. However, the large number of observations and the small size of plant formation patches relative to the size of a herd were assumed to minimize the chance of error.

PAGE 67

56 Results Diet Composition A total of 78 plant species from 25 families was recorded in the annual diet of cattle. Most of the diet consisted of grasses ( x = 80. 1 %; sd =7.2%; 20 species) and sedges ( x = 8.3%; sd = 5.1 %; 1 1 species). The families Sterculiaceae (x = 4.4%; sd = 3.0%; 7 species) and Lamiaceae (x = 2.1%; sd = 1.5%; 3 species) were the most represented among non-graminoid families. No other family comprised more than 1% of the annual diet (Table 5-1). Fewer species were consumed by cattle during the rainy and flood seasons, 62 and 63 species respectively, than during the dry season, 74 species. However, there was no significant difference in diet between seasons because cattle foraged mostly on the same grasses and sedges throughout the year (Table 5-2). The percentage use of the 1 1 most consumed species varied little among seasons, except for Mesosetum chaseae, which was consumed more in the flood season than in the rainy or dry seasons (Table 5-3) . On the other hand, the dominance of species varied by season (Kruskal-Wallis, P > 0.05). Axonopus purpusii (Short Grass), Hymenachne amplexicaulis (Marsh Pond), and M. chaseae (Scrub) were the dominant species in the diet during the rainy season (STP, P < 0.03). M. chaseae was the most important species during the flood season. M. chaseae grows in areas less subjected to flooding than H. amplexicaulis and A. purpusii (Chapter 4), which might explain its greater

PAGE 68

57 Table 5-1 . Number of species consumed and percentage consumption of major plant families (> 5%) in the diet of cattle, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition of Diet Rain Flood Dry Year Family n' % n % n % t'' Poaceae 19 80.0 20 86.0 20 74.2 20 80.1 Cyperaceae 10 7.0 9 4.5 11 13.5 11 8.3 Sterculiaceae 6 7.8 2 3.5 7 2.0 7 4.4 Lamiaceae 2 2.1 2 2.1 3 2.0 3 2.1 Others 25 3.1 30 3.9 33 8.3 37 5.1 ^ Number of species consumed by season. ^ Total number of species consumed. Average percentage of use. Table 5-2. Comparison of the cattle diet between seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Comparison rs^ n'' 95% CI P value Result Dry/rainy 0.76 29 0.546 to 0.881 <0.001 Similar diets Dry/flood 0.70 29 0.451 to 0.850 <0.001 Similar diets Rainy/flood a r» . _ 0.86 29 0.724 to 0.934 <0.001 Similar diets Spearman's Rank Correlation Index Number of food items in comparisons.

PAGE 69

58 utilization in flood season. A. purpusii was consumed more than the other species during the dry season (STP, P < 0.03). Habitat Use Cattle foraged predominantly in Short Grass during the rainy season (Table 5-4). Short Grass offered an almost homogeneous source of palatable growing grasses, which allowed the cattle to travel less in selecting food items. Cattle shifted their foraging activities to Scrub when Short Grass was inundated by flood waters and foraged on the new-growth of Short Grass after the flood receded. However, the forage was depleted rapidly because of the drought that followed the flood. Thus, cattle shifted foraging to Scrub after depleting the forage supply in Short Grass, which was indicated by the appearance of bare soil within the carpet of grass. Seasonal Diet Diversity and Niche Breadth The analysis of trophic diversity and niche breadth suggested that cattle adopted different feeding strategies according to the limitations imposed by seasonal climatic variations on food availability. New-growth of grasses provided cattle with abundant diversity and quantity of forage throughout the range in the rainy season. Cattle concentrated foraging activities in the Scrub during the flood season when a greater proportion of their diet consisted of fewer grass species, such as Mesosetum chaseae, presumably in response to a limitation in the availability of selected species (Tables 5-3 and 5-4). Consequently, trophic diversity and niche breadth were reduced from the rainy season to the flood

PAGE 70

59 Table 5-3. Percentage consumption (> 5%) by season of graminoid species to the diet of cattle as determined by fecal analysis, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition Family Species Rainy Flood Dry Average Poaceae Axonopus purpusiii 10.10 8.06 20.60 12.92 Brachiaria so 2.40 1.26 5.16 2.94 0\/nr\rinn H^nt\/lnn 2 15 5 41 2 70 3 42 Hymenachne amplexicaulis 6.96 5.43 2.86 5.08 Leersia hexandra 5.07 2.51 2.79 3.46 Mesosetum chaseae 15.20 33.40 4.84 17.81 Panicum laxum 4.62 2.82 5.53 4.32 Paspalum plicatum 5.64 2.79 2.01 3.48 P. pontanalis 5.16 5.73 6.24 5.71 Reimarochloa brasiliensis 5.23 1.99 3.20 3.47 Cyperaceae Cyperus sesquiflorus 3.05 1.05 5.67 3.26 Other Species 34.42 29.55 38.40 34.13 H" 3.15 2.79 3.31 Ba" 0.20 0.09 0.18 ^ Shannon-Weaner Index of diversity. Niche Breadth Index (standardized Levins). Table 5-4. Habitat selection of cattle, Pantanal Study Area (Caiman Ranch), 1991-1992. Habitaf Season P value MD" SG" SC' FE" Rainy Flood Dry 318.84 272.72 465.23 <0.001 <0.001 <0.001 Avoided Avoided Avoided Selected NS Avoided Selected Avoided Selected Avoided Avoided Avoided ^ Avoided = habitat used less than expected based on its availability. Selected = habitat used more than expected based on its availability. NS = no selection, habitat used in proportion to its availability (Bonferroni tests, a =0.05). Moist Depression. ^ Short Grass. Scrub. * Forest Edge.

PAGE 71

60 season (Table 5-4). Then, during the dry season, cattle consumed more of less selected food types, such as sedges, forbs, and bowse, in consequence of a reduction in biomass and palatability of grasses. The relative increase of these items in the diet caused a concomitant compensation in the trophic diversity and niche breadth. Discussion The percentage consumption of different food types as well as the selection among different plant formations by cattle varied according to the season. It is presumed that this variation was caused by differential availability and quality of selected food items in time and space. The distribution of biomass in fiber categories is spatially and temporarily dynamic, and herbivores can be expected to manipulate their food intake not only by dietary selection, but also by habitat choice and seasonal movement ( Jarman & Sinclair, 1 979; Hansen et ai, 1985; Gordon 1989a, b; Murray & Brown, 1993). Although feeding preferentially on Short Grass and Scrub, cattle in the Pantanal showed marked seasonal changes between these two formations. During the rainy season they selected Short Grass, and in the flood and dry seasons they selected Scrub. The plant formation use and selection by cattle in the Pantanal showed similar trends to that of cattle on the island of Rhum and to the African buffalo In the Serengeti (Jarman & Sinclair, 1979; Gordon, 1989a, b). Cattle and buffalo are primarily grazers because their large gut capacity and relative low energetic requirements per unit of weight allow them enough time to

PAGE 72

61 digest the carbohydrates of the plant cell-walls (Hofmann, 1973; Van Soest, 1982). On the other hand, these large bovids are forced to feed on a lower quality diet to meet their total daily energetic requirements, because highly digestible plants are much less abundant (Parra, 1978; Demment & Van Soest, 1985). Thus, buffalo occur everywhere in the Serengeti except on the short grass plains, which might have too low a biomass of forage to maintain such a large ruminant (lllius & Gordon, 1991). Similarly, cattle in the Isle of Rhum were able to use short grasslands only during the spring and summer, because of the large amount of vegetation growth. As the abundance of live material declined in the fall season, cattle left the short grasslands to feed on the tall grassland areas (Gordon, 1989a, b). The seasonal characteristic of the climate determined presumably the variation of availability and quality of forage in time and space in the Pantanal. During the rainy season cattle found optimal conditions of foraging in the Short Grass due to the higher proportion of new-growth of grass leaves to structural tissues in this plant formation than in the Scrub, which have a larger proportion of tall grasses . Demment and Van Soest (1985) compared the percentage biomass in cell wall for different grasslands and demonstrated that a greater percentage of the biomass tended to be concentrated in the high-fiber categories as the standing crop increased. This occurred because taller grasses possess a large proportion of stems and also because maturity and late-season temperatures increase the cell-wall content of grasses (Deinum & Dirven, 1971).

PAGE 73

62 Thus, it is valid to assume that Short Grass represents the best foraging ground for cattle during the growing-season. Short Grass was completely covered by water during the flood season in the Pantanal forcing cattle into the Scrub. Some species, such as A. purpusii, that were common to this habitat, became less available and cattle increased the percentage of other species in the diet. The decrease of growth and advancing maturity of grasses in the dry season made cattle to include a greater proportion of other food categories in the diet because selected species were less available and presumably less digestible. The exhaustion of selected forage in the Short Grass induced cattle to forage preferentially on the Scrub during that season as also Gordon (1989a) observed in the Isle of Rhum. Studies of cattle feeding habits in other regions of the Pantanal confirm the results of dietary preferences of the main species consumed in this study, especially A. purpusiiand M. chaseae (Allen & Vails, 1987; Pott, 1988). In addition to these two species, Pott (1982, 1986) reported that Panicum laxum was an important species for cattle in the Paiaguas and Nhecolandia regions of the Pantanal. However, P. laxum was not consumed in a greater proportion than other grass species in the Pantanal Study Area, even though it was abundantly available. In fact, its consumption was lower than Hymenachne amplexicaulis and Paspalum spp. Other grasses of secondary importance to cattle in the Pantanal Study Area, Cynodon dactylon, Leersia hexandra, Reimarochloa brasiliensis, and

PAGE 74

63 Setaria geniculata, also were considered less important to cattle by Pott (1982, 1986). On the other hand, Pott (1982) suggested that Andropogon spp, Elionurus muticus, and Trachypogon sp. were consumed regularly by cattle. The present study, however, indicated that the 3 latter species were avoided, as cattle did not consume E. muticus, and ingested little Andropogon spp. and Trachypogon sp.. Avoidance of certain plants may be related to the amount of silica and/or large percentage of cell wall in plant tissues, because the micronutrient content of plants cannot explain the full spectrum of dietary preferences exhibited by cattle. For example. Pott et al. (1987) found that A. purpusii was richer in Ca and Mg than M. chaseae, but similar in K and P contents. Also, only A. purpusii reached the Ca:P ratio (1 .9:1 .0) considered satisfactory for ruminants by the National Research Council (1976). Similar results were obtained by Brum etal. (1987) for the Pantanal's sub-region of Paiaguas, where A. purpusii showed higher contents of Cu and Zn. However, indices of preference (Chap. 08), for the top species consumed by cattle, indicated that A. purpusii was consumed less than expected in relation to its availability in the floodplain, whereas M. cliaseae was consumed in proportion to its availability. Therefore, differences in silica and digestibility between the two species may explain the observed differences in consumption. Another important aspect for cattle and range management in the Pantanal is that cattle use native forbs and browse to complement their diet.

PAGE 75

64 which can not be satisfied nutritionally by grasses only (Brum et al., 1987). Nongraminoid species averaged aproximately 12 % of the seasonal and annual diets of cattle, a value higher than reported previously (1%) (Pott, 1982; 1986). Forbs and shrubs of the Sterculiaceae family were especially significant during the rainy season. In another study, in addition to Helicteres guazumaefolia (Sterculiaceea), Pott and Pott (1987) also reported that Attalea phalerata, Cecropia pachystachya, Cordia g lab rata, Cost us sp., Smilax sp., Vitex cymosa, and few other plants not sampled in this study, were the most important forage based on nutritional value, acceptability by cattle and frequency of occurrence. Native grasses and sedges of the Pantanal did not contain the minimum levels recommended for Ca, Mg, P, Cu, and Zn during most of the year (Brum et al., 1987; Pott etal., 1987). For example, Grace (1983) recommended 10 ppm of Cu for lactating cows, but none of the grass or sedge species studied by Pott et al. (1989) achieved this levels. Deficiencies of plant micronutrients are more severe during the rainy season (Pott etal., 1989). However, Pott etal. (1989) found no significant seasonal differences of limiting micronutrients, such as Cu, in the liver of cows. Pott & Pott (1987) suggested that cattle complemented their need for Cu by consuming other plant species rich in this element, such as Cordia glabrata (Boraginaceaej, Byrsonima orbigniana (Malpighiaceae), Vernonia scabra (Compositae), Helicteres guazumaefolia (Sterculiaceae), and Chomelia obtusa (Rubiaceae).

PAGE 76

65 Collectively, these studies indicate that cattle in the Pantanal require the plant diversity provided by native pastures in order to meet their year-round nutritional needs. Furthermore, the survival of wildlife depends on the current diversified ecosystem, especially native forages. The transformation of the floodplain into mono-specific introduced pastures would create two negative effects: a decrease in sources of nutrients to cattle and a reduction of the wildlife diversity. Currently, the Pantanal has a carrying capacity three times less than the improved pastures outside the floodplain (Allen & Vails, 1987). Productivity of forage for cattle could be increased by controlling the two most important weed species in the study area: Elyonurus muticus and Schizachynum microstachyum. Allen & Vails (1987) recommended plowing to control Elyonurus muticus so as to allow fast-growing species such as Axonopus purpusii to colonize these areas; however, such a measure is unlikely to be practiced currently due to its probable high cost-benefit ratio. Cattle grazing should be maintained within carrying capacity to avoid proliferation of weed species, and pastures should be grazed in a rotational scheme both within and between seasons. Currently, cattle are kept in the same pastures year round. Movement of cattle among pastures is exercised only when seasonal flooding reaches intolerable levels. However, cattle density should be reduced in the floodplains during the flood to reduce grazing intensity to that dictated by the available carrying capacity. Cattle uproot tussocks of grasses and trample the soft soil

PAGE 77

66 during the flood causing large areas of bare soil when the water recedes. As a result, not only is carrying capacity reduced, but also, the invasion of weed species is encouraged. Presently, EMBRAPA (the Brazilian Institute for Agricultural Research) recommends a grazing quota of one cattle-unit per 3 hectares of native pastures for an extensive cattle management regime in the Pantanal (Dr. Arae, pers. comm.). However, this figure represents an average, that does not consider individual carrying capacity for the different types of plant formations that form the floodplain mosaic. Therefore, the knowledge of the carrying capacity for each of the identified plant formations (Chapter 4) will help to improve cattle management and productivity. The rational and efficient use of the native pastures of the Pantanal, plus the development of new breeds of cattle more adapted to the ecosystem, could guarantee cattle productivity (Behnke & Abel, 1996) better aligned to the maintenance of biodiversity. Ultimately, the preservation of biodiversity is important not only to the conservation of natural resources, but also to the development of ecotourism in the region.

PAGE 78

CHAPTER 6 BROWN BROCKET DEER DIET AND HABITAT USE Introduction The brown brocket deer, Mazama gouazoubira, is a small Neotropical cervid that occupies a broad variety of habitats from Mexico to northern Argentina (Avila-Pires, 1959; Czernay, 1987). Its shoulder height ranges from 350 to 610 mm, and its weight from 1 3 to 25 kg depending on the geographic region and ecosystem (Eisenberg, 1981; 1989; Czernay, 1987; Emmons, 1990; Redford & Eisenberg, 1992; Townsend, 1996). The smallest and darkest forms occur in tropical rainforests, whereas the largest and palest forms inhabit savanna habitats. Hunted for food, sport, and pelts trade (Ojeda & Mares, 1 982; Bodmer, 1989, Townsend, 1996), brown brocket deer occur in estimated densities of 0.83 individuals/km^ in the Amazon (Bodmer, 1989), 1.4-2.1/km2 in the Brazilian Cerrado (Leeuwenberg & Resende, 1994), and 0.5-2. 75/km^ in the Pantanal (Schaller, 1983; Lourival & Fonseca, 1997). Brown brocket deer are usually solitary, but may group for reproduction or for feeding when food is limited and patchy. Fawning may occur in any month of the year (Stallings, 1986; Redford & Eisenberg, 1992, Bisbal, 1994; Townsend, 1996). 67

PAGE 79

68 Brown brocket deer are found from the Atlantic coast to the Andes, and occupy a variety of ecosystems, including the forests of the northern South America and the eastern Brazilian Atlantic Forest, and the bushlands of the Venezuelan Llanos, Brazilian Cerrado, Pantanal, Paraguayan Chaco, Argentina, and Umguay (Czernay, 1987). Brown brocket deer also can survive in cultivated areas, provided forested patches are available and hunting pressure is not intense. The spectrum of ecosystems occupied demonstrates the degree of adaptation that is characteristic of this species. Brown brocket deer are considered frugivores (Bodmer, 1989; 1991), although they also consume browse and flowers (Branan etal., 1985; Stallings, 1984). On the other hand, brown brocket deer would be expected to exhibit diets reflective of the local availability of food items as suggested from the wide geographic distribution and range of habitats occupied by this species. For example, most of the plant biomass in the rainforest is located in the canopy, and, thus, is unavailable to cervids. Fallen fruits, due to their overt abundance, are perhaps the most significant source of food for brown brocket deer and to a number of other terrestrial herbivores in this ecosystem. In contrast, one may expect to find a larger proportion of browse in the brown brocket diet in open habitats such as the Pantanal, even though fruits also are present in the floodplain. Thus, the Pantanal is an ideal ecosystem to test the hypothesis that brown brocket deer are essentially frugivores.

PAGE 80

69 Objectives The following assumptions should be met if brown brockets are frugivores: Ho1. Fruits are available throughout the year. Ho2 The largest proportion of the diet consists of fruits. Ho3. The selected habitat is the forest patches where fruit trees are found. Methods Fecal Sampling and Direct Observations Five samples of brown brocket deer pellets were collected in the first week of each month, and preserved in 75% ethanol in individual plastic vials. Samples were pooled by season: rainy (November, December, and January); flood (March, April, and May); and dry (July, August, and September). Fresh samples were collected in early morning from "latrines" or known bed-sites of Individual deer at the edge of forest patches. In a few occasions, pellets were collected after observed individuals had defecated. Additionally, defecation within specific home ranges of individuals was induced by depositing pellets from one brown brocket deer into the home range of another. This method was employed after a radio-collared male was observed defecating beside the pellets of both sexes at the edge of its home range (Pinder, 1992). Monthly samples were assumed to be from adult males and females because samples were collected from sites known to be used by identified individuals (five females and six males).

PAGE 81

70 Direct observations of foraging events were used to help in identifying plant species consumed by brown brocket deer, and as a comparative method to fecal analysis. Observations were conducted daily in hourly blocks from 0600 to 1800 hrs, which coincided with the time period when data on habitat use and location of the radio-collared male were collected. Ad libitum observations of feeding events for at least 10 additional different individuals were conducted in diverse portions of the Pantanal Study Area. Microhistologic Analyses Fecal samples were washed, clarified in NaOH (10%), and mounted on four semi-permanent microhistologic slides. Distinct fecal samples were equally diluted to reach a density of three to six identifiable fragments per microscopic field. Twenty-five fields were located systematically on each slide, and were viewed at lOOx magnification for identifiable fragments. The resulting frequency data for each identified species were then transformed into the percentage composition of the diet by season (Johnson, 1982). The presence of loose trichomes (epidermal hairs) of a given plant species in the fecal samples may result in a bias for quantification of the relative contribution of each plant species in the diet. Trichomes are very abundant in some plant species and the simple quantification of the frequency of loose trichomes for every sampled microscopic field can overestimate the relative presence of that given species in the diet. Johnson (1982) recommended that only fragments of epithelium should be considered for the quantification to avoid

PAGE 82

71 this problem. Ho\A/ever, the counting of epithelium fragments only would result in an underestimation of those species in the diet of brocket and pampas deer because of the great digestibility of these particular species. I created a correction index for species of the families Malvaceae and Sterculiaceae, which possess abundant trichomes, to minimize this problem. Random samples of fecal material collected during the rainy season were quantified for 100 microscopic fields containing trichomes of species of the two families. A correction factor was then derived by dividing the frequencies of fragments with trichomes by the frequency of loose trichomes: i££^q^xioo X tricspi where: CI = correction index; freq = frequency of epithelium fragments with trichomes of species i; trie = frequency of loose trichomes of species i. The analysis revealed that 37% of the microscopic fields contained fragments of trichomes that were attached to the epithelium. This percentage was used as standard for adjusting the frequencies of Malvaceae and Sterculiaceae species in a given fecal sample. Flowers, fruits, and seeds of grasses were difficult to identify after digestion, and, therefore, their quantification was more subject to accuracy bias than the remains of grasses and woody browse. Nonetheless, the microhistologic method was used to assess the seasonal consumption of these

PAGE 83

72 food categories because it allowed for better precision in contrast to quantification by direct observations. Differences of percentage of use of species, families, and food categories were tested for significance with Kruskal-Wallis test (Sokal & Rohlf, 1969). Hierarchy of consumption among food items (species and categories) was verified with STP Mann-Whitney pairwise tests, which control the comparisonwise Type I error rate (Sokal & Rohlf, 1969). Correlations of diets between seasons was performed with the Spearman Rank of Correlation Index (Sokal & Rohlf, 1969). Trophic diversity was expressed by the Shannon-Weaner Index, and niche breadth by the Levins method as standardized by Hurlbert (Krebs, 1989). Phenoloqv of Trees The phenological stage of the most common species producing edible fruits in the floodplain was recorded from November 1991 to October 1992. A sample of 86 trees of seven species were selected randomly for determining fruiting phenology. Individual trees were marked with colored vinyl tape and visited at 30-day intervals. Data recorded on the monthly visits to each tree included the frequency of trees with flowers, the number of fruits on the ground beneath each marked tree, and the presence of brown brocket deer or peccary sign. Additonally, specimens of Tabebuia caraiba and Tabebuia heptaphylla (Bignoniaceae) also were monitored, after a brocket deer was observed eating their flowers.

PAGE 84

73 The selected species for fruiting analyses consisted of three palm trees {Acrocomia totai, Attalea phalerata, and Copernicia australis), one Rubiaceae {Genipa americana), one Myrtaceae {Psidium kennedyanum), one Cecropiaceae {Cecropia pachystachia), and one Fabaceae {Enterolobium contoiiisiliquum). A. phalerata and E. contortisiliquum grow inside tree islands. Copernicia australis inhabits the flooded plains, especially those that suffer cycles of flood and fire. The remaining species were found principally along the edge of forest patches. The data on fruiting were pooled by month for each species. To standardize the relative availability of fruits, an index of fruit availability was created: T * F FAI = *10 N where: FAI = fruit availability index; F = number of ripe fruits of each species found on the ground each month; T = number of trees of each species with ripe fruits on the ground each month; N = total number of ripe fruits on the ground during the study. Habitat Use Observations of brown brocket deer were obtained during daylight hours in all plant formations by excursions on foot and by vehicle (Chapter 5). Additionally, a radio-collared adult male was monitored (April 1991 -October 1992) to provide an equally proportioned number of observations of habitat use

PAGE 85

74 during all daylight hours. Ad libitum observations of habitat use were also recorded. To ensure independence of data, 30-minute Intervals were considered for the quantification of plant formation use (Swihart & Slade, 1985). Observations spaced temporally in this manner allowed brocket deer to have access to their entire area of use, which was approximately < 100 ha monthly (this study). Only locations performed during foraging or traveling events were used for habitat preference analysis. Availability of distinct plant formations in the Pantanal Study Area was obtained by transect sampling (Chapter 5). Results Diet Composition Microhistologic analyses of feces revealed that brown brocket deer consumed at least 139 species of 39 plant families in the Pantanal Study Area. The families that contributed most to the diet were Compositae, Euphorbiaceae, Malvaceae, Stercullaceae, and Rubiaceae. However, there was no significant difference in consumption among these five families within seasons (KruskalWallis, P > 0.05), except for the rainy season, when Stercullaceae was selected (STP, P = 0.009) (Table 6-1). Forbs and browse dominated the diet of brown brocket deer (STP, P < 0.05), comprising > 80% of their diet ( x = 84.6%, sd = 8.4%) during the rainy, flood, and dry seasons (Table 6-2). There were no significant differences in the consumption of forbs and browse within each season (STP, P > 0.05). Browse

PAGE 86

75 Table 6-1 . Percentage composition (>5%) of plant families in the diet of brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition of Diet Rainy Flood Dry Year Plant Family % n % n % n % Compositae 5 15.20 4 18.60 5 15.02 6 16.27 Euphorbiaceae 5 10.58 6 11.02 5 13.08 7 11.56 Malvaceae 1 10.62 2 11.62 2 13.56 2 11.93 Rubiaceae 6 13.56 6 18.68 3 17.96 6 16.73 Sterculiaceae 7 27.14 7 16.90 6 15.50 7 19.85 Others 71 22.90 72 23.18 63 24.88 139 23.65 Number of species in the family. See appendices for statistical tests comparing percentages. Table 6-2. Percentage composition of different food categories in the diet of brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 19911992. Rainy Flood Dry YeaK Food Category FA'^ DO' FA DO FA DO FA (398)' (119) (372) Graminoids 1.32 4.77 3.04 27.73 3.55 0.27 2.64 Small Forbs 22.67 28.14 33.39 13.45 47.01 54.84 34.36 Browse^ 58.17 35.93 54.11 50.42 37.59 37.63 49.96 Broad Leaf Forbs'' 0.04 0.00 0.73 0.00 0.12 0.00 0.30 Lianas 2.15 0.25 1.60 5.88 2.82 4.84 2.19 Grass Seeds 0.00 11.31 0.00 0.01 0.00 0.00 0.00 Flowers 0.00 0.00 0.00 0.00 0.00 2.42 0.00 Fruits 8.42 19.20 2.76 1.68 1.71 0.00 4.30 Unidentified 7.23 0.65 4.37 0.83 7.20 3.72 6.26 ^ Shrubs and tree seedlings. Alismataceae and Pontederiaceae. " Fecal analysis. Direct observations. ® Number of observations. ' Mean values among rainy, flood, and dry seasons.

