Citation
Comparative ecology of ocelot (Felis pardalis) and jaguar (Panthera onca) in a protected subtropical forest in Brazil and Argentina

Material Information

Title:
Comparative ecology of ocelot (Felis pardalis) and jaguar (Panthera onca) in a protected subtropical forest in Brazil and Argentina
Creator:
Crawshaw, Peter Gransden, 1952-
Publication Date:
Language:
English
Physical Description:
xiii, 190 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Animals ( jstor )
Carnivores ( jstor )
Ecology ( jstor )
Female animals ( jstor )
Jaguars ( jstor )
Mortality ( jstor )
National parks ( jstor )
Peccaries ( jstor )
Species ( jstor )
Transmitters ( jstor )
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF ( lcsh )
Jaguar -- Ecology -- Argentina ( lcsh )
Ocelot -- Ecology -- Argentina ( lcsh )
Wildlife Ecology and Conservation Thesis, PH. D ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 181-189).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Peter Gransden Crawshaw, Junior.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Peter Gransden Crawshaw. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022350001 ( ALEPH )
34462261 ( OCLC )

Downloads

This item has the following downloads:


Full Text











COMPARATIVE ECOLOGY OF OCELOT (Felis Dardalis)
AND JAGUAR (Panthera onca) IN A
PROTECTED SUBTROPICAL FOREST
IN BRAZIL AND ARGENTINA
















By

PETER GRANSDEN CRAWSHAW, JUNIOR


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

UNIVERSITY OF FLORIDA

1995


UNIVERSITY OF FLORIDA LIBRARIES

































Copyright 1995

by

Peter Gransden Crawshaw, Junior
































To Mel, advisor and friend Fiona,

and Mara

Without whom I never would've made it in time.













ACKNOWLEDGEMENTS


This study was generously supported by the Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renovaveis

--IBAMA, the World Wildlife Fund - WWF/US, Helisul Taxi Aereo Ltda., the Lincoln Park Zoo "Scott Neotropical Fund", Funda4io O Boticdrio de Conservagdo da Natureza, and Ilha do Sol Turismo. The Conselho Nacional de Desenvolvimento Cientifico e Tecnol6gico - CNPq awarded me with a scholarship for my academic work. The Fundaq~o Pr6-Natureza (FUNATURA) contributed with managerial aspects of the project. Several individuals in these institutions went out of their way to assure an easier flow between finances, bureaucracy, and field work. Noteworthy were Iolita M. Bampi, Francisco Neo, Jordan P. Wallauer, and Alison J. Coutinho from IBAMA; LouAnn Dietz, Victor Bullen, Carmen Pinelli, and Sally Adams from WWF-US; Steven D. Thompson from the Lincoln Zoo; Jessy Alves Pinheiro from CNPq; the brothers Celso, Moacir, and Redi Biasus from Helisul Taxi Aereo; Miguel Milano, Malu Nunes, and Fatima Soares from the Fundagdo Boticario de Proteq&o & Natureza; Alex Schorsch and the brothers Ademir and Renato from Ilha do Sol Turismo, and Maria T. PAdua from Funatura.

Foremost, I thank my committee chairman, Dr. Melvin Sunquist, for his constant support and friendship. I am iv








extremely thankful also to Drs. John F. Eisenberg, Steve R. Humphrey, John G. Robinson, Peter Feinsinger, Kent H. Redford, Jack Kaufmann, George Tanner, and David Webb. I was very fortunate to have had them on my committee, at one time or another, to help and guide me through the burdens of academia, and to relish its pleasures. Kent Redford was a constant driving force for my graduate studies at UF, since our first campfire conversations in the heart of the Pantanal, back in 1978.

Even though it meant an overall delay of two years in my graduate course, I was fortunate to work in another PhD project under Dr. Steve Humphrey. For personal and professional reasons, I had to switch back to my original Iguaqu proposal (and re-start from scratch). However, I will always cherish my memories of bobcats in the Big Cypress National Preserve in south Florida, and of the support and friendship from Steve and the boys at the Florida Fresh Water Fish and Game Commission in Naples. Dave Maehr was instrumental in helping me set up the study there.

The study in Iguaqu wouldn't have been possible without the unconditional help of the Park's director, Dr. Jose Carlos Ramos. He always made sure I had what I needed when I needed it. Terezinha S. Martinez was my buffer between rigid financial rules of the federal government in the Brasilia IBAMA offices and the flexibility required by field demands. Also in the Park's staff, Apol6nio N. Rodrigues and Salete F.








da Costa provided invaluable help and companionship. I also thank all the soldiers of the Batalhdo de Policia Florestal do Parque Nacional do Iguaqu that contributed with the project, especially Cabo Concei9do and Soldado Barreto, and the Lieutenents Almeida and Wellington.

My debt to my field assistants, Sandra M. C. Cavalcanti, Cibele B. Indrusiak, Jan K. Mahler, Lucila Manzatti, Ronaldo G. Morato, Rose L. Gasparini, Ricardo Boulhosa, and Hannes Rudiser, is immense. Their presence and reliability allowed me to fulfill my responsibilities with an overload of field and office work, as well as attending several other projects. I wish them all a brilliant professional future!

The binational character of the project was only made possible by the commitment and enthusiasm of Silvana B. Montanelli, the biologist in Iguazu National Park, Argentina. I hope we can continue to work together for many years to come. Several other people were pivotal to the study in Iguazu: the two Park's directors, Ugo Rossi and, later, HorAcio Giachino, Karina Schiaffino, Laura Malmierca, Carlos Saibene, Pedro Moreyra, Jose L. Comita, and Sergio Acosta. I also benefited from the association with the workers in the "Primate Social Structure" study undertaken in Iguazu, by Drs. Charles Janson and Alejandro Brown. I also thank Juan Carlos Chebez, head of the Wildlife Division for Northeastern Argentina, Administraci6n de Parques Nacionales, for sharing








with me his vast knowledge on the Argentine fauna, especially regarding early studies.

In the stream of students and professionals that passed through Iguaqu, I am thankful especially to Junio A. dos Santos, Leandro Silveira, Renata P. Leite, Marcelo Mazzoli, Murilo Gubert, Andre Levy, and Jim Sparks for their efforts in data collection. Renata P. Leite conducted all the work in chromatography (however frustrated) and identified predator scats for me.

Plants were identified by Marcos Sobral (Faculdade de Farm~cia, Universidade Federal do Rio Grande do Sul), Pedro Luis (Master's program, Universidade Federal de Minas Gerais), Suely (Instituto de Bot&nica de SP), Sandro M. Silva (Universidade Federal do Parana), and Stinger Guala (University of Florida). Ives Sbalqueiro (Universidade Federal do ParanA) and L. Flamarion B. Oliveira (Museu Nacional do Rio de Janeiro) identified small rodents for me.

The project gained significantly from the participation of Wanderley de Moraes, the veterinary from Itaipu Binational. I'd like to believe that all the study animals he treated, when necessary, share my appreciation of him. Also from Itaipu, I'd like to thank Leonilda, Emerson Suemitsu, Helio Fontes, and Fern&o Carbonar for their support at one time or another.

Mel and Fiona Sunquist provided "a home away from home" during the concluding stages of the writing. Their


vii








accumulated expertise and vast bibliography always at hand made all the difference. Their warm hospitality and support will never be forgotten!

The friendship, help, and advice from Rosa Lemos de Sd, Bob Godshalk, Vicky McGrath, Luciano Verdade, Jay Malcolm, and Justina Ray are deeply appreciated. Vicky McGrath also prepared all the maps of Iguaqu for me.

Finally, my gratitude to my wife Mara and to my "children" Danielle, Beatriz, and David cannot be adequately expressed in words. Not many people would agree to the system of priorities that I have to (or chose to) live by. I just hope I somehow can make it up for all the times when I should have been there for them, and wasn't...


viii















TABLE OF CONTENTS


ACKNOWLEDGEMENTS . . . . . . . . . ...

ABSTRACT . . . . . . . . . . . .

CHAPTER 1 INTRODUCTION . . . . .....

CHAPTER 2 STUDY AREA . . . . . .....

CHAPTER 3 MATERIALS AND METHODS . ....

CHAPTER 4 RESULTS: OCELOT ... ........

Radio-telemetry . . . . . . ....
Home Ranges .......... .
Movements ............ . ...
Activity .......... ... .
Social Interactions and Reproduction Dispersal . ..............
Mortality .......... . . ...
Density . . . . . . . . . . ....
Diet . . . . . . . . . . . . . . .

CHAPTER 5 RESULTS: JAGUAR ... ........

Radio-telemetry .......... ..
Home Ranges and Movements . ....
Reproduction . . . ...........
Dispersal ............
Mortality . . . . .........
Density ................ ...
Dietionships w ........
Relationships'with Humans . . .. . . . .


. . . . .. . iv

.........xi




16

. . . . . . 25

. . . . . . 25
* . . . . . . 25
. . . . . . . 25
* . . . . . . 25
. . . . . . . 29
. . . . .. . 30
* . . . . . . 31
. . . .. . 34
. . . . . . . 37 .. . . .. . 39
. . . . .. . 40
........ 78

78

.. .. . 79
. . . . . . . 81
83
86
.. . 88
88
. . . .. . 89


CHAPTER 6 DISCUSSION - COMPARATIVE ECOLOGY OF OCELOT
AND JAGUAR, WITH IMPLICATIONS FOR MANAGEMENT
AND CONSERVATION ....... ................

Management and Conservation ................

APPENDIX A - MAMMALS RECORDED IN IGUAQU ........ . APPENDIX B - SIGHTING RECORDS .... ............


116

131

145 147









APPENDIX C - TRAPPING RESULTS . ............... 149

APPENDIX D - NOTES ON SMALL CATS IN IGUAQU . . ... 167 APPENDIX E - ENVIRONMENTAL EDUCATION ACTIVITIES . 175 REFERENCE LIST ........ ................ . .. 181

BIOGRAPHICAL SKETCH ....................... 190














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

COMPARATIVE ECOLOGY OF OCELOT (Felis Dardalis)
AND JAGUAR (Panthera onca) IN A
PROTECTED SUBTROPICAL FOREST
IN BRAZIL AND ARGENTINA

By

PETER GRANSDEN CRAWSHAW, JUNIOR DECEMBER 1995


Chairperson: Dr. Melvin E. Sunquist Major Department: Department of Wildlife Ecology and Conservation (Forest Resources and Conservation)

Twenty-one ocelots (11 males, 10 females) and seven jaguars (5 males, 2 females) were captured and equipped with radio-transmitters in the neighboring Iguaqu (1,750 km2) and Iguazu (550 km2) National Parks, southwestern Brazil and northeastern Argentina.

Mean home range size for 6 adult male ocelots was 38.8 k and for 5 adult females it was 17.4 km2. Males travelled farther than females between consecutive locations (2.1 � 1.7 km and 1.3 � 1.1 km, respectively). Although more active at night, ocelots in Iguaqu displayed more daytime activity than shown in other studies. This may reflect a greater reliance on more diurnal prey, such as agoutis (Dasvtrocta azarae) and tegu lizards (Tuinambis teguixin).








Dispersal of two subadult male ocelots and one subadult female suggested difficulty in establishing new ranges within the resident population. The process involved long-range, apparently erratic movements, and in one instance, conspecif ic fights. This may have been due to a relatively high density (12 adults/100 km2) and a low turn-over rate.

Mean home range size for 4 adult male jaguars was 88.7 km?; one adult female jaguar had a home range of 70 km2. The mean linear distance between consecutive locations was greater for males than for females. Two dispersing subadult males were able to establish their adult home ranges within Iguaqu, about 60-70 km away from their natal areas. This apparent ease in finding space was likely correlated with a lower jaguar density (3.7 adults/100 km2) and high turnover rate.

Poaching accounted for most of the mortality of both species, affecting jaguars more heavily than ocelots. Genetic interchange occurred between populations from the two Parks, both species crossing the Iguaqu river during dispersal.

Mean prey-predator weight ratios for ocelot and jaguar in Iguaqu were 0.13 (mean prey weight, MPW= 1.4 kg) and 0.18 (MPW= 14.4 kg), respectively. These results are intermediate between those derived from other studies with these species.


Overall, ocelots have a better chance of survival in Iguaqu, mainly due to greater acceptance of humans of living in close proximity to this species, as opposed to jaguar.


xii








Some management recommendations are proposed. An environmental education program that includes working with neighboring land-owners (with compensation for losses due to livestock depredation where applicable) is also considered essential.


xiii














CHAPTER 1
INTRODUCTION





As an ever-increasing portion of tropical and subtropical forests gives way to human development, research on the remaining fragments set aside as national parks and reserves becomes even more critical. In spite of recent studies on this subject (Malcolm, 1993, Lovejoy et al., 1986), much remains to be learned about the natural history of the fauna and flora and the dynamics of these ecosystems, if we are to manage and preserve these islands of habitat. In this context, predators, especially the larger species, play a vital role as indicators of ecosystem integrity in these reserves.

Iguaqu National Park, in southwestern Brazil, vividly exemplifies an isolated fragment of natural habitat surrounded by intensive human activities (Figure 1-1). Created originally to protect the world-famous Igua9u Falls, the Park now represents about 80% of what remains of the once extensive subtropical forest that covered Parana state, east of the coastal Serra do Mar (Poupard et al., 1981). Now only 6% of the state consists of forest cover. Despite the proximity to burgeoning cities, such as Foz do Iguaqu and Cascavel, the








2

Park still harbors a largely intact array of its original fauna, including large predators, such as jaguar (Panthera onca) and puma (Puma concolor), and ungulates, such as peccaries (Tavassu peccari and T. tajacu) and tapir (Tapirus terrestris).