PAGE 87

76 was consumed more than forbs only during the rainy season (STP, P = 0.009). Both fecal analyses and direct observations confirmed that other food categories were of minor importance, at least quantitatively, in the diet of the brown brocket deer. Direct observations indicated also that fruits and seeds of grasses were consumed more frequently in the rainy season, whereas tree flowers were consumed more frequently in the dry season. Flowers and grass seeds were not identified in fecal samples, possibly because of a higher digestibility of these tissues (Table 6-2). The percentage consumption of graminoids and browse remained constant between seasons (STP, P > 0.05), but forbs were consumed less during the rainy season than during the flood or dry seasons (STP, P = 0.028) . Direct observations indicated that the reduced consumption of forbs during the rainy season was replaced by the increased consumption of fruits. Overall, 10 species of plants constituted the main source of food in the diet of brown brocket deer: Vernonia scabra, Wedelia brachycarpa, Caperonia castaneifolia, Euphorbia thymifolia, Sida santamarensis, Chomelia obtusa, Richardia grandiflora, Bytneria dentata, Melochia pyramidata, and Melochia villosa (Table 6-3). Percentage consumption differed among these 10 plant species for rainy and dry seasons (Table 6-3) (Kruskal-Wallis, P < 0.02). In the flood season, however, the great variance in the diet of different deer did not allow for a significant difference of consumption among the top 10 plant species (Table 6-3) (Kruskal-Wallis, P = 0.126).

PAGE 88

77 Only a few species of fruits were observed being consumed by brown brocket deer: Soroceia sprucei, Vitex cimosa, Chomelia obtusa, Genipa americana, and Acrocomia totai. Seeds of Psidium guinense were found in the stomach of a female found dead during the dry season. Finally, fragments of Eryngium ebracteatum fruits were identified through microhistologic analysis. Large seeds (> 1 cm) of Acrocomia totai and Vitex cymosa fruits were discarded by the animals after mastication of the fleshy pericarp. Only fallen flowers of Tabebuia spp., especially Tabebuia caraiba (Bignoniaceae), were observed being eaten in abundance by brown brocket deer. Diet composition did not vary much from season to season, but a few plant species were consumed in greater proportions in a specific season. Two browse species characteristic of the Forest Edge, Bytneria dentata and Meloctiia pyramidata (Sterculiaceae), were consumed in greater proportions during the rainy season than during the flood and dry seasons (Kruskal-Wallis, P < 0.01). Two other browse species characteristic of the grasslands, Cliomelia obtusa and Vernonia scabra, were consumed in greater proportions during the flood season, and a forb associated with the Scrub, Wedelia brachycarpa, was consumed the most during the dry season (Kruskal-Wallis, P < 0.07). Overall, a comparison of the total diet between seasons indicated a significant similarity among the seasons (Rs > 0.48, P < 0.001 ; Table 6-4). However, the diet composition for the dry season differed from the rainy season(Rs = 0.34, P = 0.065), when only browse and forb species were

PAGE 89

78 Table 6-3. Percentage composition (> 5%) of species by season to the diet of brown brocket deer as determined by fecal analysis, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition Family Species Rainy Flood Dry Average nnmnn^itflp 6 81 16 16 1 02 8 00 0 08 13 58 5 24 vyapfsfUiila LaolailfSliUlla 1 .uu n no ft "^7 O.OI O. 1 Q Euphorbia thymifolia 0.04 7.21 0.46 2.57 Malvaceae Sida santamarensis 12.21 11.52 8.48 10.74 Rubiaceae Chomelia obtusa 3.57 10.49 0.42 4.83 Richardia grandiflora 3.16 0.82 17.46 7.15 Sterculiaceae Bytneria dentata 14.17 2.35 0.67 5.73 Melochia pyramidata 19.29 1.57 5.00 8.62 Melochia villosa 2.87 8.33 5.09 5.43 Other Species 32.09 37.18 31.77 33.68 14.23 4.88 8.93 9.93 H'" 3.24 3.20 3.20 0.11 0.10 0.09 Plant species not identified. Shannon-Weaner index of diversity. " Niche breadth index (standardized Levins). See appendices for statistical tests comparing percentages. Table 6-4. Comparison of brown brocket deer diets (browse and forbs) between successive seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Comparison rs^ n'' 95% CI P value Result Dry/rainy 0.34 30 0.000 to 0.625 0.065 Different diets Dry/flood 0.50 39 0.224 to 0.707 0.001 Similar diets Rainy/flood a r\ 1 _ 0.50 46 0.244 to 0.690 < 0.001 Similar diets Spearman's rank correlation. Number of food items in comparisons.

PAGE 90

79 considered due to the increased contribution of forbs to the diet during the dry season (STP, P = 0.028; Table 6-2). Nevertheless, the values of the Shannon Diversity Index and Trophic Niche Width obtained for the diet were similar in the three seasons, suggesting that the dry season was not limiting for brown brocket deer in the Pantanal Study Area (Table 6-3). Availability of Fruits and Flowers Data from the phenological study suggested that fruits consumed by brown brocket deer {A. totai, G. americana, and P. kennedyanum) were relatively unavailable on the ground, except during the rainy season, although no attempt was made to quantify the biomass of fruits in the Pantanal Study Area (Figs. 6-1, 6-2). Large flowers of Tabebuia spp., which were consumed by cattle, pampas deer, and brown brocket deer, were available during the dry season (July to September). Vitex cimosa (Verbenaceae) trees produced a great number of olive-like fruits from October to December, which were also observed being consumed by brown brocket deer. P. kennedyanum (berry), E. contortisiliquum (dry fruit), and C. australis (lignlfied palm fruits) had the highest individual production per individual tree, but fruits of A. totai (pulpy mesocarp under a thin hard pericarp) were the most frequently available overtime (Table 6-5). The largest combined availability of fruits on the ground, except for dry and lignified fruits, occurred from October to February, which coincided with the rainy season. The peak of fleshy fruits abundance occurred in November, when P. kennedyanum also produced fruits

PAGE 91

80 Figure 6-1 . a) Monthly percentages of palm trees with fruits; b) monthly percentages of palm trees with fruits on the ground, Pantanal Study Area (Caiman Ranch), 1991-1992.

PAGE 92

81 a) C. pachystachia -xG. americana --E. contorticiliquum -*-P. kennedyanum b) Nov Jan Mar May Jul Sep f C. pachystachia -xG. americana --E. contorticiliquum -aP. kennedyanum Figure 6-2. a) Monthly percentages of trees with fruits; b) monthly percentage of trees with fruits on the ground, Pantanal Study Area (Caiman Ranch), 19911992.

PAGE 93

82 Table 6-5. Index of fruit availability for tree species, Pantanal Study Area (Caiman Ranch), 1991-1992. Spp/month CP' EC' GA' PK" AP' CA' AT^ Total November 0. 00 1.24 0, 04 1 06 0. 25 0. 00 1, ,70 4.29 December 0 00 0.00 0 00 0 11 0 00 0 V/. 00 0. 00 0.11 January n nn 0.03 n no 0 nn , \J\J n n 0.99 February 0. 00 0.00 0, 01 0, 00 1, 52 2. 40 0, 25 4.18 March 0. 00 0.00 0. 00 0 ,00 0. 00 15, 17 0 ,08 15.25 April 0. 00 0.00 0. 00 0. 00 0. 26 19. 65 0. 04 19.95 May 0. 00 0.00 0. 00 0 .00 0, 00 0 00 0 ,00 0.00 June 0. 00 0.00 0. 01 0. 00 0. 05 0. 00 0. 07 0.13 July 0, 00 0.82 0. 01 0, ,00 0. 00 0, 00 0 ,11 0.94 August 0. 00 0.39 0. 00 0. 00 0. 00 0. 00 0 02 0.41 September 0. 00 12.33 0. 04 0 00 0. 00 0 00 0 23 12.60 October 0. 18 16.79 0. 00 0. 00 0. 02 0. 00 0. 52 17.51 Total 0. 18 31.6 0. 11 1 17 2, 44 37 58 3 28 ^ Cecropia pachystachia ^ Enterolobium contortisiliquum Genipa americana Psidium kennedyanum " Attalea phalerata ^ Copernicia australis ^Acrocomia totai. Table 6-6. Habitat selection of brown brocket deer in the Pantanal Study Area (Caiman Ranch), 1991-1992. Habitaf Season P value MD" SG= SC' FE' Rainy 31.61 <0.0001 Avoided Avoided Selected NS Flood 47.76 <0.0001 Avoided Avoided Selected NS Dry a A : _i _ _i 26.04 <0.0001 Avoided NS NS Selected Selected = habitat used more than expected based on its availability. NS = no selection, habitat used in proportion to its availability (Bonferroni tests, P = 0.05). ^ Moist Depression. Short Grass. Scrub. Forest Edge.

PAGE 94

83 (Table 6-5). A large production of dry fruits of E. contortisiliquum and lignified fruits of C. australis occurred at the onset of the rains (September/October), and in the flood season (March/April), respectively, but there was no evidence that brown brocket deer consumed these fruits. Fruits of A. phalerata were important duhng the flood, when a number of species of birds and mammals, such as parrots {Amazona aestiva), macaws {Ara chloropterus and Anodorhynchus hyacinthinus), coatis {Nasua nasua), howler {Allouata caraya) and capuchin {Cebus appela) monkeys, foxes {Cerdocyon thous), armadillos {Dasypus spp. and Euphractus sexcintus), feral pigs, and peccaries {Tayassu pecari and Tayassu tajacu) fed abundantly on these fruits. Habitat Use Brown brocket deer avoided open vegetation, such as Short Grass and Moist Depression, in all seasons (Table 6-6). They sought protection by concealment in tree islands and thickets for rumination and sleeping. However, brown brocket deer selected Scrub in wet seasons, and Forest Edge in dry seasons for foraging activities (Bonferroni, P = 0.05). During the rainy season, 51% of 170 brown brocket deer sightings were in Scrub in contrast to 17% in Short Grass and 9% in Moist Depression. During the flood, 58% of 165 sightings were in Scrub in comparison to 17% in Short Grass and 5% in Moist Depression. In the dry season, 355 sightings indicated that brown brocket deer decreased their use of Scrub (41 %), and increased their use of Forest Edge (24%) during those months.

PAGE 95

84 Discussion Browse and forbs dominated the diet of brown brocket deer in the Pantanal. The 3 assumptions tested in this study were not consistent with the hypothesis that brown brocket deer is essentially frugivorous in the Pantanal: 1) fruits were not abundantly available throughout the year; 2) less than 20% of the diet consisted of fruits, even during the peak of fruiting; and 3) Forest Edge was not selected over the other plant formations during the rainy season, when fleshy fruits were most available. Furthermore, brown brocket deer in the Pantanal Study Area were never observed eating dry fruits as they did in Paraguay (Stallings, 1984). The diet of brown brocket deer in the Pantanal was compatible with what should be predicted by the digestive capacity of their gut (Parra, 1978; Demment & Van Soest, 1985). Because gut capacity decreases linearly with body size, but metabolic rate is proportional to the three-fourth power of body weight, smaller herbivores require relatively higher energy per unit of body weight to maintain their daily metabolic needs (Parra, 1978). This means that smaller herbivores need higher rates of turnover fermentation compared to large herbivores, to be able to produce enough volatile fatty acids used to generate energy. This higher rate of turnover can be obtained when the animal selects high-quality foods, that is, plant parts with low llgnin content and a high ratio of cell content to cell-wall. These conditions are met by the ingestion of young browse and forb leaves,

PAGE 96

85 which have a smaller percentage of cell-wall compared to grasses, and a smaller percentage of lignin than seeds (Demment & Van Soest, 1985). One condition for brown brocket deer to be frugivorous was that fruits occurred in abundance, and were reliably available throughout the year. However, most commonly, only a few trees, all species combined, would have ripe fruits available on the ground at any specific time in the Pantanal Study Area. Furthermore, only A. totai, G. americana, and few other rainy season fleshy fruits, such as V. cymosa, were consumed by brown brocket deer during the rainy season. The lignified and small fruits of C. australis, which were never observed being consumed by any animal, presumably evolved to be dispersed by water during the flood, because they have no soft mesocarp to attract mammals. The other palm fruit, A. phalerata, is too large to be swallowed by brocket deer, and has a thick pericarp that needs to be handled and peeled-off (monkeys, macaws, coatis, etc.) or crunched by powerful jaws (peccaries and tapir) before the pulp can be eaten. Guanacaste tree fruits (E. contorticiliquum) are large, covered with a hard exocarp, and rich in saponin, and, therefore, are probably consumed only by large ungulates, such as tapir and cattle, or rodents (Rodentia). Cecropia fruits are probably consumed on the trees by birds and frugivorous bats, because very few were found on the ground. Finally, the characteristics of A. totai and G. Americana fruits indicate that they evolved to be dispersed by monkeys, large birds, or extinct members of the megafauna, but not by current terrestrial mammals (Abrahamson, 1989). These fruits remain on

PAGE 97

86 the trees for long periods, and sometimes decay while still attached to the branches (pers. observ.). The few ones that fall to the ground are, in its most part, those discarded by frugivorous or seed-predator birds, because most of the forest patches in the floodplain are too small to maintain monkeys. Therefore, the Pantanal seems to do not offer the necessary conditions for exclusive frugivory for a herbivore of the size of the brown brocket deer. Previous food-habits studies on brown brocket deer, based on the percentage of fruit remains encountered in their digestive system, suggested that the species prefers fruits and palm seeds (Stallings, 1984; Branan etal. 1985; Bodmer, 1989; 1990, 1991; Bisbal 1994). However, the results from the diet analyses in this study indicated that fruits/seeds did not constitute an important item year-round in the Pantanal. Although medium-sized or large seeds of fruits are compact food packets, the time spent in digesting their covering, and possibly detoxifying the seeds, may not be cost-effective in their foraging time/energy budgets, when other foods are readily available in the savanna (MacArthur, 1972; Pyke etal., 1977; Demment & Van Soest, 1985). Seeds may contain a number of palatability inhibitors, toxins, and indigestible matter to prevent seed predation (Abrahamson, 1989). In contrast, the pulp of some fruits may contain high percentages of lipids, carbohydrates, and mineral nutrients (Abrahamson, 1989), which might explain why cervids in the Pantanal Study Area were observed eating the pulp of fruits, but discarding large seeds.

PAGE 98

87 Frugivory in brown brocket deer seems to be only a contingency or adaptation to the environment, rather than a characteristic of the species. The fact that the sub-species of brown brocket deer studied by Stallings (1984) in the Paraguayan Chaco, which ate large proportions of fruits, is the same as in the neighbouring Pantanal, supports this hypothesis. In fact, brown brocket deer seem to be as well adapted to a diet of browse and forbs as to a diet of fruits. Brown brocket deer in the savanna seem to have similar size and densities as their counterparts in the forests (Stallings, 1986; Bodmer, 1989; Leeuwenber & Resende, 1994; Townsend, 1996; this study). Unfortunately, a direct comparison between the results of this study and other studies is difficult because different methodologies were employed. Neither microhistological analysis nor stomach content research can accurately measure the percentage amount of fruits ingested. For example, it could be argued that because seeds are less digestible than flowers and leaves, the contribution of fruits in the diet could be overestimated by simply comparing the volume or weight of fruits with other items in the rumen or stomach. In the Peruvian Amazon, for example, palm fruits were ingested entirely (Bodmer, 1989; 1991). On the other hand, in this study the large seeds of palm fruits and V. cymosa were discarded in the mastication process. Therefore, if the pulp of fruits is digested quickly, then the percentage contribution of fruits to the diet could be underestimated by the fecal analysis method. Direct observations of feeding by brown brocket deer in the Pantanal Study Area suggested that fecal analyses

PAGE 99

88 underestimated the contribution of fruits to the diet, but demonstrated that fruits were not the principal food source for the diet. Furthermore, if fruits were a selected item in the diet, then Forest Edge should be the selected habitat during the period of fruit abundance. However, during the rainy season, Scrub was selected over Forest Edge, which was used only in proportion to its availability. Forest Edge was the selected plant formation only duhng the dry season, when fruit production was low. Observations on habitat use by brown brocket deer in the Pantanal paralleled the findings of Leeuwenberg and Resende (1994) for the Cerrado, where most of animal locations were obtained in the Scrub. The study conducted by Schaller (1983), in a different area of the Pantanal, also indicated the avoidance of dense forests by brown brocket deer. Most of Schaller's sightings were at forest edge, and secondarily, in what he called Cerrado Forest, the equivalent to Scrub in this study. Brown brocket deer required shelter, such as forest patches, bushes or tall grass, for rumination or resting in the Pantanal, although they selected to forage outside or at the edge of the forest patches. Brown brocket deer are solitary animals that maintain concealment from predators by using vegetation as a visual barrier, and as a hiding place when being pursued (this study). Thus, two factors seem to contribute to the use of intermediary habitats between the forest and open habitats by brown brocket deer: food choice and predator avoidance. Niche separation with pampas deer in the grasslands, and red

PAGE 100

89 brocket deer {Mazama americana) in the deciduous and gallery forests (Schaller, 1983), could also explain such habitat preferences. In conclusion, brown brocket deer in the Pantanal are not frugivorous. The comparison between the studies in the Amazon and the present study suggested that the brown brocket deer is a generalist with respect to feeding habits. The hypothesis is that brown brocket deer adapts to a diet according to the characteristics of the ecosystem. Their digestive system enables them to survive on a browse diet or a frugivorous diet (Bodmer, 1989) depending on the type of habitat occupied within their large geographic distribution (Czernay, 1987; Pinder, 1997). The greater availability of browse and forbs in the Pantanal, in contrast to more heavily forested areas, allows the brown brocket deer a large diversity of food items from which to select their diet. The only viable option for them in forested areas may be to increase the percentage of fruits in the diet, because browse is presumably sparse and rich in secondary compounds (Golley, 1978; Leigh & Windsor, 1982). The xeric leaves and stems of the shrubs in the Paraguayan Chaco may not be sufficiently digestible to allow a balanced diet without the support of protein from seeds. Stallings (1984) found that brown brocket deer even consumed dry fruits during the dry season in the Chaco. The digestive system of brocket deer species allows them to exploit fruits, as pointed-out by Bodmer (1989), and, therefore, permits them to occupy habitats where exclusive browsers or grazers would be excluded.

PAGE 101

90 This versatility in feeding habits and adaptability to diverse ecosystems provide brown brocket deer with an advantage over other South American cervids with respect to their conservation. Large areas of wilderness in South America are quickly being transformed by human activities Into agricultural lands. Only scattered and small patches of habitat remain in private ownership in much of the region below the Amazon Forest, except for governmental protected areas. However, the conservation of brown brocket deer seems to be possible because of its ability to occupy edge habitats, provided that they are protected from overhunting. Further studies on the diet and habitat use of brown brockets should be conducted in open ecosystems, such as the Brazilian Cerrado, and in the remaining fragments of the Atlantic Forest of Southeastern Brazil, to provide additional insights for the future conservation of the species.

PAGE 102

CHAPTER 7 PAMPAS DEER DIET AND HABITAT USE Introduction The pampas deer, Ozotoceros bezoarticus, is an antlered deer of the open South American habitats. Its geographic distribution ranges between the parallels 8° -36° S and 44°-62° W, from Central Brazil, Bolivian lowlands, and Paraguay to the coastal plains of Saborombom Bay in Argentina (Carvalho, 1973; Jungius, 1976; Pinder, 1993, Merino etal., 1997). Adult males weigh up to 40 kg and reach 700-750 mm at the shoulder height (Nowak & Paradise, 1983). However, few measurements of wild individuals have been published (Schaller, 1983; Jackson, 1987). An adult male captured in this research weighted 40 kg, measured 810 mm at shoulder height, and 1400 mm of total body length. Males are slightly larger than females, and Brazilian animals are relatively larger than the Argentinean and Uruguayan sub-species (Cabrera, 1943). Personal observations during this study suggested that Brazilian males may have smaller antlers than their Argentinian and Uruguayan counterparts. Few studies have been conducted on pampas deer habitat selection and food habits despite the fact that the species occupies different types of lowland grassland ecosystems, from the Atlantic Coast to the foothills of the Andes. Radio-collared pampas deer in the Brazilian "Cerrado" (savanna) used all types 91

PAGE 103

92 of vegetation, from the open grassland to the scrub (Leeuwenberg & Resende, 1994). Rodrigues (1996), using direct observations, found that forbs were the most frequent item consumed during feeding events in this ecosystem. Newgrowth of plants such as Manihot tripartita (Euphorbiaceae shrub), Anemopaegna glaucum, (Bignoniaceae vine), and Myrtaceae were important throughout the year. Flowers, browse, and grasses were of less relevance, and fruits were not observed being consumed (Rodrigues, 1996). Fecal analysis studies have been conducted in two regions of Argentina, including the coastal seasonally flooded grasslands of the Saborombom Bay (Merino, 1993) and the pampas of San Luis Province (Jackson & Giuletti, 1988), both areas of which have been altered by human activities. The first area comprised a marginal habitat at the estuary of La Plata River, and the second area was a cattle ranch dominated by pastures. Pampas deer consumed approximately equal amounts of grasses and other species in the Saborombom Bay. However, a single species of grass {Paspalum dilatatum) dominated (22%) the diet of the pampas deer in the coastal plain (Merino, 1993). Pampas deer predominantly consumed grasses (80% of the annual diet) in the plains of San Luis, browse comprised 6% of the diet, forbs 44%, and graminoids 48% (Merino, 1993). Pampas deer were considered selective grazers in San Luis, and dependent on the fresh green material of a few species (Jackson & Giuletti, 1988). Cool-season grasses were their main food from May to October. In spring

PAGE 104

93 and early summer, pampas deer selected Sorghastrum seeds, Poa ligulahs foliage, and new-growth of warm-season grasses. Summer grasses, such as Schizachyrium spp., were selected from January to May. Jackson and Giuletti (1988) suggested that grassland management was necessary to conserve pampas deer in San Luis, because pampas deer are almost exclusively grazers. Competition for food from introduced herbivores, such as cattle and horses, might endanger that small population of pampas deer. The fact that Rodrigues (1996) observed pampas deer selecting forbs, instead of the more abundant grasses in the pristine habitats of the Brazilian Cerrado, suggests that their counterparts in Argentina may be selective grazers due to reduced availability of dicots. Dicots are more rapidly fermented than monocots, and, therefore, produce metabolizable energy more quickly (Hudson, 1985). Small ruminants need to metabolize energy quickly because of their higher requirements of energy per unit body weight as compared to larger species (Hudson, 1985). On the other hand, quantification of feeding habits by direct observations, as used by Rodrigues (1996), may be misleading, considering the difficulty in identifying the species of plants eaten as well as the amount consumed. Also, direct observations tend to produce a biased sample of the population, because different individuals have distinct degrees of shyness, and accessible habitats tend to be sampled more frequently. Furthermore, feeding habits of a species

PAGE 105

94 are determined by the interaction of a number of factors such as teeth morphology, digestive anatomy, distribution of nutritious food resources in space and time, and metabolic rate (Hudson, 1985). Examples of small grazers exist in ruminant families, such as gazelles and pronghorns, but not in cervids. Pampas deer could be an exception. Thus, comparative studies using fecal analyses are required to complement and expand current knowledge on the feeding ecology of the pampas deer . Obiectives This study is the first to present data on feeding habits of pampas deer in the Pantanal. It focused on determining if pampas deer are grazers as suggested by Jackson and Giuletti (1988), or forb consumers as suggested by Rodrigues (1996). If pampas deer are grazers in the Pantanal, then: Ho1 . The largest proportion of their diet consists of grasses. On the other hand, if pampas deer prefer forbs, then; Ho2. The largest proportion of their diet consists of forbs. Methods Fecal Sampling and Analyses Accurate data collection is difficult in studies of herbivore diets in the wild. Field research based exclusively on direct observations are subject to bias such as: a) sampling an adequate number of individuals when animals are not habituated to the presence of humans; b) different habitats do not have equivalent visibility; c) accuracy in identification of specimens eaten and

PAGE 106

95 frequency of bites decrease with distance, especially if the vegetation is very diverse; and d) bite size depends on the plant species and part of the plant selected. However, direct observations can provide complementary information to fecal analyses. In particular, they help to identify those items that are completely digestible or that leave unrecognizable fragments. This chapter presents findings on the diet of pampas deer derived from microhistological analysis of fecal material, and supplements those findings with information derived from direct observations of foraging. Five samples of pampas deer pellets were collected in the first week of each month, and preserved on 75% ethanol in individual plastic vials. Samples were pooled by season: rainy (November, December, and January); flood (March, April, and May); dry (July, August, and September). Diet analyses procedures followed the same technique employed for cattle and brown brocket deer (Chapters 5 and 6). The frequency data generated for each identified species was then converted to the percentage composition of the diet by season (Johnson, 1982). Differences of percentage use of plant species, families, and food categories, among and within seasons, were tested for significance with KruskalWallis test (Sokal & Rohlf, 1969). Differences of percentage use between pairs of items were compared with Mann-Whitney U-test (Sokal & Rohlf, 1969). Hierarchy of consumption was verified for each item analyzed with STP MannWhitney pairwise tests, which control the comparison-wise Type I error rate

PAGE 107

96 (Sokal & Rohlf, 1969). Correlations of diets between seasons were performed with the Spearman Rank of Correlation Index (Sokal & Rohlf, 1969). Trophic diversity was expressed by the Shannon-Weaner Index, and niche breadth by Levins' method, as standardized by Hurlbert (Krebs, 1989). Habitat Use Observations of pampas deer were obtained during daylight hours by excursions on foot and by vehicle in all vegetation types. Additionally, a radiocollared adult male was monitored (November 1991 -February 1992) to provide an equally distributed sample of habitat use during daylight hours. Ad libitum observations also were included. Observations were analyzed only for plant formations used during foraging or traveling events. Thirthy-minute intervals were employed for the quantification of plant formation utilization when following an individual, to ensure independence of data (Swihart & Slade, 1985). The time interval was sufficient to allow individuals to move across more than one type of plant formation, although there were no data on the home range size of pampas deer to verify the adequacy of the procedure. The radio-collared male used an area of 700 ha during a period of 4 months. Results Diet Composition A total of 103 plant species, distributed among 32 families, constituted the diet of pampas deer for the three seasons combined in the Pantanal Study Area

PAGE 108

97 (Appendix B). Six or seven key species were important seasonally (Table 7-1 ), but percentage contribution was variable among fecal samples. Therefore, there was no distinct dominance among the five most represented species in rainy and flood seasons: Ludwigia longifolia, Melochia simplex, Melochia villosa, Mesosetum chaseae, and Vernonia scabra (KW, P > 0.12). In contrast, M. simplex was the dominant species consumed by pampas deer in the dry season (STP, P < 0.01). However, percentage consumption of some species varied from season to season. The consumption of M. simplex, Hydrolea spinosa, and Sida santamarensis was high during the dry season, but declined in the flood season. Aeschynomene fluminensis, Vernonia scabra (Compositae shrub), and L. longifolia (Onagraceae shrub) were consumed heavily during the rainy and flood seasons. Mesosetum chaseae was important only in the rainy season, and Pontederia sp. and Eichhornia azurae only in the flood season. Few species were consumed abundantly during two consecutive seasons. Two shrubs {V. scabra and L. longifolia) and one forb {M. villosa) exhibited high rates of consumption during both the rainy and the flood seasons (Table 7-1 ). Three forbs were consumed in similar proportions during the rainy and the dry seasons (Hyptis sp., M. simplex, and M. villosa). M. villosa was the only important overlapping species between the flood and the dry seasons, and, therefore, diet was significantly different among these two seasons. The most important families in the diet of pampas deer were Sterculiaceae (33.4%), followed by Poaceae (11.2%), and Onagraceae (10.7%).