From April 1990 through December 1994, I investigated the ecology of ocelots and jaguars in Iguaqu National Park. In 1991 the project was extended to the neighboring Iguazu National Park, in Argentina, separated by the Iguaqu river. Through the study of the natural history of these species and of the differential impact on them by human activities, I hoped to evaluate their conservation prospects in the Park. Additionally, I wished to determine if there was genetic interchange between the populations of the two Parks, and to study the dispersal patterns (sensu Schields, 1987) of subadults. Studies of these felids in other settings in Central and South America (Crawshaw and Quigley, 1989, 1991; Quigley and Crawshaw, 1992; Sunquist, 1992; Sunquist et al., 1989; Ludlow and Sunquist, 1987; Emmons, 1987, 1991; Rabinowitz and Nottingham, 1986; Konecny, 1989), usually conducted in large tracts of relatively undisturbed, continuous habitats, provided a useful comparative framework.

Chapter 2 presents a brief description of the study area. In Chapter 3, materials and methods used in the study are presented. Chapters 4 and 5 summarize the results on ocelots and jaguar, respectively. In Chapter 6, I compare the ecology








3

of these two species of felids in Iguaqu in the light of results from studies elsewhere. I conclude this chapter with a discussion on management recommendations, as they apply to the conservation of these species in Iguaqu National Park. I also included as appendices some summary results on different aspects of the study.

With the understanding that management and conservation of predators are a direct result of the local human population's attitude towards these species, the project placed considerable effort in an environmental education program in and around the Park. A summary of these activities is included as Appendix E. A similar program was developed in Iguazu National Park, by Silvana B. Montanelli, that Park's biologist.



















LVA NTA M ENTO
AE ROFOTOGR AM ETARICO A T00 AU TR'o
PAROUE NACIONAL
DO IGUAGU
ESCALA - 1100000


Plate I-Mosaic of aerial photos of Iguaqu National Park, Brazil. Note patchwork of agriculture completely surrounding the park on the Brazilian side (see also Figure 2-1).














CHAPTER 2
STUDY AREA




Igua4u National Park (Fig. 2-1) lies on the western border of Parana state, southern Brazil (250 05' to 250 41' S and 530 40' to 540 38' W), along the international boundary between Argentina and Paraguay. Presently 175,000 ha in size, the Park was created in 1939 to protect the Iguagu falls, which are considered one of the most spectacular natural sights in the continent. Situated in a region of rich soils and high human population density (45.8 inhabitants/km2; Poupard et al., 1981 - and most likely much higher at present days), the park is virtually an island of forest, completely surrounded on the Brazilian side by farmland, mostly plantations of soybean, wheat, and rice, and pastures for cattle. Unless otherwise stated, references to "the Park" will denote the Brazilian Iguaru National Park, for the remainder of this paper.

Together with the adjacent 55,500 ha Parque Nacional Iguazu, in Argentina, the Park harbors the last large tract of subtropical rainforest that once covered much of the highlands on the west side of the Serra do Mar, in southern Brazil. Although contiguous, the two Parks are administratively








6

separate. Also represented within the Park, albeit to a lesser extent, is the dwindling Brazilian pine (Araucaria) forest. A preliminary mammal .list identified 42 species of mammals, 13 (31%) of which are carnivores (Appendix A). Within the carnivores, 38% are felids. Despite being one of the oldest national parks in the country, no intensive, longterm research had ever been done in the area. Crespo (1982) conducted a study on mammals in Iguazu National Park, Argentina.

The shape of the Park resembles that of an "L" turned 900 counterclockwise, the average width (N/S) of the western half being only about 5 km (Figure 2-1). The eastern portion, with approximately 50 km on the N-S axis and 23 km on the E-W axis (roughly 1,100 km2), remains the least disturbed. There is no buffer zone or effective fencing between the Park and the surrounding private properties (Figure 1-1), along the 136 km of dry boundary of the northern perimeter of the Park. A dirt road runs along the border of the western half for less than half this distance. The remaining 304 km of the perimeter is separated by rivers: 11.5 km by the lower Iguaqu river (below the falls; mean width= 250 m) on the western end of the Park, 184 km of the upper Igua9u river (mean width= 1.0 km) which limits the Park to the south, and 107 km by the Gongalves Dias river (mean width= 10 m), on the eastern side of the Park.

The close proximity of livestock and crops to the boundary of the Park function as strong attractants to








7

wildlife. Common management problems are the invasion of crops, especially corn, by peccaries and predation of poultry, dogs, pigs, sheep, and cattle by predators. The lack of perimeter roads in a great portion of the boundary hinders efficient patrolling by Park rangers, and facilitates invasion by poachers. A large proportion of property-owners around the Park are of Italian descent, among whom hunting is a cultural legacy. In addition to illegal hunting for sport, there is also subsistence hunting by people of low income.

The study area encompassed about 80 km2 of the westernmost part of the Park, where the falls are located (Figure 2-2). Access to the falls is provided by a federal highway (BR-469), that ends at the tip of the peninsula that cuts into Argentina, at the falls. This area is visited by close to one million visitors a year (mean of 802,587 during the study period, 1990-1994), with peaks in the months of January/February and in July (Figure 2-3). During some national holidays, the number of daily visitors may exceed 10,000.

Most of the trapping and ground monitoring was conducted along the 8.8 km Pogo Preto road (PPR), which ran from the main road (BR-469) to the upper course of the Iguaqu river (Figure 2-2). The research base was located at the begining of this trail, which has been closed to the public since 1986. Three other trails also were used (Figure 2-2): the Represa trail (TR; 1.2 km); the Bananeiras trail (TB; 1.5 km); and the








8

Macuco trail (TM; 4.5 km). The latter was the least used because of heavy public visitation, being used for a jungleand boat trip, run by a concessionaire.

An 18 km, unpaved road used to cut through the Park, about halfway in the east-west length (Figure 2-1), until 1986, when state authorities decided to pave the road. A local NGO took the case to court, first at the state, and then at federal level. The final sentence ruled that not only the road should not be paved, but also that it should be closed to traffic, permanently. The closing was enforced by federal police, in October 1986. Since then, from time to time, politicians of the affected counties (mainly Medianeira and Capanema) still advocate re-opening the road, putting considerable pressure on IBAMA officials, in Brasilia . The final outcome remains to be seen; meanwhile, nature takes its course, and the road today can hardly be seen, from the ground or from the air.

The climate in the region is temperate subtropical, with mean monthly temperatures ranging from 25.70 C in February to 14.60 C in July (Crespo, 1982). The extremes during the study were -70 C and 470. Mean annual rainfall during the study period (1990-1994) was 1700 mm, with one peak in May (180 mm) and one in October (240 mm). July and August are the driest months of the year (Figure 2-4).

The vegetation in the Park consists of a rather uniform forest cover, with some regenerating patches on the western








9

half of the Park, where settlers had cleared or selectively logged the original forest. (Note: In 1975 there were over 400 families living in this area; the last resident was compensated and removed from the Park in late 1986). There are also a few small, apparently natural clearings, usually associated with wetlands.

A preliminary survey indicated 312 species of vascular plants, comprising 75 families, in Iguaqu National Park. Four families (Graminae: 24 species, 8%; Euphorbiaceae: 19 species, 6%; Compositae: 18 species, 6%; and Rubiaceae: 17 species, 5%) represented 25% of the total number of species. Twenty-seven families were represented by only one species, indicating a high diversity of plants. The vegetation in two 0.13 ha plots along the PPR, selected for small mammal trapping, was studied in more detail.

Plot A was located at km 0.7 of the PPR, and consisted of secondary forest, which was selectively logged in the past. Canopy height was between 10 and 15 meters, with emergent trees reaching 25 m (mainly Parapiptadenia riQida, Rurechtia laxiflora, Peltophorum dubium, and Avuleia leiocar )a). The dominant trees is this layer were Nectandra lanceolata and N. megaDotamica, Diatenoptervx sorbifolia, Cedrella fissilis, and Allophvlus edulis. Trees in the intermediate story had an average height between 1.5 to 4.0 m, consisting mainly of Trichilia elegans and T. caticrua, Allohvlus quaraniticus, e amalago and P. gaudichaudianum, Psychotria myriantha and










F. leiocarva. At smaller densities, Pilocarpus Dennatifolius, Ilex Daraguavensis, introduced Citrus sp., Hennecartia omDhalandra, among others, also occurred. Noteworthy was the presence of two species of bamboo Merostachys sp. and Bambusa trinii. The latter species had apparently gone through a recent die-out. The bottom story, with plants up to 1 m, was composed predominantly of ferns (Dryoteris sp. and Pteris sp., and Didimochlaena truncatula), with Hydrocotyle cf. leucoceDhala in high densities also. Pharus cf. Qlaber, Oly sp., Adiantopsis radiata, Doryopteris sp., GeoDhila macropoda, Pavonia malvacea and Saranthe sp. were also present. Both in the intermediate and lower strata, there were high densities of young individuals of the taller trees from the tree canopy, such as ParaDiDtadenia rigida, Diatenopteryx sorbifolia, CamDomanesia xanthocarpa, Patagonula americana, Balfourodendron riedelianum, and Sorocea bonplandii, among others. Vines and lianas are common in the area, mainly Pristimeria andina, Bauhinia microstachya, Wilbrandia sp., Macfadvena unruis-cati, Arrabidea cf. mutabilis, Serjania sp., Acacia sp., and Chusauea sp.. Philodendrom selloum was a common epiphyte.

Plot B was situated in primary forest, at km 4.8 of the PPR. The most conspicuous difference between the two plots was the high density of Euterpe edulis in Plot B (70% of all trees with > 10cm DBH present, n= 101), and its complete absence in plot A. The tree canopy (height from 10 to 20 m)










was composed mainly by Euterpe edulis, Cedrella fissilis, Holocalyx balansae, Alchornea sidifolia, Cabralea canjerana, Jacaratia spinosa, Nectandra megapotamica, Chrysophyllum gonorpw, and Guarea kunthiana. Less common were Plinia rivularis, Camoomanesia xanthocarpa, Cecropia sp., Ficus cf. guaranitica, Lonchocarous muehlbergianus, Inga marginata, and $yagrus romanzonfianna. In the intermediate story, with an height between 1.5 and 3.0 m, predominated Piper amalago, P. gaudichaudianum, Psychotria myriantha, P. leiocarpa, and Sorocea bonplandii; some individuals of Eugenia burkartiana and Prockia crucis are also found. In the herbaceous layer, Didymochlaena truncatula, Drvopteris sp., Pharus cf. glaber, and Geophila repens. Compared with Plot A, ground cover was notably less, Hydrocotyle cf. leucoceohala (common in Plot A) was restricted to the influence of the road (the plots started about 6 m from the road, at a right angle). There were many very young individuals of the larger trees, especially of Euterpe. There were also some lianas, including Pristimeria andina, Bauhinia mvcrostachia, and Pisonia aculeata.

Considering only the trees with > 10 cm DBH, 54 species with a total of 196 individuals were identified in Plot A, as compared to 19 species with 101 individuals in Plot B.
















Sta Teresa


to Cascave!


t:::::::::: Iguaqu National Park





Matelindia Medianer

S o Miguel do lgua >
Zga~


Cu Azul


Brazil


Argentina


Study Area


Figure 2-1--Map of Iguagu National Park showing study area.


10 km










Iguacu National
Park


Brazil


'1


Administration
Building


Iguazu National
Park


Argentina


1.5 k]a


Figure 2-2--Map of study area in Iguagu National Park, Brazil.


Iguaqu Fa













Vehicles {x 1000)


04i i i i i
Jan Feb Mar Apr May Jun


Jul Aug Sep Oct Nov Dec


1 --Visitors --X-- Vehicles I


Figure 2-3--Number of visitors and vehicles entering Iguaqu National Park, Brazil, during the study period.


120 100

80 60

40

20


Visitors (x 1000)












300


200


150


5 , I I I I I , I I I 5
J F M A M J J A S 0 N D Months


� Temperature -- E-.-Rainfall


Figure 2-4--Mean monthly temperature and rainfall for Iguaqu National Park, Brazil, 1990-1994.













CHAPTER 3
MATERIALS AND METHODS



Animals were captured using custom-made wood, wire-mesh or iron-bar box-traps of different sizes, depending on the species. Measurements (length x width x height) of traps used for ocelots were 120 x 40 x 50 cm, and for jaguar, 210 x 80 x 80 cm. Traps were placed mainly along the 8.8 km of the PPR, although several other sites on the western-most part of the Park were also trapped. The PPR was marked every 100 m with numbered posts for permanent location reference.

Live bait (chickens or laboratory rats) was housed at the back of the trap with wire mesh; once captured, the trapped animal could tear through the mesh and feed on the bait, thus reducing thrashing and stress while in the trap. Live piglets were used for two of the jaguar captures in Igua9u, and remains of dogs that had been previously killed by problem jaguars were used in two captures in Iguazu, Argentina. Capture attempts using a live goat were unsuccessful, although the jaguar circled and even climbed on the trap, without entering. The use of live bait also reduced the incidence of nontarget species, such as opossums (Didei his) and tegu lizards (Tuinambis), in the traps.








17

On four occasions, trained dogs were used on one successful jaguar recapture and one unsuccessful capture and two recapture attempts. In one of these, a problem jaguar wounded with buckshot by a local when preying on livestock killed 4 dogs and mauled another 5 during the hunt, and 1 dog was killed in another hunt. The denseness of the vegetation impeded keeping close to the dogs, after they were released on the cat's trail.

Once trapped (or treed, in the case of jaguar recaptures), animals were chemically restrained using a projectile dart shot with a CO2 pistol (Telinject U.S.A., Inc., Saugus, CA, 91350), or powder rifle (Capchur, Palmer Chemical & Equipment Co., Douglasville, GA 30133). Some were hand-injected using a noose (Ketch-All, Inc., S. Diego, CA 92104) to restrain the animal. Two drugs were used: ketamine hydrochloride (Parke-Davis & Co., Detroit, MI), and Zoletil (same as Tilazol or CI-744; Virbac do Brasil, S~o Paulo, SP 04021). Mean dosage of Ketaset for ocelots was 19.7 mg/kg � 9.9 mg/kg (range: 10.0 - 40.4; N= 12;) and for Zoletil, the mean dosage was 6.5 mg/kg � 2.4 mg/kg (range: 3.2 - 13.3; N= 17). For the jaguar, the average rate for Ketaset was 29.9 � 9.8 mg/kg (range: 18.2 - 40.4; N= 4), and for Zoletil, 7.0 � 1.0 mg/kg (range: 5.6 - 9.1; N= 16). Mean induction time with Zoletil was about 5 minutes, less than half the time required with Ketamine.








18

Anesthetized animals were examined for general body condition, measured, weighed, ear-notched and/or tattooed, photographed, and fitted with radio-collars (150-152 MHz; Telonics, Mesa, AZ 85204; and Wildlife Materials, Inc., Carbondale, IL 62901). Relative age was estimated mainly on the basis of presence of milk or permanent dentition, on tooth color and wear (juvenile, subadult, adult), and these features were used in combination with other physical characteristics, such as weight, size, sign of previous reproduction (for females), texture and color of the skin of the heel pads (Crawshaw, 1992). Biometric data on captured carnivores are given in Appendix C. Anesthetized animals were kept under observation until ambulatory.

Blood samples were collected from most animals and sent to the biochemical lab at Itaipu Binacional, for clinical and health evaluation of individuals. Blood samples from animals in the latter part of the study were additionally preserved in EZ-Blood, and sent to the National Cancer Institute, Frederick, MD, as part of a collaborative study on the genetics of neotropical felids (O'Brien et al., 1993). Results on these aspects of the study are still pending, and will be published elsewhere. Mean transmitter life for 10 ocelot radiocollars from Wildlife Materials, intended to last an average of 24 months or 720 days, was 165 � 160 days (range 6 - 507), or an average of 23.2% � 21.9% (range: 0.8 - 70.4%) of the projected operational life. Eight transmitters used in








19
six ocelots (2 M, 4 FF, including recaptures) failed after < 3 months of use. For comparison, mean operational life for 3 Telonics transmitters was 144.0% � 15.0% (range: 127% - 155%) of the projected life. Another 6 Telonics transmitters still in use have already worked, on average, 73.7% � 11.6% of their projected life. (Note: only new transmitters were considered in this analysis). I lost contact with two jaguars wearing Telonics transmitters, but I do not know if the cause was radio malfunction or if the animals were killed and the transmitters destroyed.

Radio-equipped animals were searched for from a vehicle, with a mounted omnidirectional antenna, or on foot with a directional antenna. Once a signal was heard, locations were obtained through triangulation, using the hand-held directional antenna. Given the limited ground range of the equipment, about 300m because of the dense vegetation, a light aircraft was used at approximately 14-day intervals to obtain an unbiased sample of locations. The "fly-by" method described by Mech (1983:69) was used to define animal locations from aircraft. Accuracy tests performed by project personnel indicated that locations could be described within a circle of 50 m radius (0.8 ha).

One-hundred-eighty-four flights were made, totalling approximately 460 hours (average flight time= 150 minutes, mean of 4.5 animal locations per flight, range 1 - 16).








20

Seventy-six percent of the 904 locations of ocelots and 81% of the 236 jaguar locations were aerial.

Locations were plotted on a 1:25,000 map of the study area, divided into 1.5 ha cells by a transparent grid overlay. Grid coordinates (vertical and horizontal) were assigned to each location. In the latter period of the study (subsequent to 1992), geographical coordinates were also obtained with a GPS (Global Positioning System) receiver (Transpak, Trimble Co., and Pronav 100, Garmin Co.). The GPS was particularly useful on jaguar recaptures, using dogs, and on locating transmitters on mortality mode on the ground, after initially locating the signal from the air and marking the coordinates.

Home ranges were estimated with the Home-Range (Ackerman et al., 1990) and the Mcpaal (Micro-computer Program for the Analysis of Animal Locations, Stuwe, 1985; National Zoo, Washington, DC) softwares. Results are provided for the Minimum Convex Polygon (Mohr, 1947), the Harmonic Mean (Dixon and Chapman, 1980), and the 95% Ellipse (Jenrich and Turner, 1969) methods, for comparison between methods and between studies. However, due to the particular shape of the study area, with sharp boundaries represented by the Iguaqu river (Figure 2), most of the estimates resulting from the two latter methods included areas known not to be used by some of the animals. Therefore, comparisons between individuals in this study were made using the Minimum Convex Polygon, corrected upon visual analysis of plotted locations. For








21

comparative purposes, only the home-ranges of adult animals were used. Outliers were omitted from calculations, including locations of dispersing subadult individuals.

A new method was devised to estimate home range size. The arithmetric center (AC) of the locations is determined, expressed by the means of the X and Y coordinates (White and Garrot, 1991). Then the distance of all locations to the AC is calculated. The home range is defined by a circle of which the radius is the mean distance of locations to the AC, added to 2 standard deviations. A circle thus defined encompasses, on average, 95% of the locations. This method was used herein to compare spatial organization of the home ranges and as a measure of site fidelity between years for animals with over 24 months of monitoring. The coefficient of variation (CV) was used as a measure of the dispersion of locations in relation to the AC (Ott, 1984).

Linear distances between simultaneous locations were used to analyze spacing between adult conspecifics with overlapping home-ranges. Only those locations obtained within a 40-min. period of each other were used for this analysis.

All transmitters contained two types of motion sensors (Telonics, Mesa, AZ 85204, and Wildlife Materials, Inc., Carbondale, IL 62901). One (mortality sensor) increased the signal pulse rate when the collar remained immobile for > 2 h, indicating the animal had died or dropped the collar. The other (instantaneous activity sensor) changed the pulse of the








22
signal depending on the angle of the collar (head-up, headdown). Signal pulse rate (active or inactive mode) was recorded at 15 minute intervals (Quigley et al., 1979; Sweanor, 1990). A 45 m-high fire-tower close to my house and a 25 m tower at the hotel by the falls were used to monitor activity of animals within range (< 4 km), for periods of up to 12 hours. On several occasions, the signal of animals radiocollared in Argentina were picked up from these towers. Too few activity records were obtained from the radio-tagged jaguars, however, because animals were rarely within range of these towers.

A photographic camera remote system, activated by an infrared beam (Trail Master, Shawnee, KS) was used at the beginning of the study to document the presence of the different species of carnivores, and to aid in density estimates. However, with a few remarkable exceptions, the system was too often triggered by the abundant nocturnal insect life in Iguaqu, and its use was discontinued.

All sightings of carnivores and potential prey species were recorded (Appendix B), together with location and time of day. Likewise, all roadkilled animals found in the Park and periphery were collected, and pertinent information on species, sex, age, health condition, reproductive data, date, and location was recorded. Whenever possible, the stomach and intestines were collected for analysis of contents.







23

Carnivore fecal samples were routinely collected along roads and trails, dried in large cardboard boxes with lamps permanently lighted, and stored until analysis. Contents were separated into the various components in running water over a wire mesh sieve, and identified to the lowest taxon possible, using a local reference collection of hair, bones, reptile scales, and fruit. Some items were identified at the Cap~o da Imbuia Museum, in Curitiba, PR. The predator species that left the scat was identified at the collection site, whenever possible, by association with tracks and other circumstantial evidence. When this was not possible, hairs ingested in autogrooming were identified by comparison of cuticule and medulary patterns with hairs of known samples (Pereira Leite, pers. comm.). For this analysis, hairs were selected from samples by macroscopic characteristics, clarified with xylene, and compared with known samples at 400x magnifications. Regurgitated matter (vomits) and stomach and intestinal contents were analyzed in a similar fashion.

As an additional attempt to identify predator scats, thin layer chromatography (TLC) was used (Johnson et al., 1984; Watt, 1987; Jimenez, 1993). Approximately 150 samples collected from captive animals of 14 species of carnivores were used to determine patterns (if any) of presence and behavior of 20 bile acids (12 identified, 8 unidentified). The tests were conducted at the College of Pharmacy lab, Universidade Federal do Parana, by a veterinarian with








24

extensive laboratory experience, under direct supervision from Dr. Cid Aimbird, head of the Chromatography lab. By comparing the results of the control runs and field collected scats, we concluded that the inherent variation within and between groups was too great to facilitate reliable identification of unknown predator scats. Similar results were reached by Watt (1987) and Jimenez (1993).

In an attempt to correct for differences in prey size in the diet of these predators, I multiplied the mean weight of each food item by the number of times it was found in the sample of scats, and by the mean number of individuals per scat, thus obtaining an estimate of biomass consumed for each taxa (Bt). When available, I used weights from local animals; if unavailable, weights from the literature were used. The relative importance of each prey item was then expressed as the percentage of that item in relation to the combined weight of all items.














CHAPTER 4
RESULTS: OCELOTS





Radio-Telemetry

A total of 21 ocelots (11 MM, 10 FF) was captured and radio-tagged, 15 (9 MM, 6 FF) in Iguaqu, Brazil, and 6 (2 MM, 4 FF) in Iguazu, Argentina. Information on trapping results and measurements are given in Appendix C.

The cumulative monitoring period for the 21 ocelots was 7,401 days, 3,871 (52%) for males, and 3,530 (48%) for females (Table 4-1). The total number of locations was 904, of which 695 (77%) were aerial. Mean interval between consecutive locations was 8.9 � 10.1 days (range= 1 - 85 days). Thus, there was no risk of incurring problems with serial correlation or restricted access to the animals (Ackerman et al., 1991). Locations for some of the study animals are depicted in figures 4-1 through 4-15.



Home RanQes

Male ocelots used home range areas that were about twice as large as those of females. Estimates (Minimum Convex Polygon) of home range size were calculated only for animals with over 30 locations or which were monitored for more than 25








26

3 months (Table 4-2). Outliers, or atypical movements, were eliminated from estimates of home range size, but in a few cases some distant locations were included because of repeated use of these areas at 3- to 4-month intervals. The mean home range size for 6 adult males was 38.8 � 11.8 km2 (range= 20.4

- 50.9 km2) and for 5 adult females 17.4 � 16.7 km2 (range= 3.8

- 40.4). For some animals, the cumulative area curves (Odum and Kunzler, 1955) reached an asymptotic limit after only 10 or 20 locations, whereas for others the home range sizes were still increasing at 50 locations (Figure 4-16, 4-17), suggesting that the full extent of their ranges had not been reached. This continual increase in home range size, especially for males, is likely a reflection of their occasional visits to distant parts of their range.

It is noteworthy that the three females for which the curves reached an asymptote with less than 30 locations (F03, F09, and F40), were all presumed to have small young (see Reproduction) during the period of monitoring. Even though monitored for short periods, F16 and F39 were included to show trend of large home ranges.

Home range estimates using the Harmonic Mean (HM) and 95% Ellipse methods (ELL) are included in Table 4-2, for comparison (see also Figure 4-18, 4-19). The MCP results were consistently more conservative than the other methods. The mean difference between the estimates obtained with the MCP and HM methods was 380% � 126% (range= 212% - 595%) for








27

females, and 398% � 84.4% (range= 295% - 559%) for the males. Between the 95% Ellipse and the MCP, the difference was 227% � 123% (range= 99% - 464%) for females, and 196% � 26.6% (range= 172% - 242%) for males.

Two adult females, F06 and F19, were captured early in the study and were monitored for 507 days (N= 61 locations) and 586 days (N= 68 locations), respectively, before their transmitters failed. After intervals of 572 days for F06 and 349 days for F19 without contact, they were recaptured and reequipped with new transmitters (Table 4-1). Since recapture, they have been monitored for 448 days (N= 30 locations) and 424 days (N= 42 locations), respectively. In the first periods of monitoring, F06 used an area of 39.2 km2 and F19 used 22.6 k&2. After recapture, they used areas measuring 24.0 km2 and 17.0 ki2, a reduction of 38% for F06 and 25% for F19.

Home range sizes calculated using the Circle method for 50% and 95% of locations are given in Table 4-3. The mean home range size (95%) for adult males using this method was 46.7 km2, and that for females was 18.4 km2. Males and females showed the same increase in size between the 50% and 95% estimates (z= -0.447, P=- 0.65; Mann-Whitney two-sample test); the 50% estimate was, on average, 4.3 times smaller than the 95% estimate. This indicates a more concentrated use of the area closer to the center of the home range than toward the periphery, for both males and females. The difference between








28

estimates obtained with this method and those with the MCP was 120% for males and 106% for females.

The mean distance of locations to the arithmetic center was significantly greater for adult males (2.4 � 0.6 k m; N= 7; 352 locations) than for adult females (1.5 � 0.4 km; N= 7; 346 locations; Z= -2.49, P=- 0.013, Mann-Whitney test; see Figure 4-20).

Adult males seemed to use their home ranges in a more uniform pattern, over 50% of their locations were at distances between 1.0 and 3.0 km from the AC (Figure 4-21). Females, on the other hand, concentrated their movements in areas less than 2.0 km from the AC. Subadult females show an apparent tendency to remain in a small area, with over 60% of the locations within less than 1 km from the AC. The greater distances shown for the subadult males denotes the exploratory nature of their movements.

The arithmetic centers of home ranges of same sex ocelots were widely spaced. Among all resident males, the mean distance between the AC of locations was 4.3 � 2.2 km (range= 1.4 - 8.5), and for adult females it was 4.6 � 2.2 km (range= 1.1 - 7.5; Table 4-4). A comparison of the distances between the AC of locations of radio-tagged ocelots show some potentially meaningful associations between animals. Noteworthy are the short distances between adult male M04 and M26 and M27. Among the females, the data support the assumption that F03 was the mother of F39, and F09, that of








29

F40 (see Reproduction). The distances shown between males and females confirm the association between M05 and F03 (see Reproduction), while this male was still alive (see Mortality). He had already died by the time F39 became an adult, and I believe he may have been her father. M35 was believed to associate to F09, likely being the father of F40 (see Reproduction). Other possible associations are between M45 and F06, and M27 and F03 (although from their ages, it is possible that he is F03's son with M04).