PAGE 109

98 Table 7-1 . Percentage composition (>5%) of species in the diet of pampas deer by fecal analysis, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition'' Family Species Rainy Flood Dry Average Poaceae Mesosetum chaseae 16.16 0.77 0.32 5.75 Compositae Vernonia scabra 5.19 4.28 2.62 4.03 Euphorbiaceae Caperonia castaneifolia 2.68 0.00 6.05 2.91 Hydrophyllaceae Hydrolea spinosa 1.81 0.62 9.96 4.13 Lamlaceae Hyptis spp. 5.43 2.14 5.26 4.28 Lythraceae Cuphea sp. 0.15 5.91 0.00 2.02 Malvaceae Sida santamarensis 0.52 0.07 6.47 2.35 Onagraceae Ludwigia longifolia 8.56 24.69 0.00 11.08 Pontederiaceae Eichhornia azurae 0.00 16.97 0.00 5.66 Sterculiaceae Melochia simplex 17.39 0.72 43.78 20.63 Melochia villosa 9.61 12.90 3.39 8.63 Other Species 32.5 30.93 22.15 28.53 3.08 2.82 2.39 Ba" 0.11 0.07 0.04 ^ Shannon-Weaner Index of diversity (all species included). " Niche Breadth Index for all species in the diet (standardized Levins). " See appendices for statistical tests comparing percentages.

PAGE 110

99 Sterculiaceae alone comprised more than 50% of the diet during the dry season (Table 7-2). Its percentage contribution was reduced during the flood season (KW, P = 0.01). The consumption of Poaceae was significant during the rainy season, and shared with Sterculiaceae the rank of the most consumed families of plants in that season (STP, P < 0.05). Onagraceae contributed significantly only during the flood season (STP, P < 0.03). Pontederiaceae was consumed heavily during the flood season, but its overall consumption across seasons was not as prominent as Onagraceae and Sterculiaceae. In general, the percentage consumption of food categories obtained with direct observations and fecal analyses were correlated with regards to the order of importance for each food category within seasons (Table 7-3). Significant correlations were obtained for the rainy season (Rs = 0.79, P = 0.02) and dry season (Rs = 0.90, P < 0.01). However, each method produced different percentages of consumption within food categories, which were more evident in the rainy and flood seasons than in the dry season. Seasonal changes in vegetation structure and flood level may have accounted for differential degrees of difficulty in locating pampas deer and visualizing their foraging activities. Thus, browse consumption was under-estimated by direct observations in the rainy season, and over-estimated in the dry season, whereas small forbs were over-estimated in the rainy and flood seasons (Table 7-4). Hydrophytic broadleaf forbs also were highly under-estimated by direct observations in the flood season.

PAGE 111

100 Table 7-2 Number of species consumed and percentage use (>5%) of plant families in the diet of pampas deer by season, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition of Diet Rainy Flood Dry Year Plant Family n^ %" n % n % r %^ Poaceae 16 23.7 14 6.4 10 2.8 16 11.0 Compositae 3 5.5 3 4.4 2 4.9 5 4.9 Euphorbiaceae 6 6.5 6 6.3 4 7.2 7 6.7 Fabaceae 7 5.3 7 2.3 3 3.6 9 3.7 Hydrophyllaceae* 1 1.8 1 0.6 1 10.0 1 4.1 Lythraceae 1 0.1 1 5.9 1 0.0 1 2.0 Malvaceae* 1 0.5 3 0.5 2 6.6 3 2.5 Onagraceae* 2 8.7 2 25.2 0 0.0 2 11.3 Pontederiaceae* 1 0.0 2 21.9 1 0.0 2 7.3 Sterculiaceae* 7 30.6 6 16.2 7 51.9 7 32.9 Others 30 17.3 36 10.1 19 13.1 51 13.5 ^ Number of species identified. See appendices for statistical tests comparing percentages. " Total number of species identified. ^ Average percentage of use. Percentage consumption was significantly different among seasons (KruskalWallis, P < 0.05).

PAGE 112

101 Table 7-3. Comparison of pampas deer diet composition among seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Comparison rs^ n^ 95% CI P value Result Dry/Rainy 0.60 41 0.36 to 0.77 <0.001 Similar diets Dry/Flood 0.28 38 0.00 to 0.55 0.090 Different diets Rainy/Flood a <-> 1 0.58 52 0.36 to 0.74 < 0.001 Similar diets Spearman's Rank Correlation Index. Number of food items in comparisons. Table 7-4. Percentage composition of different food items in the pampas deer diet, Pantanal Study Area (Caiman Ranch), 1991-1992. Rainy Flood Dry Average Food Category FA'^ DO' FA DO FA DO FA (1040)' (752) (669) Graminoids" 26.25 18.94 7.18 6.25 3.79 4.04 12.49 Small Forbs* 39.34 66.25 23.87 51.33 79.06 55.31 47.42 Browse^* 29.75 7.79 43.90 35.77 14.23 36.47 29.29 Broad Leaf Forbs''* 1.68 1.15 23.33 0.00 2.07 2.24 9.03 Lianas 2.09 0.29 1.72 0.00 0.47 0.00 1.45 Grass Seeds 0.00 0.58 0.00 6.65 0.00 0.00 0.00 Flowers 0.00 0.00 0.00 0.00 0.28 1.64 0.18 Fruits 0.05 5.00 0.00 0.00 0.10 0.30 0.05 Unidentified 0.84 0.00 0.00 0.00 0.00 0.00 0.10 ^ Shrubs and tree seedlings. ^ Flowers and leaves of Alismataceae and Pontederiaceae. ^ Fecal analysis. " Direct observations. ^ Number of observations. Percentage consumption was significantly different among seasons (KruskalWallis, P < 0.05).

PAGE 113

102 Selection of food categories by pampas deer appeared to track seasonal availability of new-growth (Table 7-4). Pampas deer consumed graminoids, forbs, and browse in approximately equal proportions during the rainy season, when most of the vegetation is sprouting (STP, P > 0.12). The consumption of graminoids decreased in the flood season, and was replaced by the consumption of flowers and leaves of hydrophytic broad-leaf forbs. Pampas deer concentrated feeding on small forbs in the dry season (STP, P < 0.01), possibly due to the maturity of graminoid and browse leaves. Flowers of trees {Tabebuia spp., Bignoniaceae) were eaten mostly during the dry season, when they fall in abundance after a short life on the trees. In addition to the natural fall, the action of birds increased their local availability to terrestrial herbivores by dropping the flowers to the ground after feeding on the ovaries. Fruits of the Acrocomia totai palm were observed being eaten in the dry season. The few fruits available under >A. fofa/ trees were consumed when encountered by deer, but the large seeds (3-4 cm) were not swallowed. However, the small (< 1 cm) fruits of Psidium kennedyanum, that occurred in abundance at the edge of the tree islands, were consumed heavily during a short period in the rainy season. Other fruits, such as Genipa americana, were not observed being used by pampas deer, although they were available in the area. Seeds of grasses {Axonopus purpusii and Schyzachyrium microstachyum) were observed being consumed during the flood season.

PAGE 114

103 Habitat Use Pampas deer selected open habitats, avoiding the edge of the forest except during the flood season when the grasslands were inundated. Pampas deer never entered forest patches to forage or ruminate even during the flood (Table 7-5). Short Grass was the selected plant formation, used for foraging and rumination when not flooded. Pampas deer selected to rest in the open in contrast to brov^ brocket deer. Their escape behavior consisted of running in the open, and then stopping at some distance to view the potential predator, usually behind a tall grass tussock or shrub. Fawns were born just before the onset of the rainy season, and were usually hidden within tussocks of tall grass within the Short Grass formation. Moist Depression was used principally for foraging in the rainy and dry seasons, but avoided in the flood season. Scrub was avoided during the rainy season only. As the flood invaded the lower areas, pampas deer avoided the deepest plant formations (Moist Depression), and used the higher terrain covered by Scrub and Forest Edge with increased frequency. They also continued to use the Short Grass, that was moderately covered by water. Pampas deer selected the lower areas with Short Grass as the water level decreased at the end of the flood season. However, during the dry season they continued to use the Scrub according to availability (Table 7-5). Therefore, assuming that the rainy season is the period of greatest abundance of food, the habitat constituted by Short

PAGE 115

104 Table 7-5. Plant formation selection of pampas deer in the Pantanal Study Area (Caiman Ranch), 1991-1992. Habitaf Season P value MD" SG" SC' FE' Rainy Flood Dry 26.61 14.35 26.04 <0.0001 0.0025 <0.0001 NS Avoided NS Selected Avoided NS NS Selected NS Avoided NS Avoided Avoided = plant formation used less than expected based on its availability; Selected = plant formation used more than expected based on its availability; NS = no selection, plant formation used in proportion to its availability (Bonferroni tests, P = 0.05). Moist Depression. " Short Grass. Scrub. ^ Forest Edge.

PAGE 116

105 Grass and the associated Moist Depression appeared to be the most suitable for pampas deer in the Pantanal. Discussion Fecal Analysis vs. Direct Observations The results of percentage consumption obtained by direct observations from a distance, indicated this method is not reliable for wild animals, especially if there is a great variance in accessibility of individuals among seasons. The method presents two biases. First, estimates of percentage consumption are not consistent among seasons; in some instances, percentages are either overestimated or under-estimated. Second, in seasons when there is difficulty in observing animals, as in the Pantanal Study Area, direct observations may bias the results of percentage consumption to the point of changing the rank of importance among different food items. A study of white-tailed deer foraging patterns corroborates the inaccuracy of the direct observation method (MullerSchwartze etal., 1982). Therefore, direct observations of wild animals from a distance are useful only in particular cases, such for preliminary qualitative descriptions of the dietary habits of a given species. Fecal analysis reduces these biases, because it offers a more homogeneous sample of the population's diet, permits unlimited sampling, and does not harm individuals or disrupt feeding events. However, fecal analysis has sometimes failed to accurately estimate percentage composition of diets fed to herbivores. Reasons for this inaccuracy are believed to include differential

PAGE 117

106 digestion of plant epidermis (Dearden etal., 1975; Holechek a/., 1982). For example, few cellular fragments remain after the digestion of buds and tender leaves of shrubs, which may result in an under-estimation of their relative contribution to the diet. Also, those species bearing hairs produce countless scores of loose trichomes, which tend to over-estimate the species in the diet. In contrast, grasses with their thick cellular walls permit a better preservation of fragments after digestion, and therefore a better estimate of species consumption. Development of correction equations based on in vitro or in vivo digestions can improve accuracy of the fecal analysis procedure (Dearden et al. 1975, Van Soest, 1982). In the present study, it is assumed that possible biases were not significant enough to invalidate the major conclusions of this research. However, future studies of pampas deer feeding habits should include correction factors for each of the commonest plant species consumed, to improve the accuracy of the method. Pampas Deer Diet in Relation to the Habitat Caracteristics The diet of pampas deer in the Pantanal Study Area falsified both hypotheses tested in this study. Pampas deer did not consume grasses or forbs preferentially year-round as expected. Instead, the analysis of the local rainfall regime and the related plant phenology pattern suggests that pampas deer selected for newgrowth despite the food category (graminoids, forbs, and browse). Therefore, pampas deer can be considered a concentrate selector (Hofmann, 1988) with a limited capability of fermenting monocot cell-wall.

PAGE 118

107 Although its gut anatomy has not been described yet, pampas deer must have a digestive system equivalent to cervids, and perhaps bovids, of similar diets. This gastric system is characterized by a fast passage rate, a relatively small ruminoreticulum, and a large caecum, which is adapted to a diet based mostly on plant cell contents (Demment & Van Soest, 1985; Hofmann, 1988). This research and another phenological study in the Pantanal indicated that regrowth for grasses and browse occurs in the rainy season (Schaller, 1983). Accordingly, pampas deer fed on similar proportions of graminoids, forbs, and browse during the rainy season. In the flood season, graminoids probably became more fibrous, and therefore, less nutritious (Hudson, 1985). During this season the consumption of graminoids diminished, and was replaced by the newgrowth of hydrophytic broad-leaf forbs (Alismataceae and Pontederiaceae). Hydrophytic shrubs (Onagraceae), that were sprouting during this season, were also consumed intensively. In contrast, pampas deer limited their foraging activities mainly to patches of Moist Depression duhng the dry season. These patches maintain enough moisture to allow newgrov^h of hydrophytic forbs, e.g., Melochia simplex (Sterculiaceae), and hydrophytic shrubs, e.g., Aeschynomene fluminenese (Fabaceae), even during the peak of the rainless season. Phenological patterns of vegetation in other areas where pampas deer have been studied appear to support the concept of the selectivity for newgrowth and easily digestible plant matehal. In Argentina, grass species have

PAGE 119

108 two growing seasons, which matched the seasonal foraging preferences of pampas deer (Jackson & Giuletti, 1988). Accordingly, Jackson and Giuletti (1988) considered the pampas deer to be a selective grazer. The fact that pampas deer in San Luis survived almost exclusively on a diet of grasses is probably possible because temperate grasses are in general significantly more digestible than tropical grasses (Demment & Van Soest, 1985). Temperate grasses also have equivalent percentage of cell-wall than leaves of trees and shrubs, but do not have as much lignin and secondary compounds as the latter, and therefore, they are potentially most digestible and palatable (Demment & Van Soest, 1985). However, this does not mean that pampas deer always are preferentially grazers in temperate regions. Instead, pampas deer should select for plants with higher protoplasmic protein, essential micronutrients, and carbohydrates, because, as concentrate selectors, they extract energy primarily from cell contents. In the Cerrado, where the scrub landscape is dominated by C4 rough grasses of reduced digestibility, pampas deer avoided consuming grasses yearround (Rodrigues, 1996). Instead, pampas deer selected forbs, which are chiefly found in the open grasslands along river watersheds (Rodrigues, 1996). Available literature (Leeuwenberg & Resende, 1994) coupled with personal observations of pampas deer in different regions of the Cerrado, confirm that the species selects open and moist areas in which to forage in the Cerrado. Furthermore, large concentrations of pampas deer were observed during the

PAGE 120

109 preliminary phases of this research in two burned areas of the Cerrado, which exhibited abundant and nutritious sprouting of browse and grasses. Fire is known to improve the quality of forage in the poor and aluminum toxic soils of the Cerrado (Coutinho, 1982, 1990). Key species in the diet of pampas deer in the Pantanal Study Area were: Vernonia scabra, Melochia simplex, Ludwigia spp., Hydrolea spinosa, Hyptis spp., Aeschynomene fluminensis, and Caperonia castaneifolia. All of these species were more abundant in Moist Depression, except for Vernonia scabra, as verified by the phenological study in the Pantanal Study Area. Melocliia simplex was particularly important during the dry season, when this plant species alone comprised 43.8% of the diet. Therefore, Moist Depression is fundamental for the survival of pampas deer throughout the dry season, which appears to be the most critical season in terms of nutritional limitation. Supportingly, cattle and capybaras have been reported to lose weight during the dry season in the Pantanal (Schaller & Crawshaw, 1981 ; Pott et ai, 1987). This seasonal nutritional bottle-neck may explain why pampas deer in Brazil show seasonal reproduction, whereas in Argentina and Uruguay they reproduce throughout the year (Jackson & Langguth, 1987; Redford, 1987; Pinder, 1992; Merino etai, 1997). Breeding seasons must be defined sharply in temperate and arctic environments to ensure that young are born at a time of year that facilitates prime lactation of the dam, and provides time for young to mature sufficiently before their first winter (Hudson, 1985). The same appears to

PAGE 121

110 occur in the Brazilian savannas, where rainfall instead of temperature is the environmental variable regulating food availability and nutritional quality, and ultimately, animal abundance. The fawning season for pampas deer in the Brazilian Cerrado and in the Pantanal coincides with the onset of rains (Redford, 1987, Pinder, 1992). Other species of tropical ungulates also have demonstrated breeding cycles closely associated with peaks of forage production (Phillipson, 1975: Mishra & Wemmer, 1987). Different ungulate species within the same area may have seasonal peaks of breeding, or be completely aseasonal, depending on the selected habitats and forage (Mishra & Wemmer, 1 987). The bimodal plant phenology in Argentina and Uruguay presumably allows a more balanced availability of nutritious forage throughout the year, and consequently, better environmental conditions for ensuring the survival of pampas deer fawns. Population densities of several large ungulates in the Serengeti appear to be regulated by food supplies (Sinclair, 1974; Sinclair et al., 1985;. A poor nutritional condition of females during pregnancy is reported to strongly decrease fawn survival and possibly future reproductive success of females (Murphey & Coates, 1966; Ofdetal, 1985). There is evidence to implicate insufficient food resources as cause of mortality in some situations, although many malnourished animals probably succumb to disease or predation (Crawley, 1983; Keith, 1983). The dry season is particularly important for survival of individuals and recruitment of fawns in the Pantanal, because availability and quality of food

PAGE 122

1 111 may be reduced for pampas deer during this period. Additionally, male pampas deer are growing antlers during the dry season, and females are pregnant (Pinder, 1992), which require adequate nutrition. Needed energy must largely be obtained from increased food intake, although mobilization of body reserves may contribute. Both, the growing fetus and the suckling neonate exert substantial energy demands on the mother (Ofdetal, 1985). If extrapolation from domestic ruminants is valid, one might expect pregnant pampas deer to require at least a 1/3 increase in energy requirements during late gestation, and a two-fold rise over maintenence energy needs during intense lactation (Ofdetal, 1985). Comprehensive studies of the levels of nutrients for forage species, other than grasses, are lacking for the Pantanal. However, most of the plant species studied, which have high levels of Ca, Mg, P, K, Cu, Zn, and/or crude protein (Pott & Pott, 1987) are present in the pampas deer diet, e.g., Vernonia scabra, Chomelia obtusa, Smilax sp., Helicteres guazumaefolia, Cyperus sp., Byrsonima orbignyana, and Fimbrystilis dichotoma. Of these species, Vernonia scabra has the highest value of crude protein (Pott & Pott, 1987), which could explain its status as a key species in the diet of pampas deer. This observation corroborates the hypothesis that pampas deer select nutritious forage. Therefore, an extensive study of forage quality should be conducted for several forage species in different seasons of the year, so that food resources can be better managed in benefit of cattle and deer species.

PAGE 123

112 Rangeland management practices to improve the availability and nutritional quality of forages for pampas deer should not conflict with livestock production, and if possible, benefit cattle as well because the largest income in the Pantanal ranches is derived from cattle ranching rather than tourism. On the other hand, an evaluation of the current practices for cattle management should consider their impact on wildlife before being implemented. For example, the plowing and seeding of the native pastures with introduced varieties of grasses, practiced in some portions of the Pantanal, should be avoided in Moist Depression formations, as these areas represent important food patches for pampas deer. A viable practice to improve the nutritional quality of forages in the floodplain would be a regime of prescribed fire. Small plots of Scrub isolated by fire lines could be burned in alternate years. Prescribed fire in November (growing-season) would be ideal, because it would mimic natural lightning-fire, and would improve nutrition for lactating females, which normally give birth in September/October (Pinder, 1992). Additional controlled burning in the dry season (August) might be experimented. The principal forbs consumed by pampas deer in this season would not be negatively affected because they grow in moist soils, but browse might be stimulated to sprout. Robbins and Myers (1992) speculate that plant growing-season fires promote the improvement of the herbaceous component of the ecosystem, whereas dormant-season fires favors browse production.

PAGE 124

113 Furthermore, the number of Moist Depression patches could be artificially increased in areas where very few pampas deer have become isolated by habitat fragmentation, to improve forage availability during the dry season. This could be accomplished by simply scraping off a shallow layer of soil, equivalent to the size of natural Moist Depressions on Short Grass and Scrub formations. The economic feasibility of such measure should be determined experimentally, but costs would be mostly limited to the consumption of fuel, because several ranches already possess the necessary machinery, normally used in the maintenance of roads. Outside the floodplain, where cattle management is more intensive, a few measures could help pampas deer to recolonize some of the areas where they disappeared due to the introduction of monospecific pastures; 1 ) maintenance of native vegetation along watersheds, which could provide natural diet and cover; 2) diversification of pastures by the introduction of small shrubs, such as Stylosanthes spp. (Fabaceae), that could also improve cattle's diet during the dry season; and 3) prescribed burning or mowing of selected areas. The key to the survival of the species is not only in the preservation of its habitat, but also in the search for alternatives of management for the currently modified habitats.

PAGE 125

CHAPTER 8 ECOLOGICAL SEPARATION AMONG CATTLE, PAMPAS AND BROWN BROCKET DEER Introduction Segregation along one or more niche dimensions facilitates partitioning of resources, and, thereby, ecological separation of species (MacArthur, 1972). Among sympatric ungulates, coexistence is presumably correlated with digestive anatomy and mouth dimensions (Hofmann, 1973; Owen-Smith, 1989; Gordon, 1989a; lllius & Gordon, 1991; Gross etal., 1996). On the other hand, the relative importance of the different mechanisms for niche partitioning remains controversial. Many studies of ungulates in the African savanna have indicated that coexistence is facilitated principally by dietary differences (plant growth stage and parts eaten) and to a lesser extent by the plant species per se, or by spatial and temporal differences in habitat use (Dunbar, 1978; Jarman & Sinclair, 1979; Hansen etal, 1985; Murray & Brown, 1993). Yet, simulation modeling, incorporating bioenergetic requirements of African ungulates, suggests that partitioning is achieved primarily through habitat segregation and plant parts favored (Owen-Smith, 1989). Thus, morphological differences in mouth dimensions and energy requirements lead distinct species of ungulates to 114

PAGE 126

115 achieve optimal foraging performance in those habitats where selected vegetation structures are abundant (Owen-Smith, 1989; Gordon, 1989a, b). Autoecological studies and community ecology studies on terrestrial herbivores of savannas in Africa, Asia, and Australia seem to fit the OwenSmith's model most appropriately. In all of these continents, species are primarily segregated by habitats, especially when food is limited (Dinerstein, 1987; Fox, 1989; Owen-Smith, 1989). South American forest ungulates also appear to segregate primarily by habitat types, although differences in food types (species) and categories (browse, fruits, and grass) consumed also are relevant (Branan etal., 1985; Bodmer, 1989). Likewise, in temperate South America, a study on the diet of the guanaco {Lama guanicoe) and the introduced red deer {Cervus elaphus) indicated similar patterns of segregation along the three dimensions (food types, food categories, and habitats), but inferred that differential habitat use was of greatest importance (Bahamonde et al, 1986j. Unlike African ungulates. South American herbivores appeared not to partition habitats as a consequence of the selection for different plant parts consumed. Rather, selection for distinct food categories and food species are more relevant than selection for food parts (Bahamonde etal, 1986; Bodmer, 1989j. Differences in patterns of partitioning may be a consequence of the diversity of herbivore species in each continent. South American ungulate communities have relatively few species (3-6), and a small range of body sizes (15-230 kg). In contrast, African savannas contain 15-20 species of ungulates.

PAGE 127

116 ranging from 3 kg to 6 tons. Thus, the greater the number of species using a same habitat, the greater the potential for diffuse competition and consequent pressure for further segregation along the food niche dimension. In South Amehca, the lack of competing species, allows ungulates to segregate fundamentally by habitats. Distinct degrees of niche partitioning also might be attributed to differences in diversity of vegetation (species and vegetation structure). Robinson (1986), studying capuchin monkeys in Venezuelan savannas, attributed the low number of primate species in the savanna to the low diversity of trees and reduced stratification of the forest. Forests with diversified strata and species of trees, as in the rainforests of the tropics, are able to support a greater number of arboreal and scansorial mammal species (Henry et al., 1989). Likewise, savannas with diversified vegetation types support larger number of terrestrial herbivores than ecosystems with lesser structural variations (Fox, 1989). To date, ungulate niche partitioning has not been studied in the Brazilian savannas. In this chapter, I verify the mechanisms of niche segregation among ruminants in the Pantanal Study Area, to address the question of whether ungulates in the Neotropical savanna follow the same pattern of partitioning observed in other ecosystems of the continent and globally. Specifically the hypothesis tested were:

PAGE 128

117 Hoi . Each of the studied ruminants is segregated from the other species by at least one type of resource: food species, food category, and habitat type (MacArthur, 1972). Ho2. Niche partitioning is attained primarily by habitat segregation, which is a consequence of optimal foraging for selected food items (Owen-Smith, 1989). Methods Density Estimates The TRANSAN computer program (Routledge & Fyfe, 1992) was used to estimate cervid density in the study area, as derived from data collected in 60km monthly road transects. Censuses were conducted at low speed (15 km/h) in early morning and late afternoon. Similarity and Niche Overlap Indices Methodologies for habitat use and diet analyses were described in the previous chapters. Similarity of diet between pairs of ungulates was calculated with the Percentage Similarity Index (Renkonen, 1938), which ranges from 0 (no similarity) to 100 (complete similarity). It is one of the best similarity coefficients available, because of its simplicity and its characteristic of being relatively unaffected by sample size and species diversity (Wolda, 1981; Krebs, 1989). The index is calculated as: P = E minimum (pn, pz,). where: P = Percentage similarity between samples 1 and 2.

PAGE 129

118 Pii = Percentage of plant species / in diet of ungulate species 1 . P2i = Percentage of species / in diet of ungulate species 2. Niche overlap among ruminant species was calculated with Horn's Index of Overlap (Horn, 1966), which suffers little bias due to variability in sample size and number of resources (Ricklefs & Lau, 1980; Smith & Zaret, 1982; Krebs, 1989). Horn's Index is calculated as: Ro = (I (Pi, + Pik) log(pj + pik) 1 pij log p/, 1 p* log p* ) / 2 log2 where: Ro = Horn's Index of overlap for species j and k. Pij = Proportion resource / is of the total resources utilized by species / p* = Proportion resource / is of the total resources utilized by species k. Plant Species Availabilitv and Selection Availability of plant species was estimated by surveying the vegetation, using a modification of the point-sampling method (Levy & Madden, 1933), as described in Chapter 4. This method is advantageous over "clipping and weighing" methods, because it is quicker and avoids the infra-structure necessary to dry and weigh matehal under field conditions. This method has been used in similar studies with native ungulates and cattle (Rajasekaran, 1988), and is considered especially efficient for short vegetation (Croker & Tiver, 1948; Mantovani & Martins, 1990). The preference for consumption of a plant species by deer and cattle was calculated by subtracting the proportion of the

PAGE 130

119 species available in the environment from the proportion of the species in the diet (Strauss, 1979). Friedman's test was used to verify consistency in use of food categories among the ungulates, and the STP test to verify which food categories were consumed the most by each ungulate species (Sokal & Rohlf, 1969). KruskalWallis test was employed in multi-seasonal comparisons within ungulate species for food categories (Sokal & Rohlf, 1969). The U-test verified significance of differences between consumption of food categories between seasons (Sokal & Rohlf, 1969). Results Densitv and Biomass of Cattle and Cervids A total of 40 pampas deer and 42 brown brocket deer sightings were recorded during the 1991/1992 census. The constraints on the detection function for pampas deer were: 1) maximum sighting distance of 150 m; 2) inflection point between 40% and 60%; 3) minimum tail height = 0.10% of shoulder height. Due to the lesser perceptibility, the constraints for brocket deer were: 1) maximum sighting distance of 140 m; 2) inflection point between 20% and 40%; 3) minimum tail height = 0.10% of shoulder height. Density of cattle was calculated based on the average stocking rate in the study area in 1991/1992. Cattle had the highest density among the three species of ruminants (36/km^), and almost the whole totality of biomass (8,640 kg/ km^) per unit of area (Table 8-1 ). Cattle densities increased during the peak of flood, when part

PAGE 131

120 Table 8-1 . Average density and biomass of cattle and cervids, Pantanal Study Area (Caiman Ranch), 1991-1992. Density Weight Biomass Species n/km^ % kg kg/km^ % Bos indicus 36.00 95.21 240 8,640.00 99.34 Blastocerus dichotomus 0.05 0.13 124 6.20 0.07 Ozotoceros bezoarticus 0.68 1.80 40^ 27.20 0.31 Mazama gouazoubira 1.08 2.86 22' 23.76 0.28 Total 37.81 100.00 302 8,510.96 100.0 ' Townsend, 1996. ' Schaller, 1983.

PAGE 132

121 of their foraging area was covered by > 1 0 cm of water. Pampas and brown brocket deer occurred in low densities in the Pantanal Study Area (0.68 and 1 .08/km^ respectively). The total cervid population estimated for the Pantanal Study Area was 205 (95% CI = 121-359) pampas deer and 412 (95% CI = 229653) brown brocket deer. Marsh deer {Blastocerus dichotomus) were rare in the Pantanal Study Area, and used habitats avoided by pampas and brown brocket deer, namely grassland swamps and the seasonally flooded CoperniciaTabebuia swamps. A survey by a fixed-wing aircraft (Cessna) yielded a density of 0.49 marsh deer/km^ of swamps and 0.05 individuals/km^ of flood plains (all vegetation types included), which was equivalent to a total of 23 marsh deer in the Pantanal Study Area. Dietarv Similarities Fecal analyses demonstrated that cattle, pampas deer, and brown brocket deer consumed the same food categories, but in different proportions (Fr, P < 0.01). The cumulative number of species consumed by each ruminant throughout the year indicated that 15 samples were adequate to perform the comparative analysis between and among species (Fig. 8-1 ). Cattle concentrated foraging on grasses, whereas pampas deer and brown brocket deer selected forbs and browse respectively (STP, P < 0.01) (Fig. 8-2). Similarities in consumption of food categories between cattle and the cervids were not significant, but were high between pampas and brocket deer.

PAGE 133

122 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fecal Samples —ocattle —Xpampas -obrocket Figure 8-1 . Relationship between cumulative number of plant species in fecal samples, as determined by microhistological analysis, and number of fecal samples, Pantanal Study Area (Caiman Ranch), 1991-1992.

PAGE 134

123 100 90 80 o70 Graminoid Forb Browse Fruit Cattle B Pampas deer B Brocket deer Figure 8-2. Mean percentage consumption of food categories by cattle, pampas deer, and brown brocket deer, Pantanal Study Area (Caiman Ranch), 19911992. Vertical bars indicate 95% confidence intervals on the mean percentage consumption.

PAGE 135

124 Similarities NA/ere especially high in the flood season, when pampas deer increased their consumption of browse, and brown brocket deer increased their consumption of forbs. However, a differential percentage consumption of the plant species common to both cervids resulted in low diet similarity (Table 8-2). Only three plant species were consumed simultaneously by pampas and brown brocket deer in quantities exceeding 6% within the same season: Melochia villosa (Sterculiaceae forb) in the flood season (12.9% and 8.3%, respectively); Caperonia castaneifolia (Euphorbiaceae forb) (6.0% and 8.4%, respectively); and Sida santamarensis (Malvaceae forb) in the dry season (6.5% and 8.5%, respectively). Habitat Use The percentage use of each plant formation differed significantly between pampas deer and brown brocket deer. Pampas deer used open vegetation (Moist Depression and Short Grass) more frequently, whereas brown brocket deer selected plant formations less affected by the flood, and, which had higher density of shrubs (Scrub and Forest Edge). Cattle had somewhat intermediary preferences, using the grasslands, especially the Scrub, but avoiding Moist Depression and Forest Edge (Fig. 8-3). As indicated by the frequency of use of different plant formations, pampas deer was the most prominent generalist of the 3 ungulates. In contrast, cattle probably a result of a patchy and scattered distribution of foods selected by pampas deer. 1

PAGE 136

125 Table 8-2. Percentage similarity indices for plant species and food categories consumed, and plant formations used seasonally by cattle, pampas deer, and brown brocket deer, Pantanal Study Area (Caiman Ranch), 1991-1992. Similarity of Food Species Consumption Cattle Pampas Deer Rainy Flood Dry Rainy Flood Dry Cattle Pampas deer Brown brocket deer 33.44 10.85 9.49 ~ 7.42 8.17 6.23 17.17 24.67 26.04 Similarity of Food Categories Consumption Cattle Pampas Deer Rainy Flood Dry Rainy Flood Dry Cattle Pampas deer Brown brocket deer 39.22 16.67 13.79 14.30 11.06 13.55 55.62 81.49 65.48 Similarity of Plant Formation Use Cattle Pampas Deer Rainy Flood Dry Rainy Flood Dry Cattle Pampas deer 75.91 53.92 68.03 Brown brocket deer 65.40 69.92 77.11 61.70 78.00 72.30

PAGE 137

126 0} CO Z) cn C o (1) Q. 70 60 50 40 30 20 10 4^ 0 Moist Short Grass Depression Scrub Habitat Category Forest Edge Cattle B Pampas deer S Brocket deer Figure 8-3. Mean percentage consumption of plant formations by cattle, pampas deer, and brown brocket deer, Pantanal Study Area (Caiman Ranch), 1991-1992 and brown brocket deer used Scrub with a higher proportion. This difference was I

PAGE 138

127 Use of formations by pampas deer was limited only during the flood season, when inundation almost totally covered vegetation in the Moist Depression (Fig. 8-4). Brown brocket deer, although preferentially selecting the Forest Edge and the Scrub, increased their use of other plant formations in the dry season, when newgrowth became scarce. This increase of use of open vegetation was probably correlated to the gradual increase of forbs in the diet, that occurred during the flood and dry seasons (Fig. 8-5). Cattle habitat use also was affected by flooding, and, thus, the seasonal availability of forage. Cattle used Scrub almost exclusively during the flood season, whereas the quick depletion of above-ground graminoids in the Short Grass caused a lower frequency of use of this plant formation during the dry season (Fig. 8-6). Niche Overlap and Partitionina Overlap indices for the three ungulates indicated that they were chiefly segregated from each other by the plant species they consumed, and to a lesser amount by food categories consumed, and finally by plant formations used. Dietary overlap was inversely related to the overlap in habitat use if the three species of ungulates are considered simultaneously in the analysis (Table 8-3). The more the three ungulates increased overlap in habitat use, the more they differed in diet. The largest overlap in diet occurred during the rainy season, when there was abundant newgrowth. Overlap in habitat use was minimal during this season. Cervids foraged more frequently beyond the selected habitats as newgrowth became less available, which increased their overlap in habitat use.

PAGE 139

128 50 -[ 40 m Moist Short Grass Scrub Forest Edge Depression Habitat Type Rainy B Flood Dry Figure 8-4. Mean percentage consumption of plant formations by pampas deer, Pantanal Study Area (Caiman Ranch), 1991-1992. J

PAGE 140

129 60 50 ^ 40 B 30 20 10 0 Moist Depression Short Grass Scrub Habitat Type Q Rainy B Flood Dry Forest Edge Figure 8-5. Mean percentage consumption of plant formations by brown deer, Pantanal Study Area (Caiman Ranch), 1991-1992.

PAGE 141

130 90 Moist Short Grass Scrub Forest Edge Depression Habitat Type Q Rainy H Flood Dry Figure 8-6. Mean percentage consumption of plant formations by cattle, Pantanal Study Area (Caiman Ranch), 1991-1992. Table 8-3. Niche overlap indices for cattle, pampas deer, and bro\Mi brocket deer by season, Pantanal Study Area (Caiman Ranch), 1991-1992. Niche Overlap Index Season Plant Species Food Categories Plant Formation Rainy 0.83 1.50 2.14 Flood 0.73 1.32 2.16 Dry 0.66 1.34 2.20

PAGE 142

131 Inundation also contributed to habitat-use overlap, because cattle and pampas deer used higher terrain (Scrub) more frequently. Indices of overlap for the top plant species consumed by cattle and cervids indicated that most species were consumed according to their availability (Table 8-4). Few species were simultaneously selected by two ungulate species within the same season. Melochia villosa (forb) was simultaneously selected by pampas and brown brocket deer in the flood season, and Sida santamarensis (forb) and Caperonia castaneifolia (forb) in the dry season. The three ungulates selected different plant species in the rainy season. Overall, among the grasses, cattle selected Brachiaria humidicola, Hymenachne amplexicaulis, Paspalum plicatum, Paspalum pontanalis, and one sedge Cyperus sesquiflorus. Some plant species selected by pampas deer were selected only during their grov^ng season. Mesosetum chaseae (grass) was selected in the rainy season, and Eichhornia azurae (hydrophytic forb) and Ludwigia longifolia (shrub) in the flood season. The majority of species selected during the dry season were associated with moist soils, such as Melochia simplex, Caperonia castaneifolia, and Hydrolea spinosa. Other selected species were Melochia villosa, Cuphea sp., and Sida santamarensis, which were associated with Forest Edge and Scrub most commonly (Table A-1). Brown brocket deer selected species associated mostly with Forest Edge or Scrub during the rainy and flood seasons: Sida santamarensis (forb), Melochia pyramidata (shrub), Melochia villosa (forb), Chomelia obtusa (shrub),

PAGE 143

132 Table 8-4. Seasonal food preference indices for cattle, pampas (O b.), and brown brocket deer (M.g.), Pantanal Study Area (Caiman Ranch), 1991-1992. Negative values indicate avoidance, positive values preference, and zero values no selection; values may range from -1 to +1 . rxall ly riooa ury n h ivi.g. n h M.g. KJ.D. M.g POACFAF Aynnnnn^ nnmi i^ii vj. 1 H -u. 1 u -U.UO -U. 1 o -U. ID U.UO n iR -U. ID nit; -U. 10 tsiacniana numiQicoia U.lt) U.UU U.UU U.OO U.UU 0.00 0.05 0.00 0.00 \jynoQOn uaciyion n no U.Uz U.UU U.UU 0.05 0.00 r\ r\r\ 0.00 0.03 0.00 0.00 riyrnenscnne U.U/ 0.01 0.00 0.05 0.01 0.00 0.03 0.01 0.00 ainpiexicauiis LtitiiHa nGXanuia U.UO U.U1 U.UU n no U.Uz r\ r\r\ U.UU U.UU 0.03 0.00 0.00 Mesosetum chaseae 0.02 0.17 0.00 0.01 0.01 0.00 0.05 0.00 0.00 Panicum laxum 0.01 -0.03 -0.04 -0.01 -0.04 -0.02 0.02 -0.04 -0.04 Paspalum plicatum 0.06 0.00 0.00 0.03 0.00 0.00 0.02 0.00 0.00 P. pontanalis 0.05 0.01 0.00 0.06 0.00 0.00 0.06 0.00 0.00 Reimarochloa 0.03 -0.02 -0.02 0.00 -0.01 -0.02 0.01 -0.02 -0.02 brasiliensis CYPERACEAE C. sesquiflorus 0.03 0.00 0.00 0.01 0.00 0.00 0.06 0.00 0.00 MALVACEAE Malvastrum sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 Sida santamarensis 0.01 0.01 0.11 0.01 0.00 0.12 0.00 0.06 0.08 STERCULIACEAE Bytneiia dentata 0.00 0.00 0.05 0.00 0.00 0.02 0.00 0.00 0.00 Melochia pyramidata 0.00 0.02 0.11 0.00 0.02 0.02 0.00 0.02 0.05 M. simplex -0.03 0.10 -0.03 -0.03 -0.03 -0.03 -0.03 0.40 -0.02 M. villosa 0.04 0.14 0.05 0.02 0.12 0.08 0.00 0.03 0.05 PONTEDERIACEAE Eichhomia azurea 0.00 0.00 0.00 0.00 0.17 0.00 0.00 0.00 0.00 RUBIACEAE Chomelia obtusa -0.03 -0.03 0.02 -0.03 -0.03 0.07 -0.03 -0.03 -0.03 Richardia grandiflora -0.03 -0.03 -0.01 -0.03 -0.03 -0.03 -0.03 -0.03 0.14 COMPOSITAE Vernonia scabra -0.05 0.01 0.07 -0.05 -0.01 0.11 -0.05 -0.02 -0.04 Wedelia brachycarpa 0.00 -0.01 0.01 0.00 -0.01 0.02 0.00 0.02 0.13 EUPHORBIACEAE Caperonia castaneifolia -0.01 0.01 0.00 -0.01 -0.01 0.00 -0.01 0.06 0.08 Euphorbia ttiymifolia 0.00 0.00 0.07 0.00 0.00 0.07 0.00 0.00 0.00 ONAGRACEAE Ludwigia longifolia 0.00 0.07 0.00 0.00 0.25 0.01 0.00 0.00 0.00 HYDROPHYLUCEAE Hydrolea spinosa -0.02 -0.01 -0.02 -0.02 -0.02 -0.02 -0.02 0.08 -0.02 LYTHRACEAE Cuphea sp 0.00 0.02 0.01 0.00 0.06 0.01 0.00 0.00 0.00

PAGE 144

133 and Vernonia scabra (shrub). The only selected plant species associated with open vegetation (Short Grass) during these seasons was a shrub, Euphorbia thymifolia (Table A-1 ). Similar to pampas deer, brown brocket deer selectedforbs associated with moist soils during the dry season: Richardia grandiflora, Wedelia brachycarpa, and Caperonia castaneifolia. Those plant species, however, were most commonly associated with tall vegetation, unlike the plants selected by pampas deer, except for Caperonia castaneifolia, that is associated with Moist Basin (Table A-1 ). Cattle and Deer Responses to Environmental Changes Cattle and cervids exhibited the greatest dietary diversities and niche breadths during the rainy season, which coincided with the season of surplus forage. Climate in the Pantanal affected cattle only during the flood season, when their diet was less diverse, possibly due to the inundation of the Short Grass formation. Cattle consumed nearly all above-ground graminoid biomass available in Short Grass during the dry season, when there is absence of newgrov^h, which was confirmed by the presence of bare soil among grass tussocks. Consequently, cattle selected to forage in Scrub, which harbors a higher biomass of grasses and browse. The inclusion of some browse species in the diet, as indicated by the percentages of consumption during the dry season, allowed cattle to maintain a dietary diversity comparable to the rainy season (Table 8-5).

PAGE 145

134 Table 8-5. Seasonal dietary diversity and niche breadth for cattle, pampas deer, and brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 19911992. Cattle Pampas Deer Brocket Deer H' Ba H' Ba Rainy 3.15 0.20 3.08 0.11 3.24 0.11 Flood 2.79 0.09 2.82 0.07 3.20 0.10 Dry ^ 3.31 0.18 2.39 0.04 3.20 0.09 Shannon-Weaner diversity index. ^ Levins' standardized niche breadth. Table 8-6. Percentages of food categories for cattle, pampas deer, and brown brocket deer by season, Pantanal Study Area (Caiman Ranch), 1991-1992. Cattle Pampas Deer Brocket Deer Rainy Flood Dry Rainy Flood Dry Rainy Flood Dry (%) (%) (%) (%) (%) (%) (%) (%) (%) 87.04 90.52 87.62 26.25 7.18 3.79 1.32 3.04 3.55 F" 10.00 7.00 4.41 41.02* 47.02 81.13 22.71* 34.12 47.13 2.96 2.48 5.60 31.84* 45.62 14.98 60.32 55.71 40.41 Fr'' 0.00 0.00 0.00 0.05 0.00 0.10 8.42 2.76 1.71 0.00 0.00 2.50 0.84 0.08 0.00 7.23 4.37 7.20 ^ Graminoids. ' Forbs. " Browse. Fruits. ^ Unidentified. Percentage consumption is significantly different among seasons (KruskalWallis, P < 0.05).

PAGE 146

135 Brown brocket deer managed to maintain dietary diversity and niche breadth throughout the year, by proportionally increasing the contribution of grassland forbs in the diet after the rainy season. As browse became mature, less digestible, and presumably richer in secondary compounds, brown brocket deer foraged away from the Forest Edge more frequently (Table 8-6). Thus, brown brocket deer adopted the strategy of becoming less specialized in habitat use to be able to maintain dietary diversity. Pampas deer, in contrast to brown brocket deer, were highly affected by the seasonal changes in food availability. Their dietary diversity and niche breadth became narrower when rising flood waters limited their foraging areas. Thus, pampas deer foraged mostly on hydrophytic forbs and browse during the flood season. Grasses were possibly too mature, and, therefore, less digestible to be consumed in quantity. The reduced availability of small forbs, normally encountered in Short Grass and Moist Depression, forced them to concentrate foraging on fewer plant species within these plant formations. During the dry season, pampas deer were essentially limited to feed mostly on Moist Depression, where they still found plant newgrowth. Consecutive hours of observation during this season indicated that pampas deer moved from one Moist Depression patch to the next, v^ere they fed on forbs. Moist Depression patches are found amidst Short Grass, and, to a lesser extent, also in the Scrub, what explains the high frequency of encounters of pampas deer in these two

PAGE 147

136 other plant formations during the dry season. Therefore, dry season was the period when pampas deer diet was the least diverse, and the most specialized. Discussion Resource Partitioning Ungulates in the Pantanal Study Area partitioned resources for all three dimensions (plant species, plant categories, and habitat type) investigated, and their feeding strategies indicated that there was a relationship between habitat selected and diet (Owen-Smith, 1989; Gordon, 1989a). The ruminants in the Pantanal Study Area selected different plant species, and those species were most likely to occur in those habitats selected by each ruminant (Chap. 4). The major species in the diet of pampas deer grew mainly in the closely related Moist Depression and Short Grass formations, whereas the principal species for brown brocket deer were most prominent in Scrub and Forest Edge formations. Cattle selected Short Grass in the rainy season and Scrub thereafter, presumably in response to the relative changes in availability and quality of grasses in those formations. This assumption is supported by the observation that cattle in the Isle of Rhum fed on communities containing the highly digestible live mesotrophic graminoids and forbs when these were abundant, and on communities containing live oligothrophic graminoids when the former declined in abundance. Despite the fact that dietary overlap decreased as habitat overlap increased among the three ungulates, there was no evidence that current

PAGE 148

137 competition was responsible for such segregation. Competition theory would predict that, when resources are most limiting, these cervids should diverge in feeding characteristics because of competitive effects (Smith etal., 1978). In contrast, the dyad pampas deer/brown brocket deer increased simultaneously habitat and dietary similarities, from rainy to dry season, in a way comparable to that observed for some ungulates in the African savanna (Jarman, 1971; Dunbar, 1978). In those studies, ungulate species with low diet overlap during the rainy season increased overlap in the dry season, but species with generally high diet overlap in the rainy season decreased habitat overlap in the dry season, when food availability is presumably reduced. Dunbar (1978) suggested that the simultaneous increase in diet and habitat similarities between those ungulates indicated that there was no interspecific competition. Their niche overlap was probably too low to require any further reduction during the dry season, v^ich appears to be the case with pampas deer and brown brocket deer in the Pantanal Study Area. Distinct nutritional requirements of each ruminant and their respective gut capabilities for processing plant material were more likely responsible for the segregation observed. The gut capacity of herbivores determines, in part, their capability for digestion of fiber (Demment & Van Soest, 1985; Hofmann, 1988). Large herbivores such cattle are able to survive on high fiber diets, because their larger digestive systems allows a lower passage rate necessary for the digestion of cell-wall, from where most of their energy is derived (Parra, 1978;

PAGE 149

138 Van Soest, 1982; lllius & Gordon, 1991). On the other hand, large herbivores are not expected to feed on a low fiber diet, because high quality food is rare and the amount of energy spent to search and consume these food items is prohibitive in comparison to the total daily energetic requirements of a large herbivore (Parra, 1978; Demment & Van Soest, 1985; Murray & Brown, 1993). In contrast, a small ruminant such the brown brocket deer requires less total energy, but more energy per unit of weight. A faster energetic return from digestion of plant material can be reached by the fermentation and digestion of cell contents instead of cell-wall (Van Soest, 1982). Consequently, small ruminants need a low-fiber diet, which can be digested directly by vertebrate enzymes or fermented rapidly by microbes. In contrast to graminoids, low-fiber tissue is found in leaves of trees and shrubs, which constituted the main diet of brown brocket deer in the Pantanal. Pampas deer, that weights twice as much as brown brocket deer, have presumably a larger gut capacity but require less energy relative to their body size compared to brown brocket deer. This allows a slov^r passage rate and better utilization of lower quality food than brown brocket deer. Owen-Smith's hypothesis (1989), however, does not completely explain habitat segregation among pampas and brown brocket deer. In the African savanna, most of the partitioning is achieved by selection of different plant growth stages despite extensive overlap in plant species consumed. If bioenergetic requirements were the only factor influencing habitat selection in

PAGE 150

139 the Pantanal, then pampas deer and brown brocket deer would presumably demonstrate a larger overlap in their consumption of plant species than was observed, because many species in the diet of brown brocket deer are also present in the diet of pampas deer and vice-versa. A complementary explanation for the pattern of segregation observed is that predation, past and/or present, influenced habitat choice by each cervld. Predation is a powerful force limiting the distribution of many animal populations to particular habitats. Dingoes, for example, can greatly depress kangoroo populations (Caughley et al., 1980). Many other vertebrate taxa also can limit prey species to particular habitats (Kettlewell, 1955; Wilbur etal, 1983; Robinson, 1985). Thus, different strategies of inherited behavior of predator avoidance adopted by pampas and brown brocket deer may have resulted in the observed habitat segregation. Pampas deer avoid predation by running through the open grassland, whereas brown brocket deer seeks refuge within thickets. Such limitation in use of space explains the little overlap in the consumption of plant species, because species were not equally distributed among the habitats (Chap. 4). Too much emphasis was placed on interspecific competition to explain current niche partitioning in the past (Cody & Diamond, 1975). However, there is little evidence of current competition among vertebrate herbivore species, possibly because most of the observed partitioning is most likely the result of past competition (Schroder & Rosenzweig, 1975). The few cases where

PAGE 151

140 interspecific competition seems to occur involve either the recent introduction of an exotic species into an indigenous herbivore community, or the introduction of an herbivore into a habitat different from that in which it normally exists (Fox, 1989; Wray, 1994; Putman, 1996). The Owen-Smith model (1989), on the other hand, showed that current partitioning is maintained by optimal foraging, which influences habitat selection by each species. Classification of the Cervid Feeding Strateoies Ruminants have been classified as concentrate selectors, grazers and roughage eaters, and intermediate types depending on their predominant diet and grastro-intestinal anatomy (Hofmann, 1988). Concentrate selectors, most commonly referred as browsers, include cervids whose body size range from animals as large as the moose to small bovids as the duiker {Cephalophus spp). In contrast, grazers are mostly bovids which range in size from the sheep to the African buffalo. Intermediate feeding types include representatives of cervids and bovids, which may shift diets between browse and grass depending on seasonal conditions (Homann, 1988). Bodmer (1990) expanded this classification by including frugivory as a feeding strategy distinct from the category of browsers. Frugivorous ruminants are found in tropical forests of South America, Africa, and Asia, where fruits are abundant but high quality browse for terrestrial herbivores is rare. According to this arrangement, ruminants can be classified along a continuum that ranges from fruit feeders through to browsers and to grazers.