Movements

The denseness of the forest and the lack of trails precluded regular ground following of radio-tagged animals. However, on 23 occasions I obtained more than 1 location in the same day for 3 males and for 5 females, with a mean interval between locations of 7.3 � 2.3 hours. In two instances, no movement was detected. On the remaining 21 occasions, the mean linear distance between these locations for the males was 1.9 � 0.5 km (range= 1.2 - 2.6 km; N= 4), and for the females, 0.6 � 0.5 km (range= 0.1 - 1.5 km; N= 17). The overall mean was 0.9 � 0.7 (0.1 - 2.6) km.

A one-way analysis of variance on the linear distance

(km) between consecutive locations with a one-day interval for the various age/sex groups (see Table 4.5 for descriptive statistics) failed to show significant differences (F= 2.90, P > 0.05). However, Fisher's Least Significant Difference








30

Test indicated that adult males travelled significantly greater distances than adult females and subadult females. Furthermore, pair-wise comparisons using Mann Whitney's two sample test indicated significant differences between the distances travelled by adult males x adult females (Z= 1.603, P= 0.108), adult males x subadult females (Z= 2.201, P= 0.028), subadult males x subadult females (Z= 2.417, P= 0.016), and adult females x subadult males (Z= 1.467, P= 0.143). Figure 4-22 shows the distribution of mean linear distances between consecutive locations for the different age/sex classes. Males tend to have an unpredictable movement pattern, in that they can be at any distance from the previous location. All other age/sex categories seem to have more predictable patterns.



Activity

A total of 1,115 activity readings was obtained on the study animals, with a mean of 46 � 19 (range= 4 - 78) readings per hourly interval. Overall, ocelots were more active during nighttime (1800 h to 0559 h: 41% of 627 readings) than during daytime (0600 h to 1759 h: 34% of 488 readings; Z = -2.505, P = 0.012; Two-Sample Proportion Test, Hintze, 1987). After a brief peak of activity at about 1400 h (based on 34 readings), there was a steady increase in activity level from 1600 h to 2000 h, then a decrease followed by two other peaks, a lesser one at 0200 h and a higher one at 0600 h (Figure 4-23).








31

There was no statistical difference between levels of activity of males x females. However, males were significantly more active at night than during daytime (43.6% of 289 readings and 31.4% of 210 readings, respectively; Z= 2.758, P=- 0.0058, Two-Sample Proportion Test, Hintze, 1987).


Social Interactions and Reproduction

Four of the seven adult females (F03, F06, F09, F19) were at different stages of pregnancy when captured, and F40 had recently given birth (Figure 4-24; see Appendix C, Table C-1 for dates). Birth dates were estimated from monitoring pregnant females and by back-dating from the estimated ages of young animals.

When captured on 23 August 1990, the mammae of F03 were swollen. On 9 September, I walked in to check a mortality signal from her transmitter, in very dense undergrowth, and approached the transmitter to within < 15 m. As I searched for her presumed carcass, the signal suddenly returned to the active mode and the animal rapidly left the area. On investigating the site, I found where she had been lying, in a den under a fallen log. Since the mortality mode of the transmitter is only activated after 2 hours of immobility, she must have remained that way for at least 4 hours (including the time since I first heard the mortality signal). A similar instance had once occurred with a radio-collared female bobcat (Felis rufus) that gave birth during a study in the Big








32

Cypress National Preserve, FL (Crawshaw, 1988). On 11 September, I once again followed the mortality signal from her transmitter, 0.9 km from the previous site. However, when the signal resumed the active mode, I left without further disturbing her. On 16 September, when she was recaptured, she had already given birth. I also discovered that her transmitter had failed, and I replaced it with a new one. Until 2 February 1991, when the new transmitter failed also, she was located 19 times, with a mean distance between locations of 1.7 � 0.9 (0.0 - 3.1) km. On 8 occasions she was found close to her supposed den site, and I strongly suspect she had young there. It is interesting that M05, whose home range encompassed all of the known area of F03, was located on two occasions in the den area. On one of the locations, on 24 September, he was apparently with the female. Despite intensive trapping efforts in that general area, this female was not recaptured.

Three and a half years later, on 13 and 19 February 1994, F39 and M37 were captured at the same site, within F03's former range. Their ages were estimated at approximately 4 and 1.5 years, respectively. Given that the locations of both animals (before M37's dispersal and F39's death) were concentrated over F03's former range (Figures 4-14, 4-10, and 4-7, respectively), it is tempting to suggest that F39 may have been F03's offspring, from the 1990 litter, and that one of these females was M37's mother.








33

Between 18 October and 20 December 1990, F09 was located 5 times, with a mean linear distance of 0.8 � 0.6 km (range= 0.0 - 1.8 km) between her locations, suggesting she may have young. On 19 February 1992, 15 months later, M25 was captured in the same area, at an estimated age of 1.5 years and weighing 10 kg. Unfortunately, his transmitter failed soon after capture, only 3 locations having been obtained, all of which were within F09's home range. Neither of the two were recaptured. However, on 21 September 1993, an adult male (M35, estimated age of 6 years) was captured within F09's range, and part of his home range encompassed all of that female's known range (Figures 4-5 and 4-9). In addition, F40 (estimated age of 2.0 years) was also captured in the same area, and her range overlapped entirely the northern portion of F09's former range (Figures 4-11 and 4-9, respectively). Again, judging by the combination of data on site of capture, estimated age, plotted locations, and sex, it is tempting to suggest that M35 and F09 are the parents of M25 and F40, from subsequent litters. If this is the case, it is likely that M35 is also the father of M40's litter. Hopefully, genetic analysis of blood samples from the two latter captures will shed some light on the relationship between them.












A total of 7 ocelots (4 MM, 3 FF) was captured as subadults, 3 males (M25, M37, M41) and 1 female (F02) in Iguaqu, Brazil, and 1 male (M36) and 2 females (F23, F42) in Iguazu, Argentina. The transmitter of M25 apparently failed soon after capture, after only 3 locations. M36 was killed about 13 km from his capture site, about three weeks after capture, and the collar was destroyed (see Mortality).

Three ocelots (2 MM, 1 F) crossed the Iguaqu river from one Park to the other. The two subadult males (M37 and M41) crossed from Iguaqu to Iguazu, and the female (F23), who was initially captured in Iguazu, was later recaptured in Iguaqu.

M37 was captured on 13 February 1994, and he was monitored for 183 days in the Brazilian park. During this time he was located 15 times at intervals of 11.2 � 6.6 days (range= 3 - 32 days) and he moved an average of 1.3 � 0.7 km (range= 0.3 - 2.5 km) between consecutive locations. On 15 August 1994, he was located in the Argentinean park, 0.6 km from his Brazilian location on 31 July. On 16 August, he was again back on the Brazilian side, 0.9 km from the previous location. He then remained in Iguaqu for another 107 days, with a mean linear distance of 1.0 � 0.6 km (range= 0.1 - 2.3 km) between consecutive locations (N= 9). Between 30 November and 15 December, he returned to the Argentinean side, moving an average of 7.6 � 4.9 km (range= 2.0 - 14.0 km), between 3 locations. On 15 December, he was once again located in








35

Iguagu, and remained there until 17 February 1995, during which time the mean distance between 5 consecutive locations was 2.8 � 3.0 km (range= 0.4 - 8.7 km). On 17 February, he had crossed the river one more time and has remained in Argentina since, with a mean distance between 4 locations of

3.4 � 1.6 km (0.9 - 5.4 km).

M41 was captured in Iguagu NP on 4 April 1994. He remained there for 35 days, during which time he was located 12 times with a mean distance between locations of 4.0 � 4.0 km (range= 0.3 - 15.6 km). On 9 May, he was found in Iguazu NP, Argentina, 3.1 km from the location on the previous day. When located again on 29 May, he had returned to Brazil, 4.3 km from the previous location. On 29 November, he was again located in Argentina. Over the next 16 days, he was located 5 times, at a mean interval of 5.8 � 4.2 days (range= 1 - 13 days), with a mean distance of 5.4 � 3.5 km (range= 1.0 - 9.6 km) between consecutive locations. On 24 January 1995, he was once again found in Igua9u, where he has remained there since. During this period, the mean distance between 5 locations was

1.5 � 0.6 km (range= 0.6 - 2.2 km).

F23 was first captured as a subadult in Iguazu NP on 25 January 1992, but her transmitter failed soon after capture. She was recaptured in the Brazilian park on 28 August, after being sighted feeding on the carcass of a roadkilled crabeating fox (Dusicvon thous). She was very lean (7.9 kg against 10.0 kg on the first capture), and had been recently








36

involved in a fight, presumably with a resident animal (intraspecific?). Her right ear was missing and she had open gashes on the chest and limbs. She was taken into captivity for 55 days, for treatment, and was released, after recovery, close to the recapture site on 22 October, at a weight of 12.2 kg. The mean linear distance between 21 locations during the next 217 days was 3.4 � 2.4 km (range= 0.1 - 7.7 km). After 3 April 1993, her signal was not heard despite extensive aerial searches. On 27 May 1993, the mortality signal from her collar was heard, and a location indicated the collar was in the Iguagu river, 30.5 km from her previous location. Although the collar could not be recovered, its location suggested she may have been poached for if she had drowned trying to cross back to Argentina, her carcass would have been washed to the shore. I later confirmed this theory (see Mortality).

F02 may have been another case of female dispersal. She had been captured in July 1990 on a property adjacent to Iguaqu, and was kept in captivity for about 2 weeks before I was informed. She was then radio-collared and released in the Park, at a straight-line distance of 6.0 km from her capture site. By the following day she had returned to that same area, 5.0 km from the release site. After 101 days, her collar failed. During this period the mean linear distance between 36 locations was 0.8 � 0.9 (0.0 - 5.0) km. On 24 April 1991, she was recaptured and her transmitter was








37

changed. In the 71 days this collar transmitted, before also failing, the average distance between 21 locations was 1.1 � 0.7 km (range= 0.0 - 2.2 km). On her last location, on 9 July, she had apparently followed the remnants of gallery forest along the Iguaqu river outside the Park towards the town of Foz do Iguaqu, and was found 7.0 km from her previous location, on 26 June, 1991.



Mortality

One female (F39) was killed by a bus outside of Iguaqu, on the night of 15 August 1994. The driver delivered the carcass, still with the radio-collar, at the' entrance of the Park. The radiocollar of another female, F16, was found by the main road in Iguazu NP, Argentina. Because it had been cut open, I assumed the animal was likely killed by a vehicle or shot, her whole carcass being taken.

I recently confirmed the death of two other study animals. F23 was killed on the Brazilian side by an Argentinean poacher and the collar was thrown in the Iguaqu river (see section on Dispersal), as the poacher returned to Argentina. M36 was killed by a farmer while the cat was raiding a chicken coup by the ParanA river, where it divides Argentina and Paraguay, about 13 km from his capture site. It is possible that some of the animals (F02, M25), whose signals disappeared were poached also, having had their collars destroyed. This may have been a negative by-product of local








38

publicity on the project and on radio-telemetry, since in the early days of the study I followed the signals of two collars (from a fox and a jaguar) to the poachers' homes, who had taken them as souvenirs.

The carcass of M05 was found, by following the mortality signal, in a dense bamboo undergrowth. His already decomposing body showed evidence of a fight, although the surrounding vegetation was undisturbed. After being badly wounded, the animal had likely searched for a secluded place to die. Considering the overlap between the ranges occupied by M05 and M04 (older and larger than M05), and the fact that immediately after M05's death, M04 extended his home range over all of M05's area, it seems likely that M05 died from injuries sustained in an intraspecific dispute with M04.

Following Trent and Rongstad (1974) and Heisey and Fuller (1985), I calculated survival and mortality rates for the ocelots in Iguaqu (Table 4-6) using the formula:

S=__ i-Vi and m-- l-si,

where s, is the daily survival rate, x, is the total number of transmitter-days during interval i, Y, is the total number of mortalities occurring during interval i, and mi is the daily mortality rate in interval "

Overall daily survival rate (s,) and daily mortality rate

(mi) for the radio-tagged ocelots were 0.9944 and 0.0056, respectively. Considering only full years of monitoring (1991-1994), the mean daily survival rate was 0.9914 � 0.0114








39

(Table 4-6). The cumulative mortality reached an asymptote at 24% after 600 days of monitoring (Figure 4-25), with no more known mortalities for the remainder of the study (May 1995).


Density

Six adult males were radiocollared in the study area in INPBr. One of them, M05, was killed, likely by M04, whose home range thereof encompassed the previous range of M05. The home ranges of these five males covered the entire study area. Five adult females were radiocollared and monitored. Two of the females (F06, F39) included in their ranges some small forest fragments outside of the Park boundary. Based on an empty space between the ranges of these females, I assumed there was at least one (possibly two) adult female that was not captured. Therefore, I estimated a density of 13.7 adult ocelots/100 km2 in the study area (or 1 adult ocelot/8 km2; or 0.12 adult ocelot/kM2). In addition, four subadults (3 MM, 1 F) were captured in the same area during the study period. Assuming a mean of 1 young per adult female/two years, this would give a total of 14 animals (5 adults males, 6 adult females, 3 subadults) using the 80 km2 study area, or a density of 17.5 ocelots/100 km2 (or 1 ocelot/5.7 km2). Assuming a homogeneous distribution over the 1,750 km2 of the Park, the total estimated population would be approximately 300 ocelots.