PAGE 152

141 If we accept this classification, then brown brocket deer as a species should be classified in the frugivore-browser segment of the continuum, whereas pampas deer would occupy the browser-grazer segment. Both cervids are predominantly browsers and concentrate selectors, as inferred from their diet in the Pantanal and their body size, but their digestive characteristics and possibly the absence of many competitors allow them to consume a greater proportion of other food categories. This versatility is reflected in the diversity of habitats occupied currently by brown brocket deer and pampas deer. Brown brocket deer range from tropical forests to the savanna, and pampas deer from the savanna to the temperate grasslands of South America (Avila-Pires, 1959; Redford & Eisenberg, 1992; Pinder, 1997). This versatility may have been also responsible for their survival through the several climatic changes that eliminated a number of vegetarian species during the Pleistocene era (Stehli & Webb, 1985). Conservation Implications Data collected in this research indicated that cattle do not compete for food with pampas or brown brocket deer in the Pantanal. Also, it appears that cattle also do not compete with marsh deer. Although partially a grazer, marsh deer occupy the swamps, which are avoided by cattle. The Brazilian savannas, where these four ruminant species occur sympatricaily, do not harbor native ungulates that are principally dependent on grasses. The large herbivores are mostly browsers, and those v^ich have a graminoid component in their diet are, to some degree, dependent on the vegetation of wet soils, such as capybaras.

PAGE 153

142 pampas deer, and marsh deer. As pointed out by Redford and Fonseca (1986), the Cerrado, and for that matter also the Pantanal, has a fauna derived from the gallery forests more than from the grasslands. Herbivore species totally dependent on the grasslands were eliminated during successive Pleistocenic glaciations, when the savannas possibly became too limited in size and distribution to support large populations of grassland-specialized herbivores (Stehli & Webb, 1985). Species that were able to survive were those that fulfilled their nutritional needs by utilizing habitats other than grasslands. Thus, cattle are currently occupying a niche that has no parallel in the present native fauna of the Brazilian savannas. Although there are no unoccupied areas in the Pantanal, except for permanently flooded areas, interviews with old people in the Pantanal Study Area and other regions of the Pantanal indicated that cattle production activities have not decreased abundance and diversity of wildlife. However, it is evident the impact of cattle on the narrow forest patches, which become trampled and intensively browsed especially during the dry season. Consequently, tree seedlings are prevented from substituting old trees within these forests, which over time may negatively affect a number of vertebrate species that depend on trees for protection and food. Therefore, cattle densities should be maintained compatible with the ecological balance of the native vegetation to avoid indirect negative effects on wildlife, although there are no evidences of direct competition for food among cattle and other native herbivores in the Pantanal.

PAGE 154

143 Many, if not all, of the Brazilian national parks located in the savannas were appropriated from former cattle ranches. During many years, the national park service (IBAMA) has struggled to eliminate the common ranching practice of using the park lands for the grazing of cattle. However, cattle might, in fact, bring some benefits to parks in several ways. Cattle may help to recycle nutrients of the poor savanna soils, and decrease cumulative plant biomass that causes catastrophic fires. Savanna fires tend to be harmless to the fauna and flora when little plant biomass is available (Coutinho, 1982). However, trees and animals are likely to be killed by fires as was reported by Redford (1985) in Emas National Park, when the frequency of fires is reduced and plant biomass is high. Additionally, moderate browsing of shrubs during the dormant season would promote sprouting and, therefore, provide additional newgrowth for pampas and brown brocket deer during the dry period. Summer grazing by cattle was shown to increase the quality of grass swards available to feeding red deer hinds on the Isle of Rum (Gordon, 1988). Rhodes and Sharrow (1990) reported that sheep grazing appeared to improve forage quality in autumn and spring for white-tailed deer and wapiti in Oregon, and increased the availability of highquality forage in spring. Finally, cattle would represent an alternative source of protein to large cats, which sustain low population numbers due to the complete isolation of some of these parks. Despite some of the possible benefits, the presence of cattle may result in modifications to the local vegetation. Therefore, comprehensive studies should

PAGE 155

144 be conducted by the Brazilian agricultural research department (EMBRAPA) to regulate stocking rates for each season and individual area within the parks, to avoid overgrazing. Enforcement of recommendations should be carefully conducted by park rangers to avoid possible abuses. Unfortunatelly, the Brazilian park service does not have the resources and man-power to conduct not even an experiment on such matter. Currently, most urgent needs are the increase of staff in the parks and the allocation of resources to better protect and use these areas for public education and recreation. However, the current lack of ideal conditions to conduct experiments on cattle grazing within parks should not prevent careful initiatives in that direction. Dynamic changes in vegetation constitute natural cycles of savanna ecosystems, which show a great resilience to modifications. Large scale and long term alterations in savannas are more likely influenced by climatic changes and human intervention than by the action of herbivores. Norton-Griffiths (1979) concluded that fire had a more pernicious overall effect on vegetation modifications in the Serengeti savanna than elephants and giraffes, despite of their possible high local impact on portions of the woodlands. He argued that disruptive forces, such as fire and elephants, may function to provide random perturbations, thus maintaining a mosaic of communities of peak diversity, productivity, and resilience. He further suggested that the prevention of these perturbations, as they are enforced currently in Brazilian parks, could be highly detrimental to these ecosystems in the long run.

PAGE 156

CHAPTER 9 SUMMARY AND CONCLUSIONS The present research has provided new information on the classification of plant formations in the Pantanal, and on the ecology of two native cervids, the pampas deer and the brown brocket deer. Additionally, this study detemriined the dietary overlap among cervids and cattle, for the purpose of assessing the potential impacts of cattle ranching on the conservation of these native cervids. The taxonomic and structural diversity of the Pantanal vegetation has been addressed by several authors (Hoehne, 1923; Veloso, 1948; Adamoli, 1984; Eiten, 1985; Allen & Vails, 1987; Pott, 1988). However, the current work, the first to be based on a phytosociological study, expanded knowledge of Pantanal vegetation by presenting a classification of the diverse plant formations that constitute the floodplains. Previous descriptions of the Pantanal vegetation considered only the structural profile of the vegetation and the presence of dominant species. Despite the fact that these classifications are valid as a first approach to identify differences within the macro-scale of the Pantanal ecosystem, they do not provide similar information at a micro-scale. This study of native vegetation of the floodplain in the Pantanal Study Area, revealed six distinct formations that differed significantly from each other in terms of plant species composition and abundance: Marsh Ponds, Moist 145

PAGE 157

146 Basin, Short Grass, Tall Grass, Scrub, and Forest Edge. All these formations form a mosaic and can be easily identified visually. Their recognition is relevant to the future management of the native pastures, because cattle and cervids demonstrated evident and distinct preferences for some of these formations. The present research also provided new knowledge of the feeding habits of the brown brocket deer, which had not been studied previously in a savanna ecosystem. Earlier studies have been conducted in areas of continuous forests (Branan etal., 1985; Bodmer, 1989), and in the scrub forests of the Paraguayan Chaco (Stallings, 1984), where the greater proportion of the brown brocket deer diet is comprised of fruits. This study, however, revealed that brown brocket deer are principally browsers in savanna habitats. Fruits constituted only a small proportion of their year-round diet. Additionally, the diet of brown brocket deer in the Pantanal was in accordance with the finding that they fed mostly in the open, seeking the protection of the forest or thickets to ruminate or to escape predation. Another important finding of this research was that pampas deer is not a grazer species (Jackson & Giuletti, 1988). As other bovids and cervids of similar body size, pampas deer foraged for newgrowth leaves (Dunbar, 1978; Dinerstein, 1987). During the rainy season, forbs, browse, and graminoids were prominent in the diet of the species. In other seasons, when graminoids were mature, forbs and browse dominated their diets.

PAGE 158

147 This study provided, for the first time, a quantitative data on the diet of cattle occupying the native pastures of the Pantanal. Notable was the finding that cattle complemented their predominantly graminoid diet with several species of browse and forbs, possibly due to the low nutritional value of the graminoid forages, as was suspected by previous researchers (Brum et al., 1987; Pott et al., 1987). The current practice of introducing exotic grass species to the native pastures could produce negative benefits if the practice reduces or eliminates native forb and browse species. Better results might be achieved by improving the quality and abundance of native species preferred by cattle. Among the graminoid species consumed, cattle preferred Brachiaria humidicola, Hymenachne amplexicaulis, Paspalum plicatum, Paspalum pontanalis, and Cyperus sesquiflorus. Finally, this research demonstrated that there is no evidence of competition among cattle and cervids, or between pampas deer and brown brocket deer in the Pantanal Study Area. The three ruminants are clearly segregated by distinct diet and plant formation preferences. The greatest partitioning occurred at the level of plant species consumed, which is a reflection of the different types of plant formation preferred by each ungulate. Pampas deer was associated with open vegetation (Short Grass), and, therefore, forbs occurring in moist soils were the prominent component of their diet. Brown brocket deer preferred grasslands consisting of tall grasses and shrubs (Scrub), and, consequently, their diet included principally brovy^e and forbs, which

PAGE 159

148 occurred in greatest abundance in this type of habitat. Currently, niche segregation between these two cervids is so distinct that it even allowed a larger overlap in habitat use and diet during the dry season than during the rainy season, when food is abundant. Although no removal experiments have been conducted to confirm the lack of competition between these cervids, behavior of the two species suggests that even in the absence of one species, the other would maintain the same patterns of diet and plant formation selection, presumably because of their inherited anti-predator strategies. The lack of competition between cattle and cervids makes it possible to conserve the latter on private lands, provided that portions of native habitats are maintained. Tall grass tussocks and thickets of shrubs are important for pampas and brown brocket deer, respectively, to protect newborn fawns from weather and predation. Thus, when deforesting large tracts of land, or plowing the floodplain for seeding with introduced grass species, patches of natural habitat should be left in a mosaic resembling their distribution in the floodplain. Within already established pastures, the inverse process should be promoted, namely the introduction of nutritious forb and browse species, and tall grass for cover. Newgrowth is important for both species of deer, therefore, fire management in the savanna might help to increase food availability and quality for the cervids and cattle, especially in modified habitats. The revenues gained with tourism might perhaps balance the expenditures associated with management.

PAGE 160

149 Cattle, due to their exotic status, have been banned from most Brazilian national parks. However, except for risks of increasing probabilities of spreading diseases among native ungulates, the introduction of cattle to parklands might provide some ecological advantages. The South American savannas evolved in the presence of fire and a number of extinct grazer species from diverse families of mammals. Thus, the elimination of both fire and grazers from the national parks, as currently promoted by the administration, might result in modifications in the vegetation, vvhich may not be in the best interest for the conservation of some species of plants and animals. Cattle grazing on the parks may confer a few advantages such as decreasing the accumulated plant material, which causes damaging fires, accelerating the recycling of nuthents, increasing the potential prey for large cats, and increasing park revenues from leasing fees. At the moment, touhsm in the Brazilian national parks located in savanna ecosystems is too incipient to justify the elimination of the cattle on the grounds of aesthetics. Furthermore, the presence of cattle has not deterred the everincreasing tourism industry on private lands in the Pantanal. Therefore, cattle should be considered as an additional source of management and income for those parks located in savanna habitats. Public lands must provide more immediate benefits to people than just the preservation of samples of ecosystems, because there is an increasing demand for land and a risk of encrouchment of these areas if they are perceived as wasteland.

PAGE 161

150 For obvious reasons, the maintenance of cattle should be evaluated individually on each park and guided by sound management practices, so as to avoid unwanted results. A pilot study should be conducted prior to the introduction of cattle in each area to determine the possible negative effects of the measure. The tvvo most evident drawbacks would be the overstocking of cattle and the antagonism generated towards large cats preying upon cattle. To avoid the first problem, stocking rates should be continuously monitored and established by area and season. The latter problem could be solved by refunding cattle ranchers for losses caused by predators. Native areas are becoming increasingly reduced and/or altered due to human activities. The adaptability of the cervid species to a new environment may depend on their capability to adapt their diets. The simplification of the native pastures of the Pantanal, i.e., cultivation of a few exotic grass forage species to improve cattle productivity, may not be far in the future. Therefore, the versatility shown by both pampas and brown brocket deer should be considered in creating strategies of conservation for these species in areas where development is being promoted according to the modern market rules of intensive productivity.

PAGE 162

APPENDIX A LIST OF COMMON PLANT SPECIES IN THE FLOODPLAIN

PAGE 163

CD H CO CO uu UCO H o CO (D o o o o o 00 o o CN CO o o CO 00 CO CO CJ) 00 CD 1^ co in CN CN o CO in in ^ d d d d d d d d d d d d d d d d d d d d d o o CN O CO CN o CM O ^" d d d d d CO o CO o CO o 00 in CD oo CO T — o in CN d d d d d d d d d d d CO o o o CO o CO CN CN CO 00 d d d d d d d d d o T — o o o CN o T — 00 o o o o CO in c:> CN CD o in CD d d d d o d d d d d d d d d d d d d o q CO d CD in o o (O CN in CD o o d d d d d d d d CO CO CD o CO CO CM in CN o T — d d d d d d CO 00 o o c\i CM CN O CN O a> d d d d d d d d CD O d o d CD d d o d ^ E CD O TO " X O O < < < o .5 c (1) D < CO s03 9> © o c CD O 5 8 CD O o -Q CD ?5 CD (D c: -c (to c: 0) o CO 0) 0) ±; o Q. = to T3 -a c c c ._ c <<<<<< g T3 o c c < (to CD O QQ < < CO o 3 _ XJ (D ® E Q. O ICD O E SI D (D 0) CO T3 E ._ -C O O (O c T3 <
PAGE 164

153 O H CO CO < LU O CO IO h0) o o O G OO CD 00 CO TO CN O O O 00 Td d d CD CD d d CM o ^m^rocor^coT-i^ OOOOt-OCOOO-^ d d d OO CD 00 O CD ciciocicicicicicD CO o o d CD o CO o d o CO d o o CO CD O CD d d o o d d o d TCD <» TO CN O O C o c CD > c o (D O 0) O Q. Q. CO CL 3 oooooooboooooooocjod CD S CD CD CO ^ CD :3 o O CD ^ CD c: CD s . Q) CD Ci. CO CD CD O O % o CD CO .2 2 §• CO c: CO o CD CD O O CO C: CO 0) •Q CD CO -2 O .CD CD CO S $ O CD -C O O CD •2 o CO CO CD O CO -S CJ. o cs .S 5 .to CD o O O (J O O O CD E 5 CD ^ 0 CD CD CD CD CD O O CD CD > > > > C C o o o o CO !2 I " CO :3 CD ,CD .G (D CO 55 c: CD -Q 8 P CD c: .s £: CD . S Ci. CD (D ^ sli O O O 8c: o

PAGE 165

154 O C/5 CO LU O CO CD CD CO D O O 00 .0) C 0) CO CO q o in t — q cn CO 05 CO CD o CM h-CNJ CO ONI o q q q CN CO CO CO q q CM »^ — \ V ) K ' — \ — \ QJ CO CO CD CN q o CO q cn q ^JCO f — \ 1 — \ V t ^ — ^ f — \ O ^ — ^ o o o CO o o o r-~ o CM CO o d d d d d d CO o CN o CO o o O o o d d d d d d d d 00 CO o CO CJ) CO CD o CM O in o CM o CO o o CN d d d d d d d d d d d d d d o d o d CD To d d CD To o d d CD CM d o d CN O CD O CN O CD d CJ) O d CM o CD o o d d o d 00 o o d d O CM o o o q o d d d d o o o ID CO in o o CN CN o in o d d d d d d d d d d o CO o o o CJ5 c;) CD CN NO CO CN d d d d d d d d d d d 2 « 00 c ^ CD 0) 3 CD CO O £1 b a» c 0) Q. C3. Q. Q. CO >^ >, >. CD 03 N O ^ •= T3 O 3 o o c .= Q. CO ._ t r
PAGE 166

155 CO LU LL O CO T3 o o to C 0) CO -sr 1o o o d d CD oo o o d o o d d o o d => ® O) CD £ ~ c ro nj O iT (D O O O X CN in CD CM oo oo CO CO CD CO CNl o 00 CM 00 CM o CSJ CM o o o Q Q ri *— V * ^ (— i V— ^ V — ^ ^ f — s 00 o Q CD O O o o d d d d o CD 00 CM q o O O o o o V— CM 1^ 00 oo CM q q q o o O O o o o o o o Q Q c~i * — ^ (— i \ > CO CD o o o O o d d d d d d CD 00 o O O o q d d d d d d CN CM CN O o o q d d d d o CM CN o CO CM O CM O d d d d d ;s CO (3) O o O o o d d d d d d d C3) CNJ oo in T — CD CO CO CM 00 CN CM d d d d d d d d CD O o o o o d o d CD o o d d N T3 ca Q_ b o o) ^ CO E = ^ -n "O Q. E Q. XII (0 o E I CO i 5 :3 ^ £3 CO 2 £ CD qj (D CD CD CD .3 E .o O 3; •2 0) CD N CD :3 C3) CO CD 0) CD CO O CD P iS CO o 0) CD 3; I CO q5 o ll I: 3: c: -c o CD C CD I CD b ^ CD ° E CO Q. CL CL CL C .e>,>,>, x: I I X I CO c o p > CO CD O) « C ^ CD X CD sz I— (D CD O c 3 CD CD CD O CD .to -5 I o 83:' i i i to :S tD C CD .CO 3 CD O CO ^ ^ E m ^ O I ^ ^ CD CD .b: .C3 W 0) is o CO 0) p :3 CD .£ -> -J _j -S CD "5 SO c ^ CD CD CD o CD O L_ D >^ n 2 (D "o CD CD •2' 6) y S § c: TO ?D o ^ — S ~~ CO CD ^ u s 5

PAGE 167

156 IT) o D _C C o o < CD Io H CO CO < LU CD CO IO IO m 0) o o o CO (DCNJCNJCDh--Cr)CN0O-«OCOOCOOOOt-CMCO ddcDodcbdcDciioo CD TCO o o o o (D CJ5 CD CO CM O 00 11TCO CO oooooooooo o d CM o lO O CN O CNJ Td d d 00 o o lO o o o 0) 0) CO o d d 00 d CN d d CN CO d d CD c 0) o CD O E 0) .E .E i3 CD CO i i I :3 c: .CD I -Q CD CD Q. CD CO D Q. CO E o >, o z o E CD C O T3 O d Qc§ CD _ :C CD t X o o CO w c CD Q. CO CO:i § § s ii €^ CD CD CD x: CD e O is CJ o CD CD O -S O CO :3 c CD CO :3 CD CO O O >< .o CD O E DOC O 1= O CD CL Q. CL Q. Q. CO CO CO CD CD CD CL CL Q. :3 CD c: o CD CL E CO CL CO CD CL C3) C .!= C CO D CD CD CN d CD C O = > >. CD SZ CL Q. CL Q. 2 5* :3 .CO "cD C -2 c o CD CO CD . . . CL CL CL CD C Q. .fD .§ "3 fD tD 3 . CL c: 2

PAGE 168

157 (O "O" a> C c o O < O H CO CO < LU O CO IO 00 0 D o o CO 0) G OJ-^OOOO-^LOlDOCDCDlOOOOOCOCMOOOOOOOmTCNt-0'^000'^00^1-COCOO(350CNICOt-cOOOO dddcbdcDCDddddcDdcDdddddcDd CD CO TlO d d d CO rra rO O CNJ o d d d d d CN -sr CD CM CO CO eg o o CD o o d d d d d d CM CM o d d o CO CO O CD d d CM CO o CO O d d o o d d o d o o d d cn CM d CD o sjCD CD LO O CO CO O O d d d d d d d m lo CO d d CJJCOCNICOt-Oh-lOTOOCM-«-OC0OOO ddddddddd CO CO CM Td d To O CM d d o CM CO o d d d o CO cn d «CN CO o o o d d d C3> o o d d o CM 00 lO O o o o CM O CO O O CO o d d d d CO r«o o d d CO 1o o d d o o o CO d o T— d CO o d s o CD O d 00 CO s d o o CM CO «O d d ^ c CO « C3) ^ >^ "D T3 o 5 CO CC E c CO E E _ < CO CO CO CO CO

PAGE 169

158 O 3 C C o O I < n H CO CO LU o CO o IO I00 (D D O O <0 G c^ CD Q CO f — CO CO CO CO o o o CD o o o o o O CO d (D d d d d d d d d d d d d d d d CO CO o d d d d CD CO CO CN O q q q CO d d d d d CO CN 00 CO q q q CO o o o co CD d d d d d d d d d CM ID CD CO CO o o o o o o o o o O o o d d d d d d d d d d d d d d d d T >> o O CD o c: o 0) £3. CO CO tn — CD II o "CD H X CD II CO (0 CO (0 CO O o Id CO t: II :5 CD S H II ^ Si CO CO „ CO " 2 CO T3 .9 II u " o O CO CO S >< CD = II < (D O) T3 LU 0} D C o CL U_ II TO u_ II co" II CO Q. 2 2 O

PAGE 170

APPENDIX B UNGULATE DIET COMPOSITION

PAGE 171

Table B-1 Percentage composition of species in the diet of cattle, Pantanal Study Area (Caiman Ranch), 1991-1992. Family Percentage Composition Species Rainy Flood Dry Average Andropogon hypoginus 1.71 2.55 4.13 2.80 A. selloanus 1.87 1.05 1.73 1.55 Axonopus purpusiii 10.10 8.06 20.60 12.92 Brachiaria sp. 2.40 1.26 5.16 2.94 Cynodon dactylon 2.15 5.41 2.70 3.42 Hymenachne amplexicaulis 6.96 5.43 2.86 5.08 Ichnanthus procurrens 1.14 2.35 0.83 1.44 Leersia hexandra 5.07 2.51 2.79 3.46 Mesosetum chaseae 15.20 33.40 4.84 17.81 Panicum laxum 4.62 2.82 5.53 4.32 Paspalum acuminatum 0.66 0.29 1.15 0.70 P. lineare 0.38 0.30 0.21 0.30 P. plicatum 5.64 2.79 2.01 3.48 P. pontanalis 5.16 5.73 6.24 5.71 P. virgatum 3.87 2.21 1.78 2.62 Reimaroctiloa brasiliensis 5.23 1.99 3.20 3.47 S. microstachyum 0.00 0.18 0.00 0.06 Setaria geniculata 3.44 4.02 2.23 3.23 Sporobolus jacquemontii 1.92 2.35 3.08 2.45 Trachypogon sp. 2.46 1.27 0.35 1.36 Unidentified 0.00 0.00 1 68 0 56 Cyperus brevifolius 0.06 0.00 0.05 0.04 C. haspan 0.60 0.42 1.36 0.80 C. sesquiflorus 3.05 1.05 5.67 3.26 C. surinamensis 1.46 1.30 1.43 1.40 Eleoctiaris acutangula 0.82 0.45 0.57 0.61 E. elegans 0.12 0.50 0.27 0.30 E. minima 0.32 0.24 0.82 0.46 E. nodulosa 0.05 0.06 0.86 0.32 Frimbristylis dichotoma 0.19 0.37 0.29 0.28 Rhynchospora tenuis 0.35 0.06 0.84 0.42 Unidentified 0.00 0.00 1.11 0.37 Poaceae Cyperaceae 160

PAGE 172

Table B-1 . Continued, 2. 161 Percentage Composition Family z Species Rainy Flood Dry Average Bignoniaceae Tabebuia spp. 0.00 0.00 0.60 0.20 Boraginaceae Cordia glabrata 0.00 0.00 0.10 0.03 Compositae Mikania sp., Wedelia 1.02 0.19 0.07 0.11 brachycarpa Euphorbiaceae Croton sp., Euphorbia hirta, 0.16 0.06 1.54 0 59 Sebastiania hispida Fabaceae Aeschvnomene so 0.00 0 24 0 18 0 14 Desmodium so Senna tora Stvlosanthes acuminfitfi Lamiaceae HVDf/S SDD 2 16 2 08 . \J\J 1 96 Malvaceae Malvastrum so Pavonia 0 86 1 86 0 46 1 nfi angustifolia , Sida spp. Meliaceae Tiichilia stellato-tomentosa 0.00 0.72 0 60 0 44 Molluginaceae Mollugo verticillata 0.06 0 00 0 00 0 0? Onagraceae Ludwigia inclinata 0.08 0 10 0 00 n OR Polygonaceae Polygonum sp. 0.06 0 00 0 08 0 05 Pontederiaceae Pontederia sp. 0.07 0.16 0.13 0.12 Rubiaceae Diodia kuntzei., Randia 0.06 0.19 0 07 0 11 W.I 1 armata Sapindaceae Paullinea pinnata 0 00 0 14 0 00 \J . \J\J \J,\JyJ Smilacaceae Smilax so n 14 n nn u.uu U. 1 C3 U. 1 1 Solanaceae Solanum strigilatum 0.20 0.00 0.47 0.22 Sterculiaceae Bytneria sp., IHelicteres sp. 7.77 3.47 1.98 4.41 Melochia spp., Walteria sp. Ulmaceae Celtis pubescens 0.08 0.09 0.00 0.06 Verbenaceae Uppia sp., Stchytarpheta sp. 0.08 0.28 0.35 0.24 Vitaceae Cissus sicyioides 0.06 0.00 0.07 0.04 ni^ 0.14 0.00 4.68 1.93 fruits 0.00 0.00 0.10 0.03 H'" 3.15 2.79 3.31 Ba^ 0.20 0.09 0.18 ^ Plant species not identified. " Shannon-Weaner Index of diversity (all species included). Niche Breadth Index for all species in the diet (standardized Levins).