A total of 362 carnivore scats were found in Iguagu National Park during the study period. Fifty-six (15.4%) were positively identified as ocelots scats. Eighty food items of 18 different taxa were recorded (Table 4-7), with an average of 1.4 items per scat. Mammals comprised 80% of the total number of items (Figure 4-26), with small rodents accounting for almost half of the mammals consumed in terms of frequency of occurrence (Figure 4-27), followed by opossum and armadillo (11% each), and agouti (9.4%). However, if ranked according to the estimated biomass consumed (mean weight of prey x number of individuals per scat x number of occurrences), small rodents are preceded by another eight prey items (Figure 428). The fox was known to have been taken as carrion, when F23 was seen feeding on a roadkill (see Dispersal). It is possible (and likely) that animals such as adult Procyon and Mazama and other large species are also consumed as carrion, since they are sometimes killed by cars in the park, and may at times become important alternative food sources. However, since these large prey were rarely found in ocelot scats, their overall contribution is likely to be negligible. In Figure 4-29, I considered only the most frequently taken taxa (> twice in my small sample). The prominence of armadillos, opossum, agoutis, and lizard in ocelot diet are in agreement with data on relative abundance, based on sightings, of these species in the Park (see Appendix B). All these species are








41

< 1/3 of the mean ocelot body weight, as compared to the other large items that appeared only once in the sample.












Table 4-1--Monitoring period for 21 ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina. Letter with the animal number indicate the sex; C= capture; R= recapture (a number indicates number of recaptures during that month); K= animal killed; F= transmitter failure; D= animal dropped collar.
1990 1991 1992 1993 1994
JFMAMJJASONDJFMAMJJASONDJFMANJJASONDJFMAMJJASONDJFMAMJJABOND F02 C2--F...R--F
F03 CFR---F
F06 C ---------------F ................. Rab -------------->
F09 C ------F16 C---K F19 C ----------------F .......... Rb ---------- >
F23 CF.....Ra'b .... K
F39 C ---- K F40 C ---------->
F42 C --------- >
1990 1991 1992 1993 1994
JFMAMJJAONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND M04 C --------- -D
M05 C3K
320 C ------------------- F
3125 CF
3126 C ----------------F
M27 CR-3R ---- F ...... R --------------->
M35 C --------------- >
3136 CK M37 C ---------M41 C---R ---- >
!M45 C .------->
a_ Animal taken into captivity for treatment b Collar changed











Table 4-2--Home range estimates for 21 radio-collared ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina. N= number of locations; MC= Minimum Convex Polygon method; HM= Harmonic Mean method; ELL- 95% Ellipse method.

AN # N MC HM ELL F02 59 15.6 15.6 27.8 F03 33 5.2 15.8 12.4 F06 91 40.4 136.0 44.0 F09 36 7.6 37.0 20.2 F16 10 12.1 25.7 56.2 F19 110 30.0 102.5 29.6 F23 19 27.8 127.3 112.7 F39 23 13.1 78.0 31.0 F40 40 3.4 15.2 6.9 F42 21 5.1 40.3 13.7 M04 60 20.9 59.2 47.5 M05 27 8.3 18.2 15.3 M20 38 41.5 99.7 75.2 M26 57 46.2 119.5 92.5 M27 78 50.9 119.2 87.0 M35 34 40.4 134.8 60.7 M37 32 73.1 - 94.1 M41 40 121.1 441.9 189.2 M45 26 27.7 67.4 52.6 Note: Although area estimates are given for subadult animals, they cannot be considered actual home ranges.











Table 4arithmetic and Iguaz using the 2 SD (95% AN #/AGE F02 (SA) F03 (AD) F06 (AD) F09 (AD) F16 (AD) F19 (AD) F23 (SA) F39 (AD) F40 (AD) F42 (SA) M04 (AD) M05 (AD) M20 (AD) M26 (AD) M27 (AD) M35 (AD) M37 (SA) M41 (SA) M45 (AD)


3--Mean distance (km) of locations to the center (AC) for radio-tagged ocelots in Iguaqu a National Parks, and home ranges estimates (km2) mean distance to AC (50% HR) and mean distance + HR; see Methods). No outliers were excluded.


N 61 33 91 36 10 110 25 23

42 32 53 27 63 58 83

42 43 42 26


MEAN + SD 0.9 + 1.1 1.2 + 0.5 1.8 + 1.1 1.7 + 0.6 1.8 + 0.7 1.5 + 0.9 4.2 + 5.8 1.6 + 0.8 0.7 + 0.3 1.4 + 0.9 2.6 + 1.1 1.3 + 0.7 3.2 + 1.5 2.8 + 1.6 2.8 + 1.4 2.5 + 1.5 3.0 + 3.7 3.5 + 3.5 1.9 + 1.0


RANGE

0.1 - 7.1 0.4 - 2.0 0.1 - 5.0 0.5 - 3.4 0.9 - 3.2 0.1 - 4.2

0.8 - 32.1 0.7 - 4.3 0.1 - 1.8 0.3 - 3.9 0.5 - 4.9 0.4 - 3.6 0.1 - 6.6 0.5 - 6.2 0.3 - 5.9 0.5 - 5.8

0.3 - 15.3 0.3 - 19.1 0.5 - 4.3


50% HR
2.5

4.5

10.2 9.1 10.2 7.1

55.4 8.0 1.5 6.1

21.2

5.3

32.1

24.6 24.6 19.6 28.3 38.5 11.3


95% HR
33.3

17.4 54.2 29.3

35.4 37.5 800.1

35.3

6.6

35.4 77.2

25.6 126.9 119.1

104.1 100.5

350.2 356.8 51.7


Note: Although area estimates of subadult animals are included, they cannot be considered as actual home ranges.












Table 4-4-- Distance (km) between the arithmetic centers
(AC) of home ranges of adult ocelots in Iguagu National Park, Brazil. N05 was not included in other comparisons because he was killed (by M04?, see text) before other animals were radio-collared. MALES M05 M26 M27 M35 M45 M04 4.4 2.0 1.4 6.9 4.2 M26 2.6 8.5 5.4 M27 5.9 3.0 M35 3.5


FEMALES F03 F06 F09 F39 MALES/ FEMALES
M04 M05 M26 M27
M35 M45


F06
4.5





F03

2.8 0.1 2.7 1.7 6.7 3.2


F09 6.2 3.6



F06

5.5
4.6 6.8
4.3 2.5
1.4


F39 1.1
4.9 7.0


F09

5.4 6.3 7.3
4.9 2.4 3.7


F40 6.8 3.0 1.8 7.5 F39

3.8 1.0 3.3 2.8 7.2 3.7


F40

6.7 7.0 8.5 5.9 0.8 3.8








46


Table 4-5--Mean linear distances moved between locations with a 1-day interval for 20 ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina. N= number of locations.

AGE/SEX N MEAN � SD RANGE Adult males (n= 6) 34 2.1 � 1.7 0.0 - 7.3 Adult females (n= 7) 30 1.3 � 1.1 0.0 - 4.0 Subadult males (n= 4) 15 1.9 � 1.3 0.3 - 5.8 Subadult females (n= 3) 20 1.1 � 1.0 0.0 - 5.0 All animals (n= 20) 99 1.6 � 1.4 0.0 - 7.3











Table 4-6--Daily survival and mortality rates of radiotagged ocelots in Igua9u and Iguazu National Parks, Brazil and Argentina. N1= number of animals monitored; # LOC= number of locations in the period; Di= number of animals that died; LC= transmitter failure; si= daily survival rate; m1= daily mortality rate. Adapted from Trent and Rongstad (1974) and Heisey and Fuller (1985).

YEAR # MONTHS N, # DI LC si mi LOC

1990 5 6 120 1 0 0.9916 0.0084

1991 12 8 189 1 4 0.9947 0.0053

1992 12 7 156 0 1 1.0000 0.0000

1993 12 8 79 2 2 0.9746 0.0254

1994 12 10 282 1 0 0.9964 0.0036


9 78 0 0 1.0000 0.0000


1995 5












Table 4-7--Food items found in 56 ocelot scats in Iguaqu National Park, Brazil. WGT= mean live weight (kg) of prey item; N= number of scats the item was found in; #INDs number of individuals per scat; ITEMS= percentage of total number of items found; % SCATS= percentage of scats containing that food item; ; Bet= estimated biomass of that prey item.


PREY ItEM small rodents Opossum
(Didelphis aurita)
Armadillo (Dasypus novemcinctus)
Agouti
(Dasyprocta azarae)
Invertebrates (insects= 5; land snail= 1) Rabbit
(Sylvilagus brasiliensis)
small marsupial Bird
(Cracidae, Tinamidae) Snake
Lizard
(Tupinambis teguixin)
unidentified fruit Cavy
(Cavia aperea)
Porcupine (Coendou prehensilis)
Squirrel (Sciurus aestuans)
Deer
(Mazama sp.) Fox
(Dusievon thous)b Raccoon (Procyon cancrivorus)
Grison
(Galictis cuja)
TOTAL


WGT N #IND 0.05 31 2


1.50 3.30

3.20


0.78 0.01 0.65 1.00 1.60



0.40 3.50 0.80

22.5a

6.0 10.0

1.1


a Average of the weights of the two species b Taken as carrion (see text); I Does not add to 100% because of more than 1


ITEMS 41.3


SCATS 55.3


Best 3.1


7 1 9.3 12.5 10.5 7 1 9.3 12.5 23.1 6 1 8.0 10.7 19.2


3 1 4.0


3 1 4.0 2 1 2.7 2 - 2.7 1 1 1.3 1 1 1.3 1 1 1.3

1 1 1.3 1 1 1.3 1 1 1.3 1 1 1.3


5.3 2.34 5.3 0.03 5.3 1.95 5.3 3.0
3.6 3.2 3.6
1.8 0.40 1.8 3.50 1.8 0.80 1.8 22.5 1.8 6.0 1.8 10.0 1.8 1.1


99.7 133.8c
M. americana and M. nana; item per scat.


I






















































Figure 4-1--Radio-locations (.) and capture locations (m) of an adult male ocelot (M04), Iguaqu National Park, Brazil. Monitoring period: 30/08/90 - 14/10/91.














- - t. - - - - -
e - - - - - - -


N ort


lgua~u National
Park


* * * *

* I
0 I
S
- - - -- - -- ~.* -.


Iguazu National
Park


Argentina


1.5 km


Figure 4-2--Radio-locations (.) and capture locations (a) of an adult male ocelot (M05), Iguaqu National Park, Brazil. Monitoring period: 01/09/90 - 03/11/90.


Brazil


44 b.

0*


(~. 4,..


Entrance Gate e**...


0


~













- - - - - - - - 'a ,


IguaVu National
Park


Brazil


rport


Entrance


0
0 0
0 0


Iguazu National
Park


Argentina


1.5 km


Figure 4-3--Radio-locations (.) and capture locations (m) of an adult male ocelot (M26), Iguaqu National Park, Brazil. Monitoring period: 23/05/92 - 06/10/93.


/


* *
I
*
**. ***
I
**~ -I -
I-






















































Figure 4-4--Radio-locations (.) and capture locations (U) of an adult male ocelot (M27), Igua4u National Park, Brazil. Monitoring period: 27/05/92 - present.













. ' -.%P - - I -


* lguaiu National
Park


Brazil


-.



t.. �
� 0


> Airport


0
0
0 -~ 0 I ~..0
U
4.
/


I
/
I
I
I
I


* a
**. **
*. .*


. '00


Iguazu National
Park


Argentina


1.5 km


Figure 4-5--Radio-locations (.) and capture locations (0) of an adult male ocelot (M35), Iguaqu National Park, Brazil. Monitoring period: 21/09/93 - present.






















































Figure 4-6--Radio-locations (-) and capture locations (M) of an adult male ocelot (M45), Iguaru National Park, Brazil. Monitoring period: 11/05/94 - present.






















































Figure 4-7--Radio-locations (.) and capture locations (0) of an adult female ocelot (F03), Iguaqu National Park, Brazil. Monitoring period: 23/08/90 - 02/02/91.


















- lguavu National

Park


Brazil


* 0


0 0


Y Airport


Eutrance Gate * . . **.


4.


*

*
* 0 0
(-.
0
0 4
*0
* 0%
* @0 0
*0

* 0 0
0 0
*0
0
0
0


* 0 0


* 0
* U
/ 0
* 1,

* 0/0 0


0
I
U
0I


Iguazu National

Park


Argentina


1.5 km


Figure 4-8--Radio-locations (.) and capture locations (n) of an adult female ocelot (F06), Iguaqu National Park, Brazil. Monitoring period: 15/09/90 - present.


--- - o - %0


9















- - -
- % --


Iguaqu National
Park


Brazil


IQ. ..


0 S


Y Airport


* / * I * I * I * 1 * U
**. ** I
* .*
- - - - - - - -- -- -


0
*0


Iguazu National
Park


Argentina


1.5 km


Figure 4-9--Radio-locations (.) and capture locations (m) of an adult female ocelot (F09), Iguaqu National Park, Brazil. Monitoring period: 18/10/90 - 02/07/91.





















































Figure 4-10--Radio-locations (.) and capture locations (0) of an adult female ocelot (F39), Iguaqu National Park, Brazil. Monitoring period: 19/02/94 - 15/08/94.























































Figure 4-11--Radio-locations (.) and capture locations (m) of an adult female ocelot (F40), Iguaqu National Park, Brazil. Monitoring period: 18/03/94 - present.

















I - . - - - % o


lguaVu National

Park


Brazil


\ nAirport





Entrance


0 0
0 0


.0 0
*% 00
0 000 0 0


* 0 0 0


0*













* j * I
* .* I
*S. ** U
I
- -- S..
5
- - - - - - - -


Iguazu National

Park


Argentina


1.5 km


0 0


Figure 4-12--Radio-locations (.) and capture locations (0) of an adult female ocelot (F19), Iguazu National Park, Argentina. Monitoring period: 03/08/91 - present.


* 0� .0 ** * 0.










% % d o - - -


iguaqu National " Park


BIrazil


b'.
* .


oAir rt


do
" .. .Cee6
- - as -. -m -oo~e� --e --


0
0


Iguazu National
Park


Argentina


1.5 km


Figure 4-13--Radio-locations (.) and capture locations (U) of a subadult female ocelot (F42), Iguazu National Park, Argentina. Monitoring period: 05/04/94 - present.





















































Figure 4-14--Radio-locations (.) and capture locations (N) of a subadult male ocelot (M37), Iguaqu National Park, Brazil. Monitoring period: 13/02/94 - present.















- - - % - - - -
I - - - ~ - - - -


Iguaqu National
Park


Brazil


tAirpora ~4< Entrance


0
b.

0*






0.





* 0

*.. .*
*. .*
0


0* 0 *


\* 6
0


0
, %...
/
00


/
I
I
I
I
I


.0 *


Iguazu National
Park


Argentina


1.5 km


Figure 4-15--Radio-locations (.) and capture locations (U) of a subadult male ocelot (M41), Iguaqu National Park, Brazil. Monitoring period: 04/04/94 - present.











Area (Sq. Km)


0 10 20 30 40 50 60 70 80 90 100 110 Number of Locations

-- F03 - F06 - F9 -- F16
101 F19 F F39 A. F40



Figure 4-16--Cumulative area curve for home ranges (Minimum Convex Polygon) of 7 adult female ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina.












60

50 40


Number of Locations


M04 -- M05 - M20 - M26
04 M27 - M35 -a- M45



Figure 4-17--Cumulative area curve for home ranges (Minimum Convex Polygon) for 7 adult male ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina.












AREA (km2)


140 120 100

80 60

40

20

0


F06


U I


F09


F39


F40


ANIMAL #


I = MCP fIHM M-95%ELL ECircle I


Figure 4-18--Comparison of home range estimates (km2) adult female ocelots in Iguagu and Iguazu National using different estimation methods. MCP= Minimum Polygon; HM= harmonic Mean; 95% ELL= 95% Ellipse.


for 7 Parks, Convex


II


F03


u


II


I












250 200 150 100

50

0


AREA (km2)


M04 M05 M20 M26 M27 M35 M45 ANIMAL #

MCP E HM M 95% ELL IE Circle


Figure 4-19--Comparison between home range estimates for 7 male ocelots from Iguaqu and Iguazu National Parks, Brazil and Argentina, using different estimation methods.








68



Km
5




3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


2 . . . . . ............ . . . . . . . . . . . . . . . . . . . . *.........**............... * * * ,- *' ' ' ** * * * * " * ** ...... .......




S ........... I........................ I ........................ t ......................... ...........


01
ADM (N- 7) ADF (N- 7) SAM (N- 2) SAF (N- 3) AGE/SEX


Figure 4-20--Distance (mean � 1 SD; km) from each ocelot's locations to the arithmetic center of their home ranges (see
methods). ADM: adult males; ADF: adult females; SAM: subadult males; SAF: subadult females.












PERCENT OF LOCATIONS


ADM ADF SAM SAF
AGE/SEX

Distance (Km)
(1 1/2 9M2/3 3/4 FEE )


Figure 4-21--Percent of locations at various distances from the arithmetic center of different age/sex classes of 21 ocelots in iguaqu and Iguazu National Parks, Brazil and Argentina.












PERCENT OF LOCATIONS


ADM ADF SAM SAF
AGE/SEX

Distance (Kn)
QI 1/2 ME] 2/3 3/4 )!5


Figure 4-22--Distances (km) between consecutive locations for different age/sex classes of ocelots in Iguaou and Iguazu National Parks, Brazil and Argentina.













100


80 60


PERCENT ACTIVE


0 1 1 1 1 1 1 I 1 , , ,- ,,4 . I . I --'l 0600 0800 1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 TIME OF DAY


Figure 4-23--Hourly percent activity for 1,115 readings on 21 ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina.


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


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


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












NUMBER OF OCCURRENCES


J F M A M J MONT]

PREGNANT (N" 4)


i A S 0 N D
B
=BIRTH IN- 8)


Figure 4-24--Temporal pattern of reproductive events of female ocelots in Igua4u and Iguazu National Parks, Brazil and Argentina.


I i e , , l I I












Percent


100


80


60 40 20


0


I -- Cumulative Mortality


MAnimals monitored


Figure 4-25--Cumulative mortality of radio-tagged ocelots in Iguaqu and Iguazu National Parks, Brazil and Argentina. Percent mortality and number of animals remained stable after 660 days of monitoring for the remainder of the study.


0 60 120 180 240 300 360 420 480 540 600 660 Days

































Mamn


I F, 0; I'll








4
rm
40'ea, w 4,6-4, a VoW a ..

I e 0 n.- r e V "IC 0. '. ft eAn":-";. !A 4'. b a 0 a '4 b 4.,# 0:4 0 4.:, 4:4 0: 0 q.: 0 AT. P, 0 0 0 4-:-b MA'.0 4 41:. C 0 A. 6 4 404'. 4
ik'r' pe 4'W*w'b4'.b' a 41.4T w V 6.4'. b 4 v a A 1.: 1. W., A W., P *.11. W.1 A 9 WA, 1. v v V. M F. 9 0 I'M 9.11. W., p
C A n V 0, 4 7 W'&ne b .A n ne A"o
.0 A 'p. 0A 0 40. 01, 0.41 04.1,4 W4.
4.1 47.61 1-4-T.1 --b IV. ._-I. Ord b
W.'.P'N "4o

nA 9.11. v A
4:4 W 4 p Q:.0:6 4.: 0 4: t 0 0 4: 6 4:' S 4 T 0 4 T. 0" 0 TO b 4:.p a :4.: a 4.:. w :"Ts', .:.I's'q
A'4V6 4 W &-IV& Z



vs, 74.
vv

A, Ip
l.AYOC J.A
-VOpj':* 41 "p 4-:4 4I*'; q;%,= ;-T.'IF" '; -----------MMOMMOSE.8m


V641,bllw '4 b4 V41 rv
ne 140 wmw 1.0:0 PA IMNA "MMMM,

V T41- A

0 was
w*_.k:A A k 0., P. 0 A, F 4 9 A. a
w4no
a A t 0
4wa A"OAwd v





n w a V w2fin w v


rtebrates


Figure 4-26--Percentage of the different taxa found in 56 ocelot scats in Igua9u National Park, Brazil.


Fruit Birds

















Small zodents


Porcupine Squirrel Cavy Deer


Otbet Carnivores


Agouti


Figure 4-27--Frequency of occurrence (%) of mammal species found in 56 ocelot scats, Iguaqu National Park, Brazil. Other carnivores include the crab-eating-fox, crab-eating raccoon, and grison (one occurrence each= 1.6%).


Opossum













%Estimated Biomass (kg)


201


154


Atmadillo Deer


Agoutl Opossum Raccoon Reptiles Fox Pozcupin. Small zodents


Figure 4-28--Prey items found in 56 ocelot scats in Iguaqu National Park, ranked by estimated biomass represented in scats (%). Snake and lizard were combined as reptiles.













% Estimated Biomass (kgl
AA v


Armadillo


I///E


Agouti Opossum Reptiles Rabbit


Small
rodents


Birds


Figure 4-29--Percent estimated biomass (kg) of taxa found > twice in a sample of 56 ocelot scats in Iguaqu National Park, Brazil.














CHAPTER 5
RESULTS: JAGUAR





Radio-Telemetry

Seven jaguars were captured and radio-collared, including two adult males, one adult female, three subadult males, and one subadult female. One of the adult males (M48) was captured in Iguazu National Park, Argentina; all other animals were captured in Iguaqu, Brazil. Another adult male was captured on the edge of Iguazu and translocated as a problemanimal to a provincial park south of the Argentine park. This animal was not radio-collared. Appendix C summarizes information on trapping results and morphometrics of individuals.

Five of the jaguars in Iguaqu were probably related. F17, the adult female, was presumed to be the mother of M13 and F21 (littermates) and M32 (from a subsequent litter). As the resident adult, M33 was probably the father of both litters.

A total of 236 locations was obtained on the study animals, 142 (60%) for males, and 94 (40%) for females. Adult and subadult animals were monitored about equally, with 49%








79

and 51% of locations, respectively. The cumulative monitoring period was 2,292 days, 1,764 (77%) for males and 528 (23%) for females. Overall, the mean interval between locations was 10.1 � 15.2 days (range: 0 - 102 days). Eighty percent of the locations were aerial. Tables 5-1 and 5-2 provide information on the monitoring of the study animals and Figures 5-1 through 5-6 show the individual maps with plotted locations.



Home Ranges and Movements

Home range estimates (Minimum Convex Polygon method) varied considerably for the study animals (Table 5-3), ranging from 8.8 km2 (F21) to 138 km2 (M13). Undoubtedly, some of this variation can be accounted for by differences between sex and age classes. However, the cumulative area curves (Odum and Kunzler, 1955; Figures 5-7, 5-8), show the ranges of most individuals are underestimated, due to small sample sizes and to reasons discussed below. The only exception is F17, for which an asymptote was reached at 50 locations. This female utilized a total area of 70.0 km2 during 14.5 months of monitoring.

In addition to the short monitoring period, the movements of most of the radio-tagged jaguars were influenced by human activities. Adult male M48 was translocated 8 km from his first capture site, as a problem-animal. His movements were further affected by a shotgun wound, when he resumed predation on livestock (see Mortality and Relationships with Humans).








80

The movements of M33, a male well past his prime, were likely influenced by an acquired dependency on the live baits in my traps (8 recaptures; see Appendix C). This may have been due to increased difficulty in his hunting ability, since his condition deteriorated steadily between captures. Similarly, the movements of F21, a subadult female, were influenced by a dependency on human-originated food, at the garbage dump at the hotel by the falls, and then by a short translocation following capture (see Mortality). Captured as subadults, M13 and M32 were in the process of establishing their adult home ranges when they were killed by poachers, in very similar situations (see Mortality). The signal of subadult M34's transmitter disappeared after only 21 locations, while he was still using his natal area.

As with the ocelots (see previous chapter), the dense vegetation and lack of an adequate trail system precluded routine ground monitoring of radio-tagged animals. Due to the small sample sizes, no statistics could be appropriately applied to the data.

Six of the study animals were located with a 1-day interval on 24 occasions. For all animals, the mean linear distance moved on these locations was 1.3 � 1.2 km (range: 0.2

- 5. 4 km). Overall, males tended to be found at greater distances than females on consecutive locations at increasing intervals (Figure 5-9). Table 5-2 shows the mean linear








81

distances between consecutive locations for each of the radiotagged jaguars.

Likewise, males were found more often at greater distances from the arithmetic center (AC) of their ranges than females (Figure 5-10). The distribution of locations in relation to the AC shows a tendency of greater use of the periphery of the home ranges for adult males. The adult female, on the other hand, displayed a more regular use pattern of her area (Figure 5-10).



Reproduction

On 26 September 1990, a cub was seen at Km 4.5 of the Pogo Preto road (PPR). His track width was 7 cm, and his weight was estimated at 30 kg (see Appendix C). On 19 February 1991, an adult female with two large cubs were sighted on Km 3.7 of this same road. On 19 April, M13 was captured at Km 2.5, at an estimated age between 12/14 months. On 31 July, F17 was captured at the same site, her age estimated at about 8/9 years old. On 12 September, F21 was captured behind the hotel near the falls, and her age was about the same as that estimated for M13. The distribution of their locations strongly suggests that they formed a family unit, very likely the same animals seen in February. Assuming the age estimates for the subadults are correct, their birth would have occurred between February and April of 1990, and








82

mating (- 110 days of the gestation period) from late October to December.

When captured, F17 was in about mid-term pregnancy. On 7 October 1991, 68 days later, the mortality mode of her transmitter was activated but soon resumed to the active mode. As mentioned in the previous chapter, similar instances occurred with bobcat and ocelot females prior to giving birth. The mean distance between consecutive locations for this female increased from 2.3 � 1.1 km to an asymptote of about 3.5 � 2.5 kim, between October 1991 and October 1992 (Figure 511), likely a reflection of mobility of young.

On 17 November 1992, M32 was captured at the same site as M13 and F17, with an estimated age of about 12 months. Further indication for the assumption that he was F17's son is provided by the fact that when this female was recaptured on 9 October (39 days earlier), she showed signs of a light mange, which was also present on M32 (no other animals were found with this disease). [Note: Due to a history of aggressiveness and a tendency to take domestic animals, in her old age, combined with bad physical condition when recaptured (both her upper canines had broken at the base and were badly infected), this female was removed to a zoo]. M32 was also very lean, and was locally medicated for anemia and dehydration, before release at the capture site.

M34 was captured on 14 August 1993, also at Km 2.5 of the Po9o Preto road, at an estimated age of 10-12 months (he still








83

had his milk dentition). Therefore, his birth must have occurred between August and October 1992. This fact poses an interesting question since F17, then the established female, was still accompanied by M32. Thus, she could not have been the mother of M34. And, since she was removed from the Park on 9 October, the mother of M34 must have been established in the same area occupied by F17, which leads to further the assumption that she may be F17's daughter. Tracks of an adult female have been regularly found along the PPR, and one uncollared female was darted free-ranging, from a blind, on the night of 26 July 1993, when raiding a sheep corral just outside the Park. However, the drug was not injected and the animal escaped. Intensive efforts to trap this female have so far been fruitless. Judging by tracks found on the PPR during early 1995, there is another subadult animal using the area, likely a subsequent offspring of this same female.



Dispersal

Two subadult males, M13 and M32, were monitored during dispersal. After a few long-range movements (> 5 km) outside of his natal area, M13 crossed the Iguaqu river into Iguazu National Park, in Argentina, on 24 September 1991. Assuming his birth was approximately in March 1990 (see section on Reproduction), his age would then be about 18 months. He remained in Argentina until 9 November, during which time he was located 5 times, with a mean linear distance between








84

locations (MLDBL) of 2.4 � 2.7 km. On 10 November, he was back in Iguaqu, where he remained until 13 November, with a MLDBL of 2.0 � 1.4 (0.6 - 3.8; N= 4) km. On 21 November, he was found again in Argentina, where he traversed the whole peninsula, and on 1 December, was found in Paraguay, having crossed the Parand river at a point where it is > 400 m wide. He was confined to a small island of forest, surrounded by farmland. On 17 December, he was back once again in his natal area in Iguaqu, a linear distance of 16.2 km from his previous location. On 18 January 1992, he was located 33 km northeast of where he was located on 9 January. When he was recaptured on 13 March, to change his collar, he weighed 88 kg. He had several open cuts, which he likely incurred from intraspecific fights, and his upper right canine was broken off almost at the base. Between 18 January and 16 May, the mean distance between 13 locations was 12.9 � 11.1 km (range: 1.0 - 33.0 km). On 26 May, the mortality signal was received and I located the collar at a house by the Iguaqu river, on the opposite side from the Park. After landing the helicopter at the house, I discovered M13 had been killed by the owner and his 15 year-old son while poaching white-lipped peccaries (Tayassu pecari) in the Park.