PAGE 173

162 Table B-2. Percentage composition of species in the diet of brown brocket deer, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition Family Species Rainy Flood Dry Average Poaceae Axonopus purpusiii 0.22 0.30 0.49 0.34 Cynodon dactylon 0.00 0.09 0.00 0.03 Hymenachne 0.40 0.31 0.00 0.27 amplexicaulis Ichnanthus procurrens 0.05 0.03 0.00 0.03 Leersia hexandra 0.00 0.16 0.25 0.14 Panicum laxum 0.04 1.52 0.00 0.52 Paspalum spp. 0.18 0.17 2.00 0.78 Reimarochloa brasiliensis 0.07 0.06 0.00 0.05 Setaria geniculata 0.11 0.15 0.00 0.09 Sporobolus jacquemontii 0.00 0.06 0.09 0.05 Trachypogon sp. 0.00 0.11 0.10 0.07 Cyperaceae C. surinamensis 0.15 0.00 0.43 0.19 Eleocharis acutangula 0.00 0.00 0.20 0.07 Fhmbristylis dichotoma 0.06 0.00 0.00 0.02 Rhynchospora tenuis 0.04 0.06 0.00 0.03 Acanthaceae Justicia laevilinguis 0.00 0.00 0.14 0.05 Alismataceae Echinodorus longiscapus 0.00 0.00 0.12 0.04 Amaranthaceae Alternanttiera tenella 0.00 0.04 0.00 0.01 Apocynaceae Forsteronia pubescens 0.18 0.13 0.00 0.10 Asclepiadaceae Metastelma berterianum 0.38 1.30 0.31 0.66 Bignoniaceae Tabebuia spp. 0.36 1.04 1.19 0.45 Boraginaceae Cordia glabrata 0.00 0.26 0.00 0.09 Heliotropium filiforme 0.00 0.12 0.00 0.04 Unidetermined 0.10 0.10 0.00 0.07 Compositae Arthropappus 0.24 0.10 0.16 0.17 angustifolius Baccliaris medullosa 0.06 0.00 0.00 0.02 Centraterum punctatum 0.17 0.32 0.20 0.23 Mikania sp. 0.11 0.00 0.04 0.05 Vernonia scabra 6.81 16.16 1.02 8.00 Wedelia brachycarpa 0.08 2.05 13.6 5.24 Erythroxylaceae Erythroxylum deciduum 0.00 0.06 0.00 0.02

PAGE 174

Table B-2. Continued, 2. 163 Percentage Composition Family Species Rainy Flood Dry Average Euphorbiaceae Acalypha communis 0.00 0.54 2.78 1.11 Caperonia castaneifolia 1.00 0.09 8.37 3.15 Croton spp. 1.51 1.52 0.68 1.32 Euptiorbia spp. 0.04 7.21 0.00 2.43 Sapium hassleirianum 0.00 0.06 0.09 0.05 Sebastiania hispida 1.02 0.93 0.76 0.90 Undetermined 1.45 0.60 0.46 0.84 Fabaceae Aesctiynomene spp. 0.79 0.20 0.15 0.38 Aractiis major 0.23 0.04 0.16 0.15 Bauhinia mollis 0.04 0.04 0.00 0.03 Clitoria rubiginosa 1.87 0.11 0.07 0.68 Desmodium spp. 0.31 0.30 1.21 0.61 Galactia glauscecens 0.00 0.06 0.00 0.02 Indigofera lespedezioides 0.00 0.00 1.03 0.34 Senna tora 0.36 0.00 0.00 0.12 Stylosanthes acuminata 0.00 0.10 0.00 0.03 Undetermined 0.23 0.41 0.39 0.34 Flacourticaceae Casearia aculeata 0.00 0.65 0.00 0.22 Hydrophyllaceae Hydrolea spinosa 0.04 0.00 0.00 0 01 Lamiaceae Hyptis spp. 0.33 1.69 0.00 0.67 Lythraceae Cuphea sp. 0.16 1.02 0.00 0.39 Malpighiaceae Byrsonima orbigniana 0.00 0.30 0.09 0 13 Malvaceae Malvastrum sp. 0.00 0.11 5.06 1.73 Sida santamarensis 12.21 11.52 8.48 10 74 Mel lastomataceae Tibouchina sp. 0.00 0.04 0.00 0 01 Menyanthaceae Nymphoides indica 0.00 0.06 0.00 0 02 Menispermaceae Odontocarya thamoides 0.00 0.06 0.00 0.02 Mimosaceae Mimosa spp. 0.52 0.33 0.50 0.45 Molluginaceae Mollugo verticillata 0.05 0.04 0.00 0.03 Moraceae Sorocea sprucei 0.00 0.68 0.30 0.33 Myrtaceae Psidium spp. 0.00 0.00 0.05 0.02 Onagraceae Ludwigia longifolia 0.05 0.67 0.04 0.25 Polygonaceae Coccoloba sp. 0.00 0.39 0.00 0.01 Polygonum sp. 0.00 0.00 0.27 0.09 Pontederiaceae Eichhornia azurae 0.04 0.15 0.00 0.06 Pontederia sp. 0.00 0.29 0.00 0.10

PAGE 175

Table B-2. Continued, 3. 164 Percentage Composition Family Species Rainy Flood Dry Average Portulacaceae Portulaca fluvialis 0.00 0.03 0.00 0.01 Rhamnaceae Rhamnidium elaeocarpum 0.66 0.17 0.00 0.28 Rubiaceae Chomelia obtusa 3.57 10.49 0.42 4.83 Chomelia pholiana 0.24 0.14 0.11 0.16 Diodia kuntzei 0.00 4.68 0.00 1.56 Genipa americana 0.35 1.69 0.00 0.68 Randia armata 0.07 0.89 0.00 0.32 Richardia grandiflora 3.16 0.82 17.5 7.15 Sapindaceae Paullinea pinnata 0.80 0.00 0.89 0.36 Scrophulariaceae Scoparia montevidensis 2.12 0.77 0.25 1.04 Smilacaceae Smilax spp. 1.24 0.56 0.51 0.79 Solanaceae Oestrum strigilatum 0.00 0.00 2.54 0.85 Sterculiaceae Bytneria dentata 14.17 2.35 0.67 5.73 Helicteres guazumaefolia 1.35 0.68 0.00 0.68 Melochia parvifolia 0.24 0.21 0.97 0.47 Melochia pyramidata 19.29 1.57 5.00 8.62 Melochia simplex 0.23 0.33 1.67 0.75 Melochia villosa 2.87 8.33 5.09 5.43 Waltheria americana 1.31 3.37 2.13 2.27 Ulmaceae Celtis pubescens 0.00 0.03 0.16 0.06 Umbeliferae Eryngium ebracteatum 0.09 0.34 0.00 0.14 Verbenaceae Uppia alba 0.07 0.07 0.00 0.05 Stachytarpheta 0.13 0.06 1.36 0.52 cayannensis Undetermined 0.20 0.19 0.41 0.27 Vitaceae Cissus sicyioides 0.00 0.92 0.11 0.34 Cissus spinosa 1.51 1.52 0.00 1.01 ni' 14.23 4.88 8.93 9.93 H'" 3.24 3.20 3.20 Ba'^ 0.11 0.10 0.09 ^ Plant species not identified. " Shannon-Weaner Index of diversity (all species included). Niche Breadth Index for all species in the diet (standardized Levins).

PAGE 176

165 Table B-3. Percentage composition of species in the diet of pampas deer, Pantanal Study Area (Caiman Ranch), 1991-1992. Percentage Composition Family Species Rainy Flood Dry Average Poaceae Andropogon hypogynus 0.61 0.09 0.30 0.33 A. selloanus 0.09 0.11 0.00 0.07 Axonopus purpusiii 1.63 0.91 0.20 0.91 Brachiaria humidicola 0.06 0.11 0.00 0.06 Cynodon dactylon 0.34 0.00 0.00 0.11 Hymenachne amplexicaulis 1.64 1.04 0.70 1.13 Ichnanthus procurrens 0.33 0.34 0.08 0.25 Leersia hexandra 0.92 0.30 0.25 0.49 Mesosetum chaseae 16.16 0.77 0.32 5.75 Panicum laxum 0.23 0.26 0.24 0.24 Paspalum spp. 1.10 0.87 0.28 0.87 Reimarochloa brasiliensis 0.10 0.35 0.00 0.15 Schizachynum 0.05 0.00 0.00 0.02 microstachyum Setaha geniculata 0.16 0.70 0.25 0.37 Sporobolus jacquemontii 0.21 0.08 0.00 0.09 Trachypogon sp. 0.05 0.17 0.16 0.13 Cyperaceae Cyperus spp. 1.57 0.32 0.49 0.79 Eleocharis spp. 0.50 0.27 0.52 0.43 Fhmbristylis dichotoma 0.26 0.00 0.00 0.09 Rhynchospora tenuis 0.00 0.15 0.00 0.05 Alismataceae Echinodorus longiscapus 1.51 1.42 0.00 0.98 Apocynaceae Forsteronia pubescens 0.00 0.02 0.00 0.01 Asclepiadaceae Metastelma berterianum 0.28 0.08 0.00 0.12 Bignoniaceae Tabebuia spp. 0.41 0.31 0.54 0.33 Undetermined 0.20 0.09 0.17 0.15 Boraginaceae Heliotropium filiforme 0.00 0.09 0.00 0.03 Cecropiaceae Cecropia sp. 0.14 0.05 0.14 0.11 Compositae Baccharis medullosa 0.25 0.00 0.00 0.08 Centraterum punctatum 0.05 0.05 0.00 0.03 Mikania sp. 0.00 0.12 0.00 0.04 Vernonia scabra 5.19 4.28 2.62 4.03 Wedelia brachycarpa 0.00 0.00 2.23 0.74

PAGE 177

Table B-3. Continued, 2. 166 Percentage Composition Family Species Rainy Flood Dry Average Euphorbiaceae Caperonia castaneifolia 2.68 0.00 6.05 2.91 Croton spp. 0.09 0.42 0.06 0.19 Euphorbia spp. 0.19 0.14 0.33 0.22 Phyllanthus lindbergii 0.80 0.06 0.00 0.29 Sapium hassleirianum 2.56 1.02 0.74 1.44 Sebastiania hispida 0.15 0.79 0.00 0.31 Undetermined 0.00 3.92 0.00 1.31 Fabaceae Aeschynomene spp. 4.61 1.60 3.30 3.17 Clitoria rubiginosa 0.19 0.00 0.00 0.06 Desmodium incanum 0.32 0.14 0.00 0.16 Senna tora 0.00 0.00 0.31 0.10 Stylosantlies acuminata 0.16 0.06 0.28 0.17 Undetermined 0.00 0.48 0.00 0.16 Hydrophyllaceae IHydrolea spinosa 1.81 0.62 9.96 4.13 Lamiaceae i-iyptis spp. 5.43 2.14 5.26 4.28 Lythraceae Cuphea sp. 0.15 5.91 0.00 2.02 Malpighiaceae Byrsonima orbigniana 2.35 1.58 0.19 1.37 Malvaceae Malvastrum sp. 0.00 0.08 0.16 0.08 Pavonia angustifolia 0.00 0.05 0.00 0.02 Sida spp. 0.52 0.07 6.47 2 36 Meliaceae Tricliilia stellato-tomentosa 0.04 0.00 0.20 0.08 Mellastomataceae Tibouctiina sp. 0.04 0.00 0.89 0.31 Mimosaceae Mimosa spp. 0.53 0.18 0.00 0 24 Molluginaceae Mollugo verticillata 0.63 0.05 0 71 0 46 Myrtaceae Psidium spp. 0.00 0.58 0 00 Nynphaeaceae Nynphaea sp. 0.17 0.00 0.00 0 06 Onagraceae Ludwigia inclinata 0.15 0.49 0.00 0.21 Ludwigia longifolia 8.56 24.70 0.00 11.08 Pontederlaceae Eichliornia azurae 0.00 17.00 0.00 5.72 Pontedeha sp. 0.00 4.93 1.97 2.30 Rubiaceae Chomelia obtusa 0.00 0.16 0.20 0.12 Genipa americana 0.00 0.75 0.00 0.25 Randia armata 0.10 0.11 0.38 0.20 Richardia grandiflora 0.00 0.06 0.00 0.02 Sapindaceae Paullinea pinnata 0.04 0.64 0.00 0.23 Scrophulariaceae Scopaha montevidensis 0.00 0.00 0.24 0.08

PAGE 178

Table B-3. Continued, 3. 167 Percentage Composition Family Species Rainy Flood Dry Average Smilacaceae Smilax spp. 0.23 0.00 0.07 0.10 Sterculiaceae Bytneha dentate 0.06 0.00 0.19 0.08 Helicteres guazumaefolia 0.26 0.03 1.98 0.76 Melochia parvifolia 0.96 0.12 0.70 0.59 Melochia pyramidata 2.31 2.35 1.73 2.13 Melochia simplex 17.39 0.72 43.80 20.63 Melochia villosa 9.61 12.90 3.39 8.64 Waltheha americana 0.05 0.10 0.16 0.10 Ulmaceae Celtis pubescens 0.05 0.00 0.00 0.02 Verbenaceae Lippia alba 0.97 0.30 0.00 0.42 Stachytarpheta 0.19 0.04 0.31 0.18 cayannensis Undetermined 0.06 0.00 0.00 0.02 Vitaceae Cissus sicyioides 0.64 0.80 0.27 0.57 Cissus spinosa 0.88 0.07 0.47 0.47 ni" 0.03 0.49 0.00 0.17 H'" 3.07 2.82 2.39 Ba'^ 0.11 0.07 0.04 ^ Plant species not identified. ^ Shannon-Weaner Index of diversity (all species included). " Niche Breadth Index for all species in the diet (standardized Levins).

PAGE 179

APPENDIX C ANTLER CYCLE AND REPRODUCTION

PAGE 180

Brown Brocket Deer and Pampas Deer Antler Cycles Pampas deer and brown brocket deer exhibited different phenologies in their antler cycles in the Pantanal Study Area (Fig. C-1 and C-2). Brown brocket deer were observed with polished antlers in all months of the year, and the pair of simple spiked antlers is not shed annually. In contrast, male pampas deer cast their antlers in May/June annually, and by September begin to shed the velvet from the new antlers. Male fawns born in the previous year start growing pedicels in February, which develop to spiked or bifurcated antlers in their first year. Three tines is the common pattern for adult pampas deer antlers, but two males were observed with asymmetrical antlers. An yearling had a pair of antlers consisting of a right unbranched stem and a left bifurcated antler. Another adult male had a right antler consisting of four tines. Reproduction Brown brocket deer fawns are born in any month of the year (Fig C-1 ). Although sample size was small, observations in 1991 and 1992 indicated that there was a peak of birth prior to the onset of rains, and, possibly, a secondary peak immediately after the flood season (Fig. C-3). Pampas deer, in contrast, had synchronized reproduction in the Pantanal Study Area. Females were observed being tended by males from January to March, and fawns were born from August to October at the end of the dry season. A few fawns also were born in May, after the flood season. This secondary peak is presumably consequence of an early interruption of lactation, 169

PAGE 181

170 Figure C-1 . Monthly percentages of antler stages for brown brocket deer, Pantanal Study Area (Caiman Ranch), 1991. The line shows the relative' frequencies of births. Total number of adult males = 77; total number of fawns = 19. Numbers represent individual deer and not individual observations.

PAGE 182

171 Figure C-2. Monthly percentages of antler stages for pampas deer, Pantanal Study Area (Caiman Ranch), 1991. The line shows the relative frequencies of births Total number of adult males = 98; total number of fawns = 21 . Numbers represent individual deer and not individual observations.

PAGE 183

172

PAGE 184

173 or a post-partum pregnancy, which is known to occur in other South American cervids including the marsh deer, red brocket deer, and brown brocket deer (Gardner, 1971; Chaplin, 1977; Fradrich, 1987). Few births have been observed in the Brazilian Cerrado, but the timing of fawning seem to be consistent with the population in the Pantanal Study Area (Redford, 1987). These results contrast with the reproductiive phenology in Argentinean populations of pampas deer, where fawning occurs mainly from September to November, but with births occurring also in other months of the year (Jackson & Langguth, 1987). A complicating factor in interpreting the reproduction of pampas deer in Argentina is that Jackson and Langguth (1987) combined data from different populations, each of which might have distinct reproductive profiles. Pampas deer from La Corona captive herd showed an extended antler cycle, which is consistent with a prolonged breeding season. However, wild pampas deer from Saborombon herd grew antlers during a short season (August to November), which is consistent with seasonal reproduction. Data on reproduction of pampas deer in the wild indicates that breeding cycles of pampas deer may be rather diverse due to local environmental conditions, but a greater number of populations need to be studied before any pattern can be established. Populations occupying habitats with seasons of food shortage maybe reproduce only when conditions are favorable (Richter & labisky, 1985), whereas populations in habitats with little variation in food supply maybe reproduce year-round .

PAGE 185

174 In the Pantanal, the strong seasonality of the climate certainly imposes selection against continuous reproduction in pampas deer. Fawns born during the flood season would have reduced concealment cover, which increase risk of predation by cursorial predators, caymans {Cayman crocodilus), and anacondas {Eunectes murinus). Anacondas and caymans are commonly seen on the grasslands at this time of the year. Likewise, nursing during the dry season, when food is limited, could be detrimental to the future reproductive success of the dam, and might increase her mortality risk. Milk production is known to impose a large energy and nutrient demand on the lactating female, which is far more demanding than pregnancy (Robbins etal., 1987). Fawning in late dry season and early rainy season, as observed in the Pantanal Study Area, allows acquisition of nutritious forage for lactating females, and for does and weaned fawns in late rainy season. Another period of newgrowth of graminoids and forbs also occurs in May (after the flood season), which allov^ good quality forage for the initial and most demanding period of lactation (Robbins etal., 1987). However, fawn development may be jeopardized by the reduced availability of food if they are weaned during the dry season. Fawn development require a great amount of energy and protein, specially for males (Clutton-Brock etal., 1982; Price & White, 1985), and, therefore, high rates of fawn mortality are expected to occur in the dry season, which would explain the smaller number of births recorded in May. Yet, females whose fawns die shortly after their birth would have a higher reproductive success by having a

PAGE 186

175 post-partum estrus than those females who lost their fawns and did not became pregnant until the next year. Also, birth in May could be result of an early-season reproduction by first-time breeder females. It has been documented that female fawns and first-time breeders in the wild conceive earlier or later than older females (Sadleir, 1987). Young and subordinate females would avoid competition with adult females for reproduction with dominant males if they matted earlier or later than mature females, which In turn could increase their reproductive success (Degayner & Jordan, 1987).

PAGE 187

APPENDIX D HOME RANGE SIZE

PAGE 188

Brown Brocket Deer Home Range An adult radio-collared brown brocket deer male provided data on locations and habitat use from April 1991 to November 1992. A total of 305 location/days encompassed an area of 0.97 km^ of the floodplain, that was constituted by a mosaic of all types of plant formations. Most of the locations were recorded within an area of 0.19 km^ (core area), despite the fact that different portions of the range were visited by the male in each season (Fig. D1). These seasonal outlier locations were presumably associated with mating excursions, because, in all instances, the male was observed either pursuing a female or in the vicinity of a female. In one instance, the male defecated on the fresh droppings of a female within minutes after she had left the area, which may represent some sort of communication between the sexes regarding their physiological condition (Muller-Schwarze, 1987). The frequent observation of seven identified males and seven identified females occupying areas adjacent to the collared male suggested that its home range size was typical for the species in the floodplain. Identified individuals were never observed very far from previous location sites. The lack of frequent observations of other adult males within the core area of the collared male suggested that there might be some spacing among adult individuals of the same sex. This hypothesis was supported by observations of other brown brocket deer. Additionally, an agnostic encounter between two males was observed, when the collared male was chased back to his own core area at full 177

PAGE 189

178 speed. In this instance, no female was observed near the two males, therefore, this aggressive behavior presumably was not related to mating. Captive adult males also tend to be aggressive, frequently attacking people that enter their pens (pers. observ.). In another instance, a young adult male was observed marking a shrub with his antlers, and then pawing the ground on the edge of a forest patch near the border of the home range of the collared male. Minutes after this male had left the area, the collared male arrived and marked the same shrub v^th his antlers, pavy^d the ground beneath, and finally defecated on the pawed area. Again, no female was involved in the contest. Interestingly, when this intruding male was a sub-adult in velvet, it visit the home range of the collared male without being challanged. Other sub-adult males and females also seem to be tolerated by adult brown brocket deer in other portions of the study area, as they were some times observed foraging not far away from each other. Dung piles were produced in the core area of the collared male. Other dung piles found outside his home range area were associated with areas where identified individuals were frequently located, which supported the hypothesis of spatial segregation among individuals. Dung piles are presumably related to stable and restricted home ranges, and are also produced by other small cervids such as Moschus and Muntiacus (Muller-Schwarze, 1987). Pampas Deer Home Range An adult and dominant male pampas deer was radio-collared in November 1991 and died in February 1992. The male had shown no signs of disease or injury, although his body condition appeared to be slightly debilitated

PAGE 190

179 disease or injury, although his body condition appeared to be slightly debilitated (ribs visible), after attending several females. Possibly, this decrease in body condition, related to his mating effort, predisposed him to predation. The 38 location/days obtained for this male yielded a four-month home range of 5.94 km^ Thirteen locations obtained during the flood season demonstrated that the deer did not move from his range because of water (Fig. D-2).

PAGE 191

180 1 1-9 locations 10-15 locations 16-20 locations 21-25 locations > 25 locations 250 m rainy season flood season dry season Figure -D1 . Adult male brown brocket deer home range, Pantanal Study Area (Caiman Ranch), Apr 1991 -Nov 1992.

PAGE 192

181 —i1— -— f — ir \ \ \ 1 : ! ^ -4— I 1 location 2 locations 3 locations 4 locations 5 locations 500 m rainy season flood season Figure D-2. Adult male pampas deer home range, Pantanal Study Area (Caiman Ranch), Nov 1991 -Feb 1992.

PAGE 193

APPENDIX E BONFERRONI TESTS FOR HABITAT USE AND AVAILABILITY

PAGE 194

Table E-1. Simultaneous Bonferroni confidence intervals for utilization of vegetation types by cattle, Pantanal Study Area (Caiman Ranch), 1991-1992. Vegetation type Expected Used Bonferroni intervals Results RAINY n' = 500 n = 1,531 Scrub 0.35 0.38 0.35 < P < 0.42 NS' Short grass 0.25 0.51 0.48 < P < 0.54 Selected Moist shallows 0.25 0.10 0.08
PAGE 195

184 Table E-2 Simultaneous Bonferroni confidence intervals for utilization of vegetation types by brown brocket deer, Pantanal Study Area (Caiman Ranch), 1991-1992. vegeiaiion type txpeciea useo Bonferroni intervals Kesuiis R AIMV n — ouu n — 1 /u OCI uu u.oo u.on 0.42
PAGE 196

185 Table E-3. Simultaneous Bonferroni confidence intervals for utilization of vegetation types by pampas deer, Pantanal Study Area (Caiman Ranch), 19911992. Vegetation type Expected Used Bonferroni intervals Results RAINY n' = 500 n = 440 Scrub 0.35 0.27 0.22 < P < 0.32 Avoided Short grass 0.25 0.39 0.33
PAGE 197

APPENDIX F KRUSKAL-WALLIS TESTS FOR DIET COMPOSITION

PAGE 198

Table F-1 . Kruskal-Wallis test for differences in percentage use of 10 grass species by cattle within seasons, Pantanal Study Area (Caiman Ranch), 19911992. Season df KW P value Rainy 9 29.56 < 0.001 Flood 9 28.74 < 0.001 Dry 9 23.48 0.005 Table F-2. Kruskal-Wallis test for differences in percentage use of grass species by cattle among three seasons, Pantanal Study Area (Caiman Ranch), 19911992. Season df KW P value A. purpusii 2 5.58 0.021 M. chaseae 2 7.76 0.061 P. pontanalis 2 0.18 0.914 Table F-3. Kruskal-Wallis test for differences in percentage use of five plant families by brown brocket deer within seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Season df KW P value Rainy 4 13.61 0.0087 Flood 4 2.81 0.5908 Dry 4 0.78 0.9417 187

PAGE 199

188 Table F-4. Kruskal-Wallis test for differences in percentage use of 10 plant species by brown brocket deer within seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Season df KW P value Rainy 9 36.94 < 0.001 Flood 9 13.91 0.126 Dry 9 19.66 0.020 Table F-5. Kruskal-Wallis test for differences in percentage use of food categories by brown brocket deer among three seasons, Pantanal Study Area (Caiman Ranch), 1991-1992 Food Category df KW P value Small Forbs 2 6.95 0.031 Graminoids 2 9.68 0.008 Browse 2 2.34 0.310 Lianas 2 4.05 0.132 Table F-6. Kruskal-Wallis test for differences in percentage use of 10 plant families by pampas deer within seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Season df KW P value Rainy 9 37.88 < 0.0001 Flood 9 33.47 0.0001 Dry 9 40.34 < 0.0001 Table F-7. Kruskal-Wallis test for differences in percentage use of plant five plant species by pampas deer within seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Season df KW P value Rainy 4 3.72 0.445 Flood 4 7.23 0.124 Dry 4 17.83 0.003

PAGE 200

189 Table F-8. Kruskal-Wallis test for differences in percentage use of four food categories by pampas deer within seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Season df KW P value Rainy 3 13.59 0.004 Flood 3 12.30 0.006 Dry 3 16.22 0.001 Table F-9. Kruskal-Wallis test for differences in percentage use of plant families by pampas deer among three seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Plant Family df KW P value Poaceae 2 9.74 0.008 Hydrophilaceae 2 11.18 0.004 Lamiaceae 2 12.09 0.002 Malvaceae 2 8.88 0.012 Onagraceae 2 10.57 0.005 Pontederiaceae 2 13.29 0.001 Sterculiaceae 2 8.78 0.012 Table F-10. Kruskal-Wallis test for differences in percentage use of food categories by pampas deer among three seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Food Category df KW P value Graminoids 2 9.68 0.008 Small forbs 2 6.95 0.031 Browse 2 2.34 0.310

PAGE 201

APPENDIX G STP AND SPEARMAN RANK CORRELATION TESTS

PAGE 202

Table G-1 . STP nonparametric multiple comparisons of the consumption of the most important plant species by cattle during the rainy season, Pantanal Study Area (Caiman Ranch), 1991-1992. A^ pd A M 0.347' H 0.076 0.028 P 0.009 0.009 0.028 ^ Axonopus purpusii. Mesosetum chaseae. " Hymenachne amplexicaulis. Paspalum pontanalis. * P value for differences in consumption of plant species by fecal analysis. Table G-2. STP nonparametric multiple comparisons of the consumption of the most important plant species by cattle during the flood season, Pantanal Study Area (Caiman Ranch), 1991-1992. A^ hF F A M 0.047" H 0.174 0.028 P 0.174 0.028 0.602 Axonopus purpusii. Mesosetum chaseae. " IHymenachne amplexicaulis. Paspalum pontanalis. ^ P value for differences in consumption of plant species by fecal analysis. 191

PAGE 203

192 Table G-3. STP nonparametric multiple comparisons of the consumption of the most important plant species by cattle during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. A^ A M 0.009' H 0.009 0.174 P 0.028 0.465 0.076 C 0.009 0.465 0.117 0.917 ^ Axonopus purpusii. '° Mesosetum chaseae. Hymenachne amplexicaulis. Paspalum pontanalis. * Cyperus sesquiflorus. ' P value for differences in consumption of plant species by fecal analysis.