The cumulative mean distance (� 1 SD; km) of locations to the arithmetic center of M13's natal area increased steadily and seemed to be levelling off when he was killed (Figure 512). The coefficient of variation (CV; Ott, 1984; see








85

methods), plotted on the Y2 axis of the figure, indicates the random nature of his movements.

During his first 9 locations following capture on 17 November 1992, M32 covered an area of 5.4 k2, presumably still in his natal area. After 21 January 1993, the signal of his transmitter could not be found within the study area. By gradually increasing the area searched during flights, I found him on 22 February on the eastern sector of the Park, 64 km east of his last location. Assuming he was born in September 1991 (see Reproduction), his age then was 16 months. His dispersal may have been precipitated by the removal of his mother, F17, from the Park in October. His next 10 locations encompassed an area of 308 km2, with a mean linear distance between locations (MLDBL) of 9.0 � 3.6 km (range: 4.4 - 16.7 km). His movements for the next 10 locations decreased to a mean of 3.4 � 1.2 km (range: 1.8 - 4.9 km), and were confined to an area of 17 km. In his next 10 locations, he increased the area used, as a result of two long range movements (17.5 and 10.2 km) to an area where he was later seen twice (during recapture attempts) with an adult female. Therefore, he was already becoming an established adult, as indicated by the asymptote reached in the mean distance from the AC of his natal area and on the CV of his locations (Figure 5-13). He was killed by poachers in June 1994.

As a comparison to the dispersal patterns shown in Figures 5-6 and 5-7 for the Iguaqu males, I applied the same








86

method to a dispersing subadult male from another study (Figure 5-14; data from Crawshaw and Quigley, 1991). The greater number of locations for this male shows the expected pattern of increasing mean distances and SD from the AC of the natal area until an asymptote is reached when the animal becomes established in an area. Conversely, the CV shows an initial increase in the dispersion of locations as the animal disperses, then gradually decreasing as it becomes attached to a specific area.

Subadult male M34 would have provided another example of dispersal, but I lost contact with his transmitter while he was still in his natal area (see Mortality).


Mortality

Three of the 7 radio-collared jaguars were killed by poachers within the limits of Iguaqu National Park. Two males (M13 and M32, brothers of subsequent litters of F17) were shot by hunters waiting at blinds for other game species (peccaries, Tavassu Decari and _T. taJacu, and paca, Agouti paca). The female, F21, was shot apparently as a trophy in a contracted hunt in the most densely visited part of the Park. Being habituated to humans, she regularly used a rustic shed of a concessionaire as an off-hours refuge, and was killed a short distance from it.

An uncollared adult male (Appendix C, Table C-2) was blinded by a small-grain cartridge shot on September 1993, and








87

was killed by a truck on the highway that borders the northeastern section of the Park. A large jaguar cub (circa 20 kg) was killed by a vehicle on this same stretch of highway on 30 March 1995.

In addition, an uncollared large male (Appendix C, Table C-2), captured as a problem-animal in a private property neighboring Iguazu NP and translocated to Urugua-i Provincial Park, to the south of Iguazu, was shot by locals within one week of his release.

The transmitter signals of M33 and M34 disappeared prematurely, and they were not found, despite extensive aerial searches. It is possible that M34, a subadult male, dispersed out of our searching range. However, this is unlikely since very little undisturbed habitat remains outside the Park on the Brazilian side. All these areas and those in Argentina were thoroughly searched from the air. M33 was an old, established animal, and his transmitter had still another full year of battery life at the time he disappeared. The low probability of Telonics transmitters failure (see Material and Methods) and the history of high mortality of jaguar in the area suggests that these animals were likely killed and their collars destroyed.

The overall daily survival (si) and mortality rates (mi) for the radio-tagged jaguars were 0.9788 and 0.0212, respectively (see Mortality section in the previous chapter, for explanation of method). If we assume that both animals




Full Text

PAGE 1

COMPARATIVE ECOLOGY OF OCELOT (Felis pardalis ) AND JAGUAR f Panthera onca) IN A PROTECTED SUBTROPICAL FOREST IN BRAZIL AND ARGENTINA By PETER GRANS DEN CRAWSHAW, JUNIOR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1995 UNIVERSITY OF FLORIDA LIBRARIES

PAGE 2

Copyright 1995 by Peter Gransden Crawshaw, Junior

PAGE 3

To Mel, advisor and friend Fiona, and Mara Without whom I never would 've made it in time.

PAGE 4

ACKNOWLEDGEMENTS This study was generously supported by the Institute Brasileiro do Meio Ambiente e dos Recursos Naturals Renovaveis — IBAMA, the World Wildlife Fund WWF/US, Helisul Taxi Aereo Ltda., the Lincoln Park Zoo "Scott Neotropical Fund", Funda
PAGE 5

extremely thankful also to Drs. John F. Eisenberg, Steve R. Humphrey, John G. Robinson, Peter Feinsinger, Kent H. Redford, Jack Kaufmann, George Tanner, and David Webb. I was very fortunate to have had them on my committee, at one time or another, to help and guide me through the burdens of academia, and to relish its pleasures. Kent Redford was a constant driving force for my graduate studies at UF, since our first campfire conversations in the heart of the Pantanal, back in 1978. Even though it meant an overall delay of two years in my graduate course, I was fortunate to work in another PhD project under Dr. Steve Humphrey. For personal and professional reasons, I had to switch back to my original Igua^u proposal (and re-start from scratch) . However, I will always cherish my memories of bobcats in the Big Cypress National Preserve in south Florida, and of the support and friendship from Steve and the boys at the Florida Fresh Water Fish and Game Commission in Naples. Dave Maehr was instrumental in helping me set up the study there. The study in Iguaq:u wouldn't have been possible without the unconditional help of the Park's director, Dr. Jose Carlos Ramos. He always made sure I had what I needed when I needed it. Terezinha S. Martinez was my buffer between rigid financial rules of the federal government in the Brasilia IBAMA offices and the flexibility required by field demands. Also in the Park's staff, Apolonio N. Rodrigues and Salete F. V

PAGE 6

da Costa provided invaluable help and companionship. I also thank all the soldiers of the Batalhao de Policia Florestal do Parque Nacional do Iguagu that contributed with the project, especially Cabo Concei^ao and Soldado Barreto, and the Lieutenents Almeida and Wellington. My debt to my field assistants, Sandra M. C. Cavalcanti, Cibele B. Indrusiak, Jan K. Mahler, Lucila Manzatti, Ronaldo G. Morato, Rose L. Caspar ini, Ricardo Boulhosa, and Hannes Riidiser, is immense. Their presence and reliability allowed me to fulfill my responsibilities with an overload of field and office work, as well as attending several other projects. I wish them all a brilliant professional future! The binational character of the project was only made possible by the commitment and enthusiasm of Silvana B. Montanelli, the biologist in Iguazu National Park, Argentina. I hope we can continue to work together for many years to come. Several other people were pivotal to the study in Iguazu: the two Park's directors, Ugo Rossi and, later, Horacio Giachino, Karina Schiaffino, Laura Malmierca, Carlos Saibene, Pedro Moreyra, Jose L. Comita, and Sergio Acosta. I also benefited from the association with the workers in the "Primate Social Structure" study undertaken in Iguazu, by Drs. Charles Janson and Alejandro Brown. I also thank Juan Carlos Chebez, head of the Wildlife Division for Northeastern Argentina, Administracion de Parques Nacionales, for sharing vi

PAGE 7

with me his vast knowledge on the Argentine fauna, especially regarding early studies. In the stream of students and professionals that passed through Iguagu, I am thankful especially to Junio A. dos Santos, Leandro Silveira, Renata P. Leite, Marcelo Mazzoli, Murilo Gubert, Andre Levy, and Jim Sparks for their efforts in data collection. Renata P. Leite conducted all the work in chromatography (however frustrated) and identified predator scats for me. Plants were identified by Marcos Sobral (Faculdade de Farmacia, Universidade Federal do Rio Grande do Sul) , Pedro Luis (Master's program, Universidade Federal de Minas Gerais) , Suely (Institute de Botanica de SP) , Sandro M. Silva (Universidade Federal do Parana) , and Stinger Guala (University of Florida) . Ives Sbalqueiro (Universidade Federal do Parana) and L. Flamarion B. Oliveira (Museu Nacional do Rio de Janeiro) identified small rodents for me. The project gained significantly from the participation of Wanderley de Moraes, the veterinary from Itaipu Binational. I'd like to believe that all the study animals he treated, when necessary, share my appreciation of him. Also from Itaipu, I'd like to thank Leonilda, Emerson Suemitsu, Helio Pontes, and Fernao Carbonar for their support at one time or another. Mel and Fiona Sunquist provided "a home away from home" during the concluding stages of the writing. Their vii

PAGE 8

accumulated expertise and vast bibliography always at hand made all the difference. Their warm hospitality and support will never be forgotten! The friendship, help, and advice from Rosa Lemos de Sa, Bob Godshalk, Vicky McGrath, Luciano Verdade, Jay Malcolm, and Justina Ray are deeply appreciated. Vicky McGrath also prepared all the maps of Iguagu for me. Finally, my gratitude to my wife Mara and to my "children" Danielle, Beatriz, and David cannot be adequately expressed in words. Not many people would agree to the system of priorities that I have to (or chose to) live by. I just hope I somehow can make it up for all the times when I should have been there for them, and wasn't... viii

PAGE 9

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARATIVE ECOLOGY OF OCELOT (Felis Eardalis) AND JAGUAR (Panthera onca) IN A PROTECTED SUBTROPICAL FOREST IN BRAZIL AND ARGENTINA By PETER GRANSDEN CRAWSHAW, JUNIOR DECEMBER 1995 Chairperson: Dr. Melvin E. Sunquist Major Department: Department of Wildlife Ecology and Conservation (Forest Resources and Conservation) Twenty-one ocelots (11 males, 10 females) and seven jaguars (5 males, 2 females) were captured and equipped with radio-transmitters in the neighboring Igua?u (1,750 km^) and Iguazu (550 km^) National Parks, southwestern Brazil and northeastern Argentina. Mean home range size for 6 adult male ocelots was 38.8 km^ and for 5 adult females it was 17.4 km^. Males travelled farther than females between consecutive locations (2.1 ± 1.7 km and 1.3 ± 1.1 km, respectively). Although more active at night, ocelots in Igua?u displayed more daytime activity than shown in other studies. This may reflect a greater reliance on more diurnal prey, such as agoutis f Dasvprocta azarae) and tegu lizards ( Tupinambis teguixin) . xi

PAGE 10

Dispersal of two sxibadult male ocelots and one subadult female suggested difficulty in establishing new ranges within the resident population. The process involved long-range, apparently erratic movements, and in one instance, conspecific fights. This may have been due to a relatively high density (12 adults/100 km^) and a low turn-over rate. Mean home range size for 4 adult male jaguars was 88.7 km^; one adult female jaguar had a home range of 70 km^. The mean linear distance between consecutive locations was greater for males than for females. Two dispersing subadult males were able to establish their adult home ranges within Iguagu, about 60-70 km away from their natal areas. This apparent ease in finding space was likely correlated with a lower jaguar density (3.7 adults/100 km^) and high turnover rate. Poaching accounted for most of the mortality of both species, affecting jaguars more heavily than ocelots. Genetic interchange occurred between populations from the two Parks, both species crossing the Iguaqru river during dispersal. Mean prey-predator weight ratios for ocelot and jaguar in Iguagu were 0.13 (mean prey weight, MPW= 1.4 kg) and 0.18 (MPW= 14.4 kg), respectively. These results are intermediate between those derived from other studies with these species. Overall, ocelots have a better chance of survival in Igua^u, mainly due to greater acceptance of humans of living in close proximity to this species, as opposed to jaguar. xii

PAGE 11

Some management recommendations are proposed. An environmental education program that includes working with neighboring land-owners (with compensation for losses due to livestock depredation where applicable) is also considered essential . xiii

PAGE 12

TABLE OF CONTENTS ACKNOWLEDGEMENTS iv ABSTRACT xi CHAPTER 1 INTRODUCTION 1 CHAPTER 2 STUDY AREA 5 CHAPTER 3 MATERIALS AND METHODS 16 CHAPTER 4 RESULTS: OCELOT 25 Radio-telemetry 25 Home Ranges 25 Movements 29 Activity 30 Social Interactions and Reproduction 31 Dispersal 34 Mortality 37 Density 39 Diet 40 CHAPTER 5 RESULTS: JAGUAR 78 Radio-telemetry 78 Home Ranges and Movements 79 Reproduction 81 Dispersal 83 Mortality 86 Density 88 Diet 88 Relationships with Humans 89 CHAPTER 6 DISCUSSION COMPARATIVE ECOLOGY OF OCELOT AND JAGUAR, WITH IMPLICATIONS FOR MANAGEMENT AND CONSERVATION 116 Management and Conservation 131 APPENDIX A MAMMALS RECORDED IN IGUA?U 145 APPENDIX B SIGHTING RECORDS 147 ix

PAGE 13

APPENDIX C TRAPPING RESULTS 149 APPENDIX D NOTES ON SMALL CATS IN IGUAgiU 167 APPENDIX E ENVIRONMENTAL EDUCATION ACTIVITIES . . 175 REFERENCE LIST 181 BIOGRAPHICAL SKETCH 190 X

PAGE 14

CHAPTER 1 INTRODUCTION As an ever-increasing portion of tropical and subtropical forests gives way to human development, research on the remaining fragments set aside as national parks and reserves becomes even more critical. In spite of recent studies on this subject (Malcolm, 1993, Lovejoy et al., 1986), much remains to be learned about the natural history of the fauna and flora and the dynamics of these ecosystems, if we are to manage and preserve these islands of habitat. In this context, predators, especially the larger species, play a vital role as indicators of ecosystem integrity in these reserves. Iguagu National Park, in southwestern Brazil, vividly exemplifies an isolated fragment of natural habitat surrounded by intensive human activities (Figure 1-1) . Created originally to protect the world-famous Iguagu Falls, the Park now represents about 80% of what remains of the once extensive subtropical forest that covered Parana state, east of the coastal Serra do Mar (Poupard et al., 1981). Now only 6% of the state consists of forest cover. Despite the proximity to burgeoning cities, such as Foz do Igua?u and Cascavel, the 1

PAGE 15

2 Park still harbors a largely intact array of its original fauna, including large predators, such as jaguar (Panthera onca) and puma ( Puma concolor) , and ungulates, such as peccaries ( Tayassu peccari and T. tajacu) and tapir ( Tapirus terrestris ) . From April 1990 through December 1994, I investigated the ecology of ocelots and jaguars in Iguagu National Park. In 1991 the project was extended to the neighboring Iguazu National Park, in Argentina, separated by the Iguazu river. Through the study of the natural history of these species and of the differential impact on them by human activities, I hoped to evaluate their conservation prospects in the Park. Additionally, I wished to determine if there was genetic interchange between the populations of the two Parks, and to study the dispersal patterns (sensu Schields, 1987) of subadults. Studies of these felids in other settings in Central and South America (Crawshaw and Quigley, 1989, 1991; Quigley and Crawshaw, 1992; Sunquist, 1992; Sunquist et al., 1989; Ludlow and Sunquist, 1987; Emmons, 1987, 1991; Rabinowitz and Nottingham, 1986; Konecny, 1989), usually conducted in large tracts of relatively undisturbed, continuous habitats, provided a useful comparative framework. Chapter 2 presents a brief description of the study area. In Chapter 3, materials and methods used in the study are presented. Chapters 4 and 5 summarize the results on ocelots and jaguar, respectively. In Chapter 6, I compare the ecology

PAGE 16

3 of these two species of f el ids in Iguaqiu in the light of results from studies elsewhere. I conclude this chapter with a discussion on management recommendations, as they apply to the conservation of these species in Iguagu National Park. I also included as appendices some summary results on different aspects of the study. With the understanding that management and conservation of predators are a direct result of the local human population's attitude towards these species, the project placed considerable effort in an environmental education program in and around the Park. A summary of these activities is included as Appendix E. A similar program was developed in Iguazu National Park, by Silvana B. Montanelli, that Park's biologist.

PAGE 17

4 Plate I~Mosaic of aerial photos of Iguagu National Park, Brazil. Note patchwork of agriculture completely surrounding the park on the Brazilian side (see also Figure 2-1).

PAGE 18

CHAPTER 2 STUDY AREA Iguaqu National Park (Fig. 2-1) lies on the western border of Parana state, southern Brazil (25" 05' to 25" 41' S and 53° 40' to 54" 38' W) , along the international boundary between Argentina and Paraguay. Presently 175,000 ha in size, the Park was created in 1939 to protect the Iguagu falls, which are considered one of the most spectacular natural sights in the continent. Situated in a region of rich soils and high human population density (45.8 inhabitants/km^; Poupard et al., 1981 and most likely much higher at present days) , the park is virtually an island of forest, completely surrounded on the Brazilian side by farmland, mostly plantations of soybean, wheat, and rice, and pastures for cattle. Unless otherwise stated, references to "the Park" will denote the Brazilian Iguaqru National Park, for the remainder of this paper. Together with the adjacent 55,500 ha Parque Nacional Iguazu, in Argentina, the Park harbors the last large tract of subtropical rainforest that once covered much of the highlands on the west side of the Serra do Mar, in southern Brazil. Although contiguous, the two Parks are administratively 5

PAGE 19

6 separate. Also represented within the Park, albeit to a lesser extent, is the dwindling Brazilian pine (Araucaria) forest. A preliminary mammal list identified 42 species of mammals, 13 (31%) of which are carnivores (Appendix A) . Within the carnivores, 38% are felids. Despite being one of the oldest national parks in the country, no intensive, longterm research had ever been done in the area. Crespo (1982) conducted a study on mammals in Iguazu National Park, Argentina. The shape of the Park resembles that of an "L" turned 90° counterclockwise, the average width (N/S) of the western half being only about 5 km (Figure 2-1) . The eastern portion, with approximately 50 km on the N-S axis and 23 km on the E-W axis (roughly 1,100 km^) , remains the least disturbed. There is no buffer zone or effective fencing between the Park and the surrounding private properties (Figure 1-1) , along the 136 km of dry boundary of the northern perimeter of the Park. A dirt road runs along the border of the western half for less than half this distance. The remaining 304 km of the perimeter is separated by rivers: 11.5 km by the lower Iguagu river (below the falls; mean width= 250 m) on the western end of the Park, 184 km of the upper Iguagu river (mean width= 1.0 km) which limits the Park to the south, and 107 km by the Gon^alves Dias river (mean width= 10 m) , on the eastern side of the Park. The close proximity of livestock and crops to the boundary of the Park function as strong attractants to

PAGE 20

7 wildlife. Common management problems are the invasion of crops, especially corn, by peccarie's and predation of poultry, dogs, pigs, sheep, and cattle by predators. The lack of perimeter roads in a great portion of the boundary hinders efficient patrolling by Park rangers, and facilitates invasion by poachers. A large proportion of property-owners around the Park are of Italian descent, among whom hunting is a cultural legacy. In addition to illegal hunting for sport, there is also subsistence hunting by people of low income. The study area encompassed about 80 km^ of the westernmost part of the Park, where the falls are located (Figure 2-2). Access to the falls is provided by a federal highway (BR-469) , that ends at the tip of the peninsula that cuts into Argentina, at the falls. This area is visited by close to one million visitors a year (mean of 802,587 during the study period, 1990-1994), with peaks in the months of January/February and in July (Figure 2-3). During some national holidays, the number of daily visitors may exceed 10,000. Most of the trapping and ground monitoring was conducted along the 8.8 km P090 Preto road (PPR) , which ran from the main road (BR-469) to the upper course of the Iguagu river (Figure 2-2) . The research base was located at the begining of this trail, which has been closed to the public since 1986. Three other trails also were used (Figure 2-2): the Represa trail (TR; 1.2 km); the Bananeiras trail (TB; 1.5 km); and the

PAGE 21

8 Macuco trail (TM; 4.5 km). The latter was the least used because of heavy public visitation, being used for a jungleand boat trip, run by a concessionaire. An 18 km, unpaved road used to cut through the Park, about halfway in the east-west length (Figure 2-1) , until 1986, when state authorities decided to pave the road. A local NGO took the case to court, first at the state, and then at federal level. The final sentence ruled that not only the road should not be paved, but also that it should be closed to traffic, permanently. The closing was enforced by federal police, in October 1986. Since then, from time to time, politicians of the affected counties (mainly Medianeira and Capanema) still advocate re-opening the road, putting considerable pressure on IBAMA officials, in Brasilia . The final outcome remains to be seen; meanwhile, nature takes its course, and the road today can hardly be seen, from the ground or from the air. The climate in the region is temperate subtropical, with mean monthly temperatures ranging from 25.7° C in February to 14.6** C in July (Crespo, 1982). The extremes during the study were -7° C and 47". Mean annual rainfall during the study period (1990-1994) was 1700 mm, with one peak in May (180 mm) and one in October (240 mm) . July and August are the driest months of the year (Figure 2-4) . The vegetation in the Park consists of a rather uniform forest cover, with some regenerating patches on the western

PAGE 22

9 half of the Park, where settlers had cleared or selectively logged the original forest. (Note: In 1975 there were over 400 families living in this area; the last resident was compensated and removed from the Park in late 1986) . There are also a few small, apparently natural clearings, usually associated with wetlands. A preliminary survey indicated 312 species of vascular plants, comprising 75 families, in Iguaq:u National Park. Four families (Graminae: 24 species, 8%; Euphorbiaceae : 19 species, 6%; Compositae: 18 species, 6%; and Rubiaceae: 17 species, 5%) represented 25% of the total number of species. Twenty-seven families were represented by only one species, indicating a high diversity of plants. The vegetation in two 0.13 ha plots along the PPR, selected for small mammal trapping, was studied in more detail. Plot A was located at km 0.7 of the PPR, and consisted of secondary forest, which was selectively logged in the past. Canopy height was between 10 and 15 meters, with emergent trees reaching 25 m (mainly Parapiptadenia rigida . Ruprechtia laxiflora, Peltophorum dubium, and Apuleia leiocarpa) . The dominant trees is this layer were Nectandra lanceolata and N. meqapotamica, Diatenopteryx sorbifolia . Cedrella fissilis . and Allophylus edulis. Trees in the intermediate story had an average height between 1.5 to 4.0 m, consisting mainly of Trichilia eleqans and T. catigua, Allophvlus guaraniticus . Piper amalago and P. gaudicha udianum . Psvchotria mvriantha and

PAGE 23

10 p. leiocarpa . At smaller densities, Pilocarpus pennatif olius ^ Ilex paracmayensis . introduced Citrus sp. , Hennecartia omphalandra . among others, also occurred. Noteworthy was the presence of two species of bamboo Merostachys sp. and Bambusa trinii. The latter species had apparently gone through a recent die-out. The bottom story, with plants up to 1 m, was composed predominantly of ferns ( Drvoteris sp. and Pteris sp. , and Didimochlaena truncatula) , with Hvdrocotyle cf . leucocephala in high densities also. Pharus cf . glaber, Olyra sp. , Adiantopsis radiata, Doryopteris sp. , Geophila macropoda . Pavonia malvacea and Saranthe sp. were also present. Both in the intermediate and lower strata, there were high densities of young individuals of the taller trees from the tree canopy, such as Parapiptadenia rigida . Diatenopteryx sorbifolia, Campomanesia xanthocarpa . Pataaonula americana . Bal f ourodendron riedelianum . and Sorocea bonplandii, among others. Vines and lianas are common in the area, mainly Pristimeria andina . Bauhinia microstachva . Wilbrandia sp., Macfadyena uncmis-cati . Arrabidea cf . mutabilis . Seriania sp. , Acacia sp . , and Chuscmea sp . . Philodendrom selloum was a common epiphyte. Plot B was situated in primary forest, at km 4.8 of the PPR. The most conspicuous difference between the two plots was the high density of Euterpe edulis in Plot B (70% of all trees with > lOcm DBH present, n= 101) , and its complete absence in plot A. The tree canopy (height from 10 to 20 m)

PAGE 24

11 was composed mainly by Euterpe edulis . Cedrella f issilis . Holocalvx balansae . Alchornea sidifolia . Cabralea canierana . Jacaratia spinosa . Nectandra meaapotamica . Chry sophy 1 lum aonocarpum . and Guarea kunthiana . Less common were Plinia rivularis . Campomanesia xanthocarpa . Cecropia sp. , Ficus cf . guaranitica . Lonchocarpus muehlbergianus . Inga marainata . and Syagrus r oman z on f i anna . In the intermediate story, with an height between 1.5 and 3.0 m, predominated Piper amalago, P. gaudichaudianum . Psychotria myriantha . P. leiocarpa . and Sorocea bonplandii ; some individuals of Eugenia burkartiana and Prockia crucis are also found. In the herbaceous layer, Didymochlaena truncatula. Dry opt er is sp. , Pharus cf. glaber . and Geophila repens . Compared with Plot A, ground cover was notably less, Hvdrocotyle cf. leucocephala (common in Plot A) was restricted to the influence of the road (the plots started about 6 m from the road, at a right angle) . There were many very young individuals of the larger trees, especially of Euterpe . There were also some lianas, including Pristimeria andina, Bauhinia mvcrostachia . and Pisonia aculeata . Considering only the trees with > 10 cm DBH, 54 species with a total of 196 individuals were identified in Plot A, as compared to 19 species with 101 individuals in Plot B.

PAGE 25

12 Figure 2-1 — Map of Iguagu National Park showing study area.

PAGE 26

13 Figure 2-2— Map of study area in Igua9u National Park, Brazil.

PAGE 27

14 120 Visitors (x 1000) Vehicles (x 1000) 20 Figure 2-3 — Number of visitors and vehicles entering Igua9u National Park, Brazil, during the study period.

PAGE 28

15 Figure 2-4 — Mean monthly temperature and rainfall for Iguacu National Park, Brazil, 1990-1994.

PAGE 29

CHAPTER 3 MATERIALS AND METHODS Animals were captured using custom-made wood, wire-mesh or iron-bar box-traps of different sizes, depending on the species. Measurements (length x width x height) of traps used for ocelots were 120 x 40 x 50 cm, and for jaguar, 210 x 80 x 80 cm. Traps were placed mainly along the 8.8 km of the PPR, although several other sites on the western-most part of the Park were also trapped. The PPR was marked every 100 m with numbered posts for permanent location reference. Live bait (chickens or laboratory rats) was housed at the back of the trap with wire mesh; once captured, the trapped animal could tear through the mesh and feed on the bait, thus reducing thrashing and stress while in the trap. Live piglets were used for two of the jaguar captures in Iguagu, and remains of dogs that had been previously killed by problem jaguars were used in two captures in Iguazu, Argentina. Capture attempts using a live goat were unsuccessful, although the jaguar circled and even climbed on the trap, without entering. The use of live bait also reduced the incidence of nontarget species, such as opossums (Didelphis) and tegu lizards ( Tupinambis ^ , in the traps. 16

PAGE 30

17 On four occasions, trained dogs were used on one successful jaguar recapture and one unsuccessful capture and two recapture attempts. In one of