PAGE 204

193 Table G-4. STP nonparametric multiple comparisons of the consumption of the most important plant families by brown brocket deer during the rainy season, Pantanal Study Area (Caiman Ranch), 1991-1992. Po C E F H L M 0 Pn POa C" 0.028' 0.047 0.754 F' 0.047 0.917 0.754 H' 0.009 0.076 0.028 0.117 0.007 0.007 0.007 0.013 0.007 0.009 0.009 0.009 0.028 0.047 0.095 O'' 0.117 0.917 0.754 0.917 0.602 0.317 0.295 Pn' 0.005 0.005 0.005 0.005 0.005 0.058 0.019 0.019 S' 0.251 0.009 0.009 0.009 0.009 0.007 0.009 0.047 0.005 ^ Poaceae. Compositae. " Euphorbiaceae. Fabaceae. * Hidrophilaceae. ' Lamiacea. ' Malvaceae. ^ Onagraceae. ' Pontederiaceae. ' Sterculiaceae. P value for differences in consumption of browse by fecal analysis.

PAGE 205

194 Table G-5. STP nonparametric multiple comparisons of the consumption of the most Important plant families by brown brocket deer during the flood season, Pantanal Study Area (Caiman Ranch), 1991-1992. Po C E F H L M 0 Pn POa 0.1 74'^ 0.251 0.465 0.016 0.047 0.754 0.009 0.009 0.016 0.076 0.602 0.117 0.347 0.076 0.009 0.009 0.009 0.046 0.046 0.116 0.016 0*^ 0.009 0.009 0.028 0.009 0.009 0.009 0.009 Pn' 0.117 0.117 0.174 0.117 0.117 0.117 0.028 0.917 S' 0.174 0.009 0.076 0.009 0.009 0.117 0.009 0.117 0.465 ^ Poaceae. Compositae. " Euphorbiaceae. Fabaceae. * Hidrophilaceae. ' Lamiacea. ^ Malvaceae. Onagraceae. ' Pontederiaceae. ' Sterculiaceae. P value for differences in consumption of browse by fecal analysis.

PAGE 206

195 Table G-6. STP nonparametric multiple comparisons of the consumption of the most important plant families by brown brocket deer during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. Po C E F H M POa 0.754' 0.117 0.347 0.917 0.917 0.251 0.009 0.076 0.174 0.047 0.251 0.465 0.754 0.251 0.174 0.009 0.009 0.009 0.009 0.009 0.009 ^ Poaceae. Compositae. Euphorbiaceae. Fabaceae. * Hidrophilaceae. ' Malvaceae. ^ Sterculiaceae. ' P value for differences in consumption of browse by fecal analysis.

PAGE 207

196 Table G-7. STP nonparametric multiple comparisons of the consumption of food categories by brown brocket deer during the rainy season, Pantanal Study Area (Caiman Ranch), 1991-1992. B^ Fo'' Fr= U' B Fo 0.0090 Fr 0.0090 0.0283 G 0.0090 0.0090 0.1172 L 0.0090 0.0090 0.2087 0.7540 U 0.0090 0.0090 0.7540 0.0090 0.0163 ^ Browse. " Forbs. " Fruits. Graminoids. * Lianas. ' Unidentified non-graminoid vegetative material. ° P value for differences in consumption between food categories by fecal analysis. Table G-8. STP nonparametric multiple comparisons of the consumption of food categories by brown brocket deer during the flood season, Pantanal Study Area (Caiman Ranch), 1991-1992. B^ Fo" Fr" G^ C lF B Fo 0.1172 Fr 0.0090 0.0090 G 0.0090 0.0090 0.9168 L 0.0090 0.0090 0.8340 U 0.0090 0.0090 0.4647 ^ Browse. ' Forbs. " Fruits. Graminoids. ^ Lianas. 0.9168 0.6015 0.3472 ^ Unidentified non-graminoid vegetative material. ^ P value for differences in consumption between food categories by fecal analysis.

PAGE 208

197 Table G-9. STP nonparametric multiple comparisons of the consumption of food categories by brown brocket deer during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. B^ Fo'' Fr= U' B Fo 0.3472 Fr 0.0090 0.0090 G 0.0163 0.0090 0.4020 L 0.0163 0.0090 0.5296 0.5296 U 0.0283 0.0090 0.0283 0.1745 0.1745 ^ Browse. " Forbs. " Fruits. Graminoids. * Lianas. ^ Unidentified non-graminoid vegetative material. ^ P value for differences in consumption between food categories by fecal analysis. Table G-10. Nonparametric multiple comparisons by STP of the consumption of browse by brown brocket deer among seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Rainy Flood Dry Rainy Flood 0.4647" Dry a _ 1 0.1172 0.3472 P value for differences in consumption of browse by fecal analysis. Table G-1 1 . Nonparametric multiple comparisons by STP of the consumption of forbs by brown brocket deer among seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Rainy Flood Dry Rainy Flood 0.1172" Dry a r-. 1 0.0283 0.4647 P value for differences in consumption of browse by fecal analysis.

PAGE 209

198 Table G-12. Nonparametric multiple comparisons by STP of the consumption of fruits by brown brocket deer among seasons, Pantanal Study Area (Caiman Ranch), 1991-1992. Rainy Flood Dry Rainy Flood 0.2087' Dry 0.1425 0.8340 P value for differences in consumption of browse by fecal analysis. Table G-1 3. STP nonparametric multiple comparisons of the consumption of the most important plant species by pampas deer during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. Ca' Ms'= S' v' Ca H 0.174^ Ms 0.009 0.009 Mv 0.251 0.047 0.009 S 0.917 0.174 0.009 0.347 V 0.251 0.009 0.009 1.000 0.209 ' Caperonia castaneifolia. Hydrolea spinosa. " Melochia simplex. Melochia villosa. ^ Sida santamarensis. ' Vernonia scabra. ^ P value for differences in consumption of plant species by fecal analysis.

PAGE 210

199 Table G-14. STP nonparametric multiple comparisons of the consumption of the most important plant families by pampas deer during the rainy season, Pantanal Study Area (Caiman Ranch), 1991-1992. R' S" c E 0.1745' M 0.4647 0.1172 R 0.9168 0.4647 0.3472 S 0.0090 0.0090 0.0090 0.0090 ^ Compositae. Euphorbiaceae. " Malvaceae. " Rubiaceae. ® Sterculiaceae. ' P value for differences in consumption between families. Table G-15. STP nonparametric multiple comparisons of the consumption of the most important plant families by pampas deer during the flood season, Pantanal Study Area (Caiman Ranch), 1 991 -1 992. C' E^ R^ S' 0 E 0.1745' M 0.6015 0.7540 R 0.9168 0.2506 0.3472 S an 0.9168 0.2506 0.3472 0.6015 ^ Compositae. Euphorbiaceae. ^ Malvaceae. Rubiaceae. ^ Sterculiaceae. ' P value for differences in consumption between families.

PAGE 211

200 Table G-16. STP nonparametric multiple comparisons of the consumption of the most important plant families by pampas deer during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. R" S" c E 0.9168' M 0.9168 0.9168 R 0.9168 0.7540 0.9168 S 0.7540 0.3472 1.0000 0.3472 ^ Compositae. Euphorbiaceae. Malvaceae. Rubiaceae. * Sterculiaceae. ' P value for differences in consumption between families. Table G-17. STP nonparametric multiple comparisons of the consumption of food categories by pampas deer during the rainy season, Pantanal Study Area (Caiman Ranch), 1991-1992. G' G F 0.117" B 0.754 0.028 H 0.009 0.009 0.009 ^ Graminoids. " Small forbs. " Shrub and tree seedlings. Flowers and leaves of Alismataceae and Pontederiaceae. " P value for differences in consumption of food categories by fecal analysis.

PAGE 212

201 Table G-18. STP nonparametric multiple comparisons of the consumption of food categories by pampas deer during the flood season, Pantanal Study Area (Caiman Ranch), 1991-1992. G F 0.009^ B 0.047 0.016 H 0.117 0.016 0.602 ^ Graminoids. " Small forbs. " Shrub and tree seedlings. Flowers and leaves of Alismataceae and Pontederiaceae. * P value for differences in consumption of food categories by fecal analysis. Table G-19. STP nonparametric multiple comparisons of the comsumption of food categories by pampas deer during the dry season, Pantanal Study Area (Caiman Ranch), 1991-1992. G' f'' B= G F 0.009' B 0.016 0.009 H 0.173 0.009 0.009 ^ Graminoids. ' Small forbs. ^ Shrub and tree seedlings. Flowers and leaves of Alismataceae and Pontederiaceae. * P value for differences in consumption of food categories by fecal analysis. Table G-20. Spearman Rank Correlation test between the composition of the seasonal diet of pampas deer obtained by direct observation and by fecal analysis. Comparison rs^ n" 95% CI P value Rainy 0.81 8 0.258 to 0.965 0.014 Flood 0.48 8 0.000 to 0.886 0.228 Dry a r» 1 0.90 8 0.527 to 0.982 0.002 Number of food categories in comparisons.

PAGE 213

REFERENCES ABRAHAMSON, W.G. 1989. Plant Animal Interactions. McGraw-Hill, Inc. , New York, N.Y. 480 pp. ADAMOLI, J. 1986. Fitogeografia do Pantanal. Anais do simposio sobre recursos naturals e socio-economicos do Pantanal, EMBRAPA, Brasilia 5:105-106. ALHO, C.R.J., T.E. LACHER, Z.M.S. CAMPOS, and H.C. GONQALVES. 1987a. Ecologia de capivara {Hydrochaeris hydrrochaeris, Rodentia) do Pantanal: I. Habitats, densidades e tamanho de grupo. Revista Brasileira de Biologia, 47:87-97. ALHO, C.R.J., T.E. LACHER, Z.M.S. CAMPOS, and H.C. GONQALVES. 1987b. Ecologia de capivara {Hydrochaeris hydrrochaeris, Rodentia) do Pantanal: II. Atividade, sazonalidade, uso do espapo e manejo. Revista Brasileira de S/o/ogr/a 47:99-1 10. ALHO, C.R.J., T.E. LACHER, and H.C. GONQALVES. 1988. Environmental degradation in the Pantanal ecosystem. Bioscience 38:164-171. ALLDREDGE, J.R. and J.T. RATTI. 1992. Further comparison of some statistical techniques for analysis of resource selection. J. Wildl. Manage. 56:1-9. ALLEN, A C. and J.F.M. VALLS. 1987. Recursos forrageiros nativos do Pantanal Mato-grossense (Allen and Vails, eds.). EMBRAPA-CENARGEN, Documentos 8, 339 pp. ALMEIDA, A. 1976. Hunting jaguar in the Mato-Grosso. Stanwill Press, England, 194 pp. 2p2

PAGE 214

203 AVILA-PIRES, F.D. de. 1959. As fomnas sul-americanas do "veado-vira". Anais da Academia Brasileira de Ciencias 31 : 547-556. BANKS, V. 1991 . The Pantanal: Brazil's forgotten wilderness. Sierra Club, San Francisco, Calif. 254 pp. BAHAMONDE, N., S. MARTIN, and A.P. SBRILLER. 1986. Diet of guanaco and red deer in Neuquen Province, Argentina. J. Range Manage. 39(1 ):22-23. BECCACECI, M.D. 1994. A census of marsh deer in Ibera Natural Reserve, its Argentine stronghold. Oryx 28: 1 31 -1 34. BECCACECI, M.D. and M.L. MERINO. 1994. Dieta del Ciervo de los Pantanos en la Reserva Ibera, Corrientes, Argentina. Pp 129-130 in: Population and Habitat Viability Assessnrtent Report for Cervo-do-Pantanal (Blastocerus dichotomus) (Pinder, L. and U.S. Seal eds.). lUCN/SSC Conservation Breeding Specialist Group, MN. BEHNKE, R. and N. ABEL. 1996. Stocking rates for African pastoral systems. World Animal Review 2:9-1 7. BIANCHINI, J.J. and J.C.L. PEREZ. 1972. Informe sobre la situacion actual del ciervo de las pampas Ozotoceros bezoarticus celer Cabrera 1 943 en la Provincia de Buenos Aires. Acta Zool. Lilloana 39:149-157. BISBAL. E.F.J. 1994. Population biology and diet of the brocket deer (Mazama spp.) (Artiodactyla: Cervidae) in Venezuela. Revista de Biologia Tropical 42:305313. BODMER, R.E. 1989. Frugivory in Amazon ungulates. Ph.D. dissertation. University of Cambridge, U.K., 158 pp. BODMER, R E. 1990. Responses of ungulates to seasonal inundations in the Amazon floodplain J. Trop. Ecol. 6:191-201. BODMER, R.E. 1991. Strategies of seed predation in Amazonian ungulates. Biotropica 23:255-261 .

PAGE 215

204 BONINO, N. and A.P. SHRILLER. 1991. Composicion botanica de la dieta del guanaco (Lama guanlcoe) en dos ambientes contrastantes de Tierra del Fuego, Argentina. Ecologia yAusfra/ 1:97-102. BRANAN, W.C., M.C. WERKHOVEN, and R.L. MARCHINTON. 1985. Food habits of brocket and white-tailed deer in Suriname. J. Wildl. Manage. 49:972-976. BRASIL. 1974. Estudos hidrologicos da Bacia do Alto Paraguai. Technical Report, 1 . Ministerio do Interior, Departamento Nacional de Obras e Saneamento, Brasilia, D.F., Brazil. 359 pp. BRASIL. 1982. Projeto RADAMBRASIL, 27. Ministerio das Minas e Energia, Secretaria-Geral, Brasilia, D.F., Brazil. 448 pp. BRUM, P A R. de, J.C. de SOUSA, J.A. COMASTRI FILHO, and I.L. de ALMEIDA. 1987. Deficiencias minerals de bovinos na sub-regiao dos Paiaguas, no Pantanal Mato-grossense. Pesq. Agrop. Bras. 22:10391048. BYERS, C.R., R.K. STEINHORST, and P R. KRAUSMAN. 1984. Clarification of a technique for analysis of utilization-availability data. J. Wildl. Manage 48:1050-1053. CABRERA, A. 1943. Sobre la sistematica del venado y su vahacion individual y geografica. Rev. Mus. La Plata, Zool. 3:5-41. CARVALHO, C.T. 1973. 0 veado-campeiro. Situagao e distribuigao (Mammalia, Cervidae). Boletim Tecnico Instituto Florestal, Sao Paulo 7:9-22. CAUGHLEY, G., G.C. GRIGG, J. CAUGHLEY, and G.J.E. HILL. 1980. Does dingo predation control the densities of kangoroos and emus? Austr Wildl Res 7:1-12. CHAPLIN, R.E. 1977. Deer. Blandford Press, Dorset, U.K., 218 pp. CLUTTON-BROCK, T.H., F.E. GUINNESS, and S.D. ALBON. 1982. Red Deer: Behavior and Ecology of Two Sexes. The University of Chicago Press, Chicago, III. 378 pp. CODY, M L. and J.M. DIAMOND. 1975. Ecology and Evolution of Communities. Belknap Press, Harvard University Press, Cambndge, MA, 545 pp.

PAGE 216

205 CONRAD, J.H., LR. McDOWELL, G.L ELLIS, and J.K. LOOSLI. 1985. Minerals para ruminantes em pastejo em regioes tropicais. University of Florida, Gainesville, 90 pp. COUTINHO, L.M. 1982. Ecological effects of fire in the Brazilian Cerrado. Pp 273291 in Ecology of Tropical Savanr)as{Hunt\ey , B.J. and B.H. Walker, eds.). Ecological Studies Vol. 42, Springer-Verlag, Berlin. COUTINHO, L.M. 1990. Fire in the ecology of the Brazilian Cerrrado. Pp. 82-105 in Fire in the Tropical Biota (J.G. Goldammer, ed.). Springer-Verlag, Berlin, Germany. CRAWLEY, M.J. 1983. Herbivory: the dynamics of animal-plant interactions. University of California Press, Bekerley, Calif. 437 pp. CRAWSHAW, P.G. and H.B. QUIGLEY. 1984. Ecologia do jaguar Panthera onca no Pantanal. Typed report, on file IBDF/DN (IBAMA), Brasilia, Brasil, 1 10 pp. CROCKER, R.L. and N.S. TIVER. 1948. Survey methods in grassland ecology. J. Sr. Grassld Soc. 3:1-26. CZERNAY, S. 1987. Spiesshirsctie und pudus: die Gattungen Mazama undPudu. A. Ziemsen Verlag, Wittenberg Lutherstadt, 84 pp. De AZCARATE, T. 1981. Sociobiologia y manejo del chiguire. Donana Acta Vertebrata 7:1 -228. DEARDEN, B.L., R E. PEGAU, and R.M. HANSEN. 1975. Precision of miaohistological estimates of ruminant food habits. J. Wildl. Manage 39 402407. DEGAYNER, E.J. and P A JORDAN. 1987. Skewed fetal sex ratios in white-tailed deer: evidence and evolutionary speculations. Pp. 178-188 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, DC. DEINUM, B. and J.G.P. DIRVEN. 1971. Climate, nitrogen, and grass. The influence of age on chemical composition and in vitro digestibility of maize (Zea mayx) and tall fescue {Festuca arundinacea). Neth. J. Agnc. Sci 19:264-272. DEMMENT, M.W. and P.J. VAN SOEST. 1985. A nutritional explanation for bodysize patterns of ruminant and nonruminant herbivores The Am Nat 125:641-672.

PAGE 217

206 DINERSTEIN, E. 1987. Deer, plant phenology, and succession in the lowland forests of Nepal. Pp. 272-288 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.. DUNBAR, R.I.M. 1978. Competition and niche separation in a high altitude herbivore community in Ethiopia. E.Afr. Wildl. J. 16:183-199. EISENBERG, J.F. 1981. Ttie mammalian radiations. The University of Chicago Press, Chicago, III. 509 pp. EISENBERG, J.F. 1987. The evolutionary history of the Cervidae with spetial reference to the South American Radiation. Pp 60-64 in Biology and Management of ttie Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C. EISENBERG, J.F. 1989. Mammals of the A/eofrop/cs Vol 1. The University of Chicago Press, Chicago, III. 449 pp. EITEN, G. 1972. The Cerrado vegetation of Brazil. Bot. Rev. 38:201-341. EITEN, G. 1985. Vegetation near Santa Teresinha, NE Mato Grosso. Acta Amazonica 15:275-301. ELLIOTT III, H.W. and R.H. BARRETT. 1985. Dietary overlap among axis, fallow, and black-tailed deer and cattle. J. Range. Manage. 38:546-550. EMMONS, L.H. 1990. Neotropical rainforest mammals. The University of Chicago Press, Chicago, III. 281 pp. ERASMUS, T., B.L. PENZHORN, and N. FAIRALL. 1978. Chemical composition of faeces as an index of veld quality. S. Afric J. Wild! Res. 8:19-24. ESCOBAR, A. and E.G. JIMENEZ. 1976. Estudio de la competencia alimenticia de los herbivores mayores del llano inundable con especial referenda al chiguire {Hydrochaehs hydrochaehs). Agr Tropical. 26:215-227. FOX, B.J. 1989. Community ecology of macropodoids. Pp. 89-104 in Kangaroos, Wallabies and Rat-kangoroos (Grigg, G., P. Jarman, and I. Hume, eds.). Surrey Beatty & Sons Pty Limited, Chipping Norton, Australia. FRADRICH, H. 1987. The husbandry of tropical and temperate cervids in West Berlin Zoo. Pp. 422^28 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.

PAGE 218

207 GARDNER, A.L. 1971 . Postpartum estrus in a red brocket deer Mazama americana from Peru. J. Mammal. 52:623-624. GOLDSMITH, F.B., CM. HARRISON, and A.J. MORTON. 1986. Description and analysis of vegetation. Pp. 437-524 in Methods in Plant Ecology (Moore and Chapman, eds.). Blackwell Scientific Publications, Oxford, U.K.. GOLLEY, F.B. 1978. Forest leaf production. Pp. 17-22 in The Ecology of Arboreal Folivores (Montgomery G.G., ed.). Smithsonian Institution Press, Washington, D.C.. GOMES, L. 1997. Santuario Ameagado. Vfeya 30:58-73. GOODALL, D.W. 1952. Some considerations in the use of point quadrats for the analysis of vegetation. Aust. J. Sdent. Res. Sen B. 5: 1 -41 . GORDON, I.J. 1988. Facilitation of red deer grazing by cattle and its impact on red deer performance. J. Appl. Ecol. 25:1-10. GORDON, I.J. 1989a. Vegetation community selection by ungulates on the Isle of Rhum. III. Determinants of vegetation community selection. J. Appl. Ecol. 26:65-79. GORDON, I.J. 1989b. Vegetation community selection by ungulates on the Isle of Rhum. II. Vegetation community selection. J. Appl. Ecol. 26:53-64. GRACE, N.D. 1983. Cooper. Pp. 56-66 in The mineral requirements of grazing ruminants (Grace, N.D. ed.). Palmerston North, NewZeland Society of Animal Production, New Zeland. GREEN, M. 1987. Diet composition and quality in hymalayan musk deer based on fecal analysis. J. Wildl. Manage. 51:880-892. GROSS, J.E., P.U. ALKON, and M.W. DEMMENT. 1996. Nutritional ecology of dimorphic herbivores: Digestion and passage rates in Nubian ibex. Oecologia 107(2): 170-178. HANLEY, T.A. and K.A. HANLEY. 1982. Food resource partitioning by sympatric ungulates on Great Basin rangeland. J. Range. Manage 35:152-158. HANSEN, R.M., M M. MUGAMBI, and S.M. BAUNI. 1985. Diets and trophic ranking of ungulates of the northern Serengeti. J. Wildl. Manage. 49 823829.

PAGE 219

208 HANSON, W.R. and GRAYBILL, F. 1956. Sample size in food-habits analyses. J. Wildl. Manage. 20:64-68. HEINONEN, S., H. CHAVES, R. MALETTI, E. KRAUCZUK, and G. CAVIA. 1989. Operative "Guazu-ti" Prlmera etapa. Fundacion Vida Silvestre Argentina, Capitulo Missiones, Posadas. 12 pp. HENRY, S.R., A.K. LEE, and A. P. SMITH. 1989. The trophic structure and species richness of assemblages of arboreal mammals in Australian forests. Pp. 229-240 in Patterns in the Structure of Mammalian Communities (Morris, D.W., Z. Abramsky, B.J. Fox, and M.R. Wlllig, eds.). Texas Tech University Press, Lubbock, Tex. HERRERA, E. and D. MACDONALD. 1989. Resource utilization and territoriality in group-living capybaras (Hydroclnoerus hydrochaehs). J. Anim. Ecol. 58:667679. HOEHNE, F.C.. 1923. Phytophysionomia do Estado de Mato-Grosso e ligeiras notas a respeito da composigao e disthbuigao da sua flora. Melhoramentos, Sao Paulo, 104 pp. HOEHNE, F.C. 1936. 0 grande Pantanal de Mato Grosso. Bol. Agric, Sao Paulo 1936:443-470. HOFMANN, R.R. 1973. Ttie Ruminant Stomach: Stomach Structure and Feeding Habits of East African Game Ruminants. Kenya Literature Bureau, Nairobi, Kenya, 349 pp. HOFMANN, R.R. 1988. Anatomy of the gastro-intestinal tract. Pp. 14-43 in The Ruminant Animal: Digestive Physiology and Nutrition (D.C. Church, ed). Prentice Hall, Englewoods Cliffs, N.J.. HOLECHEK, J.L., B. GROSS, S.M.DABO, and T. STEPHENSON. 1982. Effects of sample preparation, growth stage, and observer on microhistological analysis of herbivore diets. J. Wildl. Manage. 46:281-568. HOMOLKA, M. and M. HEROLDOVA. 1992. Similarity of the results of stomach and faecal contents analyses in studies of the ungulate diet. Folia Zoologica 41 : 1 93-208. HORN, H.S. 1966. Measurement of "overlap" in comparative ecological studies Am Nat 100:419-424.

PAGE 220

209 HUDSON, R.J. 1985. Body size, energetics, and adaptive radiation. Pp. 1-24 in Bioenergetics of Wild Herbivores {Hudson, R.J., and R.G. White, eds.). CRC Press Inc., Boca Raton, Fla.. HURLBERT, S.H. 1971 . The non-concept of species diversity: A critique and alternative parameters. Ecology 52:577-586. IBGE. 1977. Geografia do Brasil: Regiao Centro-Oeste, Vol IV. SERGRAF, Rio de Janeiro, 364 pp. ILLIUS, A.W. and I.J. GORDON. 1991. Prediction of intake and digestion in ruminants by a model of rumen kinetics integrating animal size and plant charactenstics. J. Agric. Scienc. 116:145-157. JACKSON, J.E. 1987. Ozotoceros bezoarticus. Mammalian Spec/es 295: 1-5. JACKSON, J.E. and J.D. GIULETTI. 1988. The food habit of pampas deer {Ozotoceros bezoarticus celer) in relation to its conservation in a relict natural grassland in Argentine. Biol. Conserv. 45: 1 -1 0. JACKSON, J.E. and A. LANGGUTH. 1987. Ecology and status of the pampas deer {Ozotoceros bezoarticus) in the Argentinian pampas and Uruguay. Pp 402409 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.. JARMAN, P.J. 1971. Diets of large mammals in the woodlands around Lake Kariba. Oeco/og/a 8: 1 57-1 78. JARMAN, P.J. and A R E. SINCLAIR. 1979. Feeding strategy and the pattern of resource partitioning in ungulates. Pp 130-163 in Serengeti: Dynamics of an Ecosystem (A.R.E. Sinclair and M. Norton-Griffiths, eds.). University of Chicago Press, Chicago, III.. JOHNSON, M.K. 1982. Frequency sampling for microscopic analysis of botanical compositions. J. Range. Manage. 35:541-542. JORGENSON, J. 1986. Notes on the ecology and behaviour of capybaras in Northeastern Colombia. Vida Silvestre Neotropical 1:31-40. JUNGIUS, H. 1976. Status and distribution of Threatened deer species in South America. Pp:203-217 in Report to SSC/IUCN Deer Specialist Group W.W.F. Yearbook. KEITH, L.B. 1983. Role of food in hare population cycles. 0//fos 40:385-395.

PAGE 221

210 KETTLEWELL, H.B.D. 1955. Selection experiments on industrial melanism in the lepidoptera. Heredity 9:323-342. KREBS, C. 1989. Ecological Methodology. Harper Collins Publishers, New York, 645 pp. LEEUWENBERG, F. and S. L. RESENDE. 1994. Ecologia de Cervideos na Reserva Ecoiogica do IBGE/DF: Manejo e Densidade de Popula96es. Cadernos de Geociencias 1 1 : 89-95. LEIGH, E.G. and D.M. WINDSOR. 1982. Forest production and regulation of primary consumers on Barro Colorado Island. Pp. 1 1 1-122 in Ttie Ecology of a Tropical Forest: Seasonal Rhythms and Long-term Changes {l.e\Qh, E.G., A.S. Rand, and D.M. Windsor, eds.). Smithsonian Institution Press, Washington, D.C.. LEW, E.B. 1927. The grassland of New Zealand. Progress of regrassing experiments on deteriored hill county in Whangamomona County I. Interim results of burning and sowing trials: seed mixture recommended for second bums. N.Z.J. Aghc 34:145-164. LEVY, E.B. and E.A. MADDEN. 1933. The point method of pasture analysis. N.Z.J. Aghc 46:267-279. LOURIVAL, R.F.F., and G.A.B. FONSECA. 1997. Analise de sustentabilidade do modelo de caga tradicional, no Pantanal da Nhecolandia, Corumba, MS. Pp 123-172 in Manejo e Conservagao de Vida Silvestre no Brasil (ValladaresPadua, C, R.E. Bodmer & L. Cullen Jr. eds.). CNPq, Brasilia, Brasil. LUDWIG, J.A. and J.F. REYNOLDS. 1988. Statistical Ecology: A pnmer on methods and computing. John Wiley and Sons, New York, N.Y. 337 pp. MACARTHUR, R.H. 1972. Geographical Ecology: Patterns in the Distnbution of Species. Harper & Row, Publishers, New York, N.Y. 269 pp. MACDONALD, D. 1981. Dwindling resources and the social behaviour of capybaras (Hydrochaehs hydrochaehs) (Mammalia). J. Zool. (London) 194:371-391. MANTOVANI, W. and F.R. MARTINS. 1990. O metodo de pontes. Acta Bot Bras 4:95-122. MARCUM, C.L. and D O. LOFTSGAARDEN. 1980. A nonmapping technique for studying habitat preferences. J. Wild! Manage. 44:963-968.

PAGE 222

211 MARTINEZ, A M., V. MOLINA, S.F. GONZALEZ, S.J. MARROQUIN, and CH.J.NAVAR. 1997. Observations of white-tailed deer and cattle diets in Mexico. J. Range. Manage. 50:253-257. MAURO, R.A. 1 993. Abundanda e padrao de distribuigao de cervo-do-pantanal Blastocerus dichotomus (llliper. 1815), no Pantanal Mato-Grossense. M.Sc. thesis, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil. 48 pp. MCCLOSKEY, L.R. 1970. The dynamics of the community associated with a marine scleractinian coral. Int. Rev. Gesamten Hydrobiol. 55:13-81. MERINO, M. 1993. Dieta del "venado de las pampas" {Ozotoceros bezoarticus ce/er Cabrera 1943) en la Reserva de Vida Silvestre "Campos del Tuyu", Bahia de Saborombon, Provincia de Buenos Aires, Argentina: Datos Preliminares. Pp. 150-157 in Population and Habitat Viabilty Assessment for the Pampas Deer (Ozotoceros bezoarticus) (Gonzalez, S., M. Merino, M. Gimenz-Dixon, S. Ellis, and U.S. Seal eds.). lUCN/SSC Conservation Breeding Specialist Group, MN, U.S.A.. MERINO,L.M, S. CONZALES, F. LEEUWENBERG, F.H.G. RODRIGUES, L. PINDER, and W.M. TOMAS. 1997. Veado-campeiro {Ozotoceros bezoarticus). Pp. 42-55 in Biologia e Conservagao de Cervideos SulAmericanos: Blastocerus, Ozotoceros e Mazama (J.M.B. Duarte, ed.). FUNEP, Jaboticabal, SP, Brazil. MILLER, F.W. 1930. Notes on some mammals of southern Matto Grosso, Brazil. J. Mammal. 11:10-22. MISHRA, H.R. & C. WEMMER. 1987. The comparative breeding ecology of four cervids in Royal Chitwan National Park. Pp. 259-271 in Biology and Management of ttie Cen/idae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.. MULLER, R.N. 1982. Vegetation patterns in the mixed mesophytic forest of eastern Kentucky. Eco/ogy 63: 1901 -191 7. MULLER-SCHWARZE, D., B. STAGGE, and C. MULLER-SCHWARZE. 1982. Play behavior: persistence, decrease and energetic compensation during food shortage in deer fawns. Sc/ence 21 5:85. MULLER-SCHWARZE, D. 1987. Evolution of cervid olfactory communication. Pp. 223-234 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.

PAGE 223

212 MURPHEY, D A. and J.A. COATES. 1966. Effects of dietary protein on deer. Trans. North Am. Wildl. and Nat. Resour Conf. 21:129-138. MURRAY, M.G., and D. BROWN. 1993. Niche separation of grazing ungulates in the Serengeti: an experimental test. J. Anim. Ecol. 62:380-389. NATIONAL RESEARCH COUNCIL. 1976. Nutrient requirements of beef cattle. Committee on animal nutrition. Subcommittee on beef cattle nutrition, Washington, EUA. National Academy of Science, Washington, D.C. 56 pp. NEU. C.W., C.R. BYERS and J.M. PEEK. 1974. A technique for analysis of utilization-availability data. J. Wildl. Manage. 38:541-545. NORES, J.G. 1984. Contenido de algunos elementos trazas en praderas naturales uruguayas. R, Fac. Agron. 35:23-25. NORTON-GRIFFITHS, M. 1979. The influence of grazing, bro\A«ing, and fire on the vegetation dynamics of the Serengeti. Pp. 310-352 in Serengeti: Dynamics of an Ecosystem (Sinclair, A. R E. and M. Norton-Griffiths eds.). The University of Chicago Press, Chicago. NOWAK,R.M. and J.L. PARADISO. 1983. Walker's mammals of the world. Johns Hopkins Univ. Press, Baltimore, Md. 2:569-1362. OFDETAL, O.T. 1985. Pregnancy and lactation. Pp. 215-238 in Bioenergetics of Wild Herbivores {Hudson, R.J., and R.G. White, eds.). CRC Press Inc., Boca Raton, Fla.. OJEDA, R.A. and M A. MARES. 1982. Conservation of South American Mammals: Argentina as a paradigm. Special Publ. Pymatuning Lab. of Ecol. 6:505-539. OWEN-SMITH, N. 1989. Morphological factors and their consequences for resource partitioning among African savanna ungulates: a simulation modelling approach. Pp 155-165 in Patterns in the Structure of Mammalian Communities (Morris, D.W., Z. Abramsky, B.J. Fox, and M.R. Willig, eds.). Texas Tech University Press, Lubbock, Tex.. PARRA, R. 1978. Comparison of foregutand hindgut fermentation in herbivores. Pp 205-230 in The Ecology of Arboreal Folivores (G.G. Montgomery, ed.). Smithsonian Institution Press, Washington, D.C. PHILLIPSON, J. 1975. Rainfall, primary productivity, and 'carrying capacity' of Tsavo National Park (East Kenya). E. African Wildl J. 13:171-201.

PAGE 224

213 PINDER, L. 1992. Comportamento social e reprodutivo dos veados campeiro e catingueiro. Anais de Etologia 10:167-173. PINDER, L. 1993. Status of pampas deer in Brazil. Pp. 157-162 in Population and Habitat Viabilty Assessment for the Pampas Deer (Ozotoceros bezoarticus) (Gonzalez, S., M. Merino, M. Gimenz-Dixon, S. Ellis, and U.S. Seal, eds.). lUCN/SSC Conservation Breeding Specialist Group, MN, U.S.A. PINDER, L. 1994. Marsh deer seasonal movements and home range size. Pp. 121127 in Population and Habitat Viability Assessment Report for Cervo-doPantanal (Blastocerus dicfiotomus) (Pinder, L. and U.S. Seal, eds.). lUCN/SSC Conservation Breeding Specialist Group, MN, U.S.A.. PINDER, L. 1996. Marsh deer Blastocerus dichotomus population estimate in the Parana river, Brazil. Biol. Conserv. 75:87-91. PINDER, L. 1997. Veado-catingueiro {Mazama gouazoubira, Fisher 1814). Pp 60-67 in Biologia e Conservagao de Cervideos SulAmericanos: Blastocerus, Ozotoceros e Mazama (J.M.B. Duarte, ed.). FUNEP, Jaboticabal, SP, Brazil. PINTO, M.N. 1994. Cerrado: Caracterizagao, ocupagao e perspectivas. Edunb, Brasilia, D.F., Brazil, 681 pp. POR, F.D. 1995. The Pantanal of Mato Grosso (Brazil). Kluwer Academic Publishers, London, U.K. 122 pp. POTT, A. 1 982. Pastagens das sub-regioes dos Paiaguas e da Nhecolandia do Pantanal Mato-grossense. EMBRAPA-UEPAE, Corumba, Circular Tecnica 10, 49 pp. POTT, A. 1986. Pastagens no Pantanal. Anais FEALQ 1986:413-441. POTT, A. 1988. Pastagens do Pantanal. EMBRAPA-CPAP, documentos 7, Corumba, 58 pp. POTT, A., Z. CAMPOS, and C. ALHO. 1986. Plantas da dieta da capivara de Nhecolandia, Pantanal; observagoes iniciais. Resumes do Congresso Nacionalde Botanies 27 (Ouro Preto), 262 pp. POTT, E.B.. 1986. Teores de minerals e proteina bruta em plantas forrageiras da parte alta de Corumba, MS. EMBRAPA-CPAP technical report, Corumba MS, Brazil, 6 pp.

PAGE 225

214 POTT, E.B., P A R. BRUM, I.L. DE ALMEIDA, J.A. COMASTRI FILHO, and J.F. DYNIA. 1987. Nutrigao mineral de bovinos de corte no Pantanal MatoGrossense. Pesq. Agropec. Bras. 22:1093-1109. POTT E.B. and A. POTT. 1987. Niveis de nutrientes em plantas nao-gramineas pastejadas por bovinos na sub-regiao dos Paiaguas, do Pantanal MatoGrossense. Pesq. Agropec. Bras. 22:1293-1299. POTT E.B., I.L. de Almeida, P A R. de BRUM, J.A. COMASTRI FILHO, A. POTT, and J.F. DYNIA. 1989. Nutrigao mineral de bovinos de corte do Pantanal Mato-Grossense: 2. Miaonutrientes na Nhecolandia (Parte Central). Pesq. agropec. bras. 24:109-126. PRANCE, G.T. and G.B. SCHALLER. 1982. Preliminary study of some vegetation types of the Pantanal, Mato Grosso, Brasil. Brittonia 32:223-251. PRICE, M.A., and R.G. WHITE. 1985. Growth and development. Pp. 183-213 in Bioenergetics of Wild Herbivores {Hudson, R.J., and R.G. White, eds.). CRC Press, Boca Raton. PUTMAN, R.J. 1984. Facts from faeces. Mammal Rev. 14:79-97. PUTMAN, R.J. 1996. Competition and Resource Partitioning in Temperate Ungulate Assemblies. Chapman and Hall, London, 131 pp. PYKE, G.H., H.R. PULLIAM, and E.L. CHARNOV. 1977. Optimal foraging: a selective review of theory and tests. Q. Rev. Biol. 52:137-154. QUINTANA, R.D., S. MONGE, and A.S.M. ALVAREZ. 1994. Feeding habits of capybara {Hydrochaeris hydrochaeris) in afforestation areas of the Lower Delta of the Parana River, Argentina. Mammalia 58:569-580. RAJASEKARAN, G. 1988. Food habits of black buck (Antelope cervicapra) of Point Calimeer Wildlife Sanctuary. M.Sc. Thesis, Bharathidasan University, India, 26 pp. RATTER, J.A., A. POTT, V.J. POTT, C.N. da CUNHA, and M. HARIDASAN. 1989. Observations on woody vegetation types in the Pantanal and at Corumba, Brazil. Notes RBG Edinb. 45:503-525. REDFORD, KH. 1985. Emas National Park and the plight of the Brazilian cerrados Oryof 19:210-214.

PAGE 226

215 REDFORD, KH. 1987. The pampas deer {Ozotoceros bezoarticus) in central Brazil. Pp. 410-414 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, D.C.. REDFORD, K.H. 1 997. A floresta vazia. Pp 1-22 in Manep e Conservagao de Vida Silvestre no Bras// (Valladares-Padua, C, R.E. Bodmer, and L. Cullen Jr., eds.). CNPq, Brasilia, DF, Brazil. REDFORD, K.H. and J.F. EISENBERG. 1992. Mammals of ttie Neotropics\/o\. 2. The University of Chicago Press, Chicago, III. 430 pp. REDFORD, K.H. and G.A.B. FONSECA. 1986. The role of gallery forests in the zoogeography of the cerrado's non-volant mammalian fauna. Biotropica 18:126-135. RENKONEN, 0. 1938. Statisch-okologische Untersuchungen uber die terrestiche kaferwelt der finnischen bruchmoore. Ann. Zool. Soc. Bot. Fenn. Vanamo 6:1-231. RESENDE, S.L. and F. LEEUWENBERG. 1992. Estudos de Cervideos na Bacia do Taquara. Final report to CNPq, Brasilia, 50 pp. RHODES, B.D. and S.H. SHARROW. 1990. Effect of grazing by sheep on the quantity and quality of forage available to big game in Oregon coast range. J. Range Manage. 43:235-237. RICHTER, A.R. and R.F. LABISKY. 1985. Reproductive dynamics among disjunct white-tailed deer herds in Florida. J. Wildl. Manage. 49:964-971 . RICKLEFS, R.E., and M. LAU. 1980. Bias and dispersion of overlap indices: results of some Monte Carlo simulations. Eco/ogy 61:1019-1024. RIZZINI, C.T. 1979. Tratado de Fitogeografia do Brasil: aspectos ecologicos e floristicos, Vol 2. HUCITEC Ltda., Sao Paulo, Brasil, 374 pp. ROBBINS, C.T, O.T OFTEDAL, and K.I. O'ROURKE. 1987. Lactation, early nutrition, and hand-rearing of wild ungulates, with spetial reference to deer. Pp 429-442 in Biology and Management of the Cervidae (CM. Wemmer, ed.), Smithsonian Institution Press, Washington, DC. ROBBINS, L.E. and R.L. MYERS. 1992. Seasonal effects of prescribed burning in Florida: a review. Tall Timbers Res. Inc., Misc. Publ. 8. 96 pp.

PAGE 227

216 ROBERTSON, P A, G.T. WEAVER, and J.A CAVANAUGH. 1 978.Vegetation and tree species patterns near the northern terminus of the Southern Floodplain Forest. Eco/. Monogr. 48:249-267. ROBERTSON, P A, M.D. MACKENZIE, and L.F. ELLIOT. 1984. Gradient analysis and classification of the woody vegetation for 4 sites in southern Illinois and adjacent Missouri. Vegetatio 58:87-104. ROBINSON, J.G. 1986. Seasonal variation in use of time and space by the wedgecapped capuchin monkey, Cebus olivaceous: implications for foraging theory. Smithsonian Contrib. to Zool., 431, 60 pp. ROBINSON, J.G. and K.H.REDFORD. 1986. Body size, diet, and population density of Neotropical forest mammals. Am. Nat. 128:665-680. ROBINSON, S.K. 1985. Coloniality in the yellow-rumped cacique as a defence against nest predators. /Aa/( 102:506-519. RODRIGUES, F.H.G. 1996. Histoha natural e biologia comportamental de veadocampeiro (Ozotoceros t)ezoarticus) em cerrado do Brasil Central. M.Sc. Thesis, Universidade Estadual de Campinas, Sao Paulo, Brasil, 89 pp. ROUTLEDGE, R.D. and D A. FYFE. 1992. Confidence limits for line transect estimates based on shape restrictions. J. Wildl. Manage. 56:402-407. SADLEIR, R.M.F.S. 1987. Reproduction of female cervids. Pp. 123-144 in Biology and Management of the Cervid{C.M. Wemmer, ed.) Smithsonian Institution Press, Washington, D.C.. SCHALLER, G.B. 1976. Report on a wildlife survey in northern Argentina and in the Emas National Park, Brasil. Mimeographed report to world Wildlife Fund, Merges, 17 pp. SCHALLER, G.B. 1983. Mammals and their biomass on a Brazilian ranch. Arq. Zool., Sao Paulo 31 : 1 -36. SCHALLER, G.B. and P. CRAWSHAW. 1981. Social dynamics of capybaras population. Saugetierkundliche Mitteilungen 29:3-16. SCHALLER, G.B. and J.M.C VASCONCELOS. 1978. A marsh deer census in Brazil. Oryx 14:345-351. SCHRODER, G.D. and, M L. ROSENZWEIG. 1975. Perturbation analysis of competition and overlap in habitat utilization between Dipodomys ordii and Dypodomys merhami. Oecologia 19:9-28.

PAGE 228

217 SICK, H. 1983. Migragdes e aves na America do Sul Continental. CEMAVE, Brasilia, DF, Brazil. 311pp. SILVESTRE FILHO, D.F. and N. ROMEU. 1974. Caracteristicas e Potencialidades do pantanal Matogrossense. IPEA, Brasilia, Brasil, 220 pp. SINCLAIR, A.R.E. 1974. The natural regulation of buffalo populations in East Africa. IV. The food supply as a regulating factor, and competition. E. Afr. Wildl. J. 12:291-311. SINCLAIR, A.R.E., H. DUBLIN, and M. BORNER. 1985. Population regulation of Serengeti wildebeest: a test of the food hypothesis. Oecologia (Berl.) 65:266-268. SMITH, E.P. and T.M. ZARET. 1982. Bias in estimating niche overlap. Ecology 63:1248-1253. SMITH, J.N.M., P R. GRANT, B R. GRANT, I. ABBOTT, and L.K. ABBOTT. 1978. Seasonal variation in feeding habits of Darwin's finches. Ecology 59:1137-1150. SOKAL, R.R. and F.J. ROHLF. 1969. Biometry. Freeman Press, San Francisco, Calif., 859 pp. SPARKS, D R. and J.C. MALECHEK 1968. Estimating percentage dry weight in diets using a miaoscope technique. J. Range. Manage. 21 :264-265. STALLINGS, J.R. 1984. Notes on feeding habits of Mazama gouazoubira in the Chaco Boreal of Paraguay. Biotropica 16:155-157. STALLINGS, J.R. 1986. Notes on the reproductive biology of the grey brocket deer {Mazama gouazoubira) in Paraguay. J. Mammal. 67:172-175. STEHLI, F.G., and S.D. WEBB. 1985. The Great American Biotic Interchange. Plenum Press, New York, N.Y. 532 pp. STRAUSS, R.E. 1979. Reliability estimates for Ivlev's electivity index, the forage ratio, and proposed linear index of food selection. Trans. Am Fish Soc 108:344-352. SUDO, H. 1974. Metodologia de classificagoes climaticas. Boletim do Departamento de Geografia, Faculdade de Filosofia Ciencias e Letras de Presidente Prudente, Sao Paulo, SP, Brazil, 4,5,6:57-81.

PAGE 229

218 SWIHART, R.K. and N.A. SLADE. 1985. Testing for independence of observations in animal movements. Ecology 66: 1 1 76-1 1 84. TER BRAAK, C.J.F. 1989. CANOCO an extension of DECORANAto analyze species-environment relationships. Hydrobiologia 184:169-170. THOMAS, A.S. 1960. Changes in vegetation since the advent of myxomatosis. J. Ecol. 48:287-306. THOMAS, D.L. and E.J. TAYLOR. 1990. Study designs and tests for comparing resource use and availability. J. Wildl. /Wanage.54(2):322-330. THOMPSON, S. 1987. Sampling for estimating multinomial proportions. Am. Stat. 41:42-46. TOMAS, W.M., M.D. BECCACECI, and L. PINDER. 1997. Cervo-do-Pantanal {Blastocerus dichotomus). Pp. 24-40 in Biologia e Conservagao de Cervideos SulAmericanos: Blastocerus, Ozotoceros e Mazama (J.M.B. Duarte, ed.). FUNEP, Jaboticabal, SP, Brazil. TOWNSEND, W. R. 1996. Nyao Ito: Caza y Pesca de los Siriono. Institute de Ecologia, Universidad Mayor de San Andres, FUN-ECO, La Paz, Bolivia, 1 30 pp. UNEP United Nations Environmental Programme. 1991/1992. Environmental report. Oxford, U.K. 407 pp. VAN SOEST, P.J. 1982. Nutritional Ecology of the Ruminant. 0 & B Books. Corvallis, Ore. 374 pp. VELOSO, H P. 1948. Considera?6es sobre a vegetagao do Estado de Mato Grosso: (II) notas preliminares sobre o pantanal e zonas de transigao. Boletim Geografico, Rio de Janeiro 6:798-943. VOSS, W.A., F.R. dos SANTOS BREYER, G.C. MATTES, and H.G. KONRAD. 1981 . Constatagao e observagao de uma populagao residual de Blastocerus dichotomus {WWger, 1811)(Mammalia, Cervidae). Iheringia, Zoologia 59:25-36. WESTMAN, W.E. 1980. Gaussian analysis: identifying environmental factors influencing bell-shaped species distributions. Ecology 61 :733-739. WHITMANN, W.C. and E.I. SIGGEIRSON. 1954. Comparison of line interception and point contact methods in the analysis of mixed grass range vegetation Eco/ogy 35:431 ^36.

PAGE 230

219 WILBUR, H.M., P.J. MORIN, and R.N. HARRIS. 1983. Salamander predation and the structure of experimental communities: anuran responses. Ecology 64:1423-1429. WILLIAMS, W.T., J.M. LAMBERT, and G.N. LANCE. 1966. Multivariate methods in plant ecology V. Similarity analyses and information-analysis. J. Ecol. 54:427^5. WILLIAMS, W.T., G.N. LANCE, L.J. WEBB, and L.G. TRACEY. 1973. Studies in numerical analysis of complex rain-forest communities. J. Ecol. 61 :47-70. WOLDA, H. 1981. Similarity, indices, sample size and diversity. Oecologia 50:296302. WRAY, S. 1994. Competition bet\A/een muntjac and other herbivores in a commercial coniferous forest. £)eer 9:237-242.

PAGE 231

BIOGRAPHICAL SKETCH Laurenz Pincler\A/as born on 22 September 1958 in Sao Paulo, Brazil. He attended Universidade de Sao Paulo from 1978-1981, and received his Bachelor of Science, with a major in zoology, in 1981 . He then attended a postbaccalaureate course in Nuclear Biosciences at the Universidade do Rio de Janeiro, from which he graduated in 1982. He attended the graduate program at the Universidade do Estado do Rio de Janeiro from 1983-1986, and was graduated with a Master of Science degree in 1986. In 1987, he worked for the Golden Lion Tamarin Project of the Smithsonian Institute as associate researcher and field administrator at the Reserva Biologica de P090 das Antas, Rio de Janeiro. In 1988, he entered a doctoral studies program at the University of Florida, and was graduated with a Doctor of Philosophy degree in 1997. He is currently a member of the Deer Specialist Group of lUCN. 220

PAGE 232

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and its fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ronald F. Labisky, Chair ^ Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and its fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Eisenberg Katharine Ordway Professor of Ecosystem Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and its fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Kent H. RedfcJrd Associate Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and its fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. George W. Tanner Professor of Wildlife Ecology and Conservation

PAGE 233

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and its fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. S. David Webb Distinguished Research Curator of Zoology This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 1997 t)ean. College^ of Agriculture Dean, Graduate School


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 EQVFU9AVB_FWBPHI INGEST_TIME 2015-04-13T20:36:00Z PACKAGE AA00029969_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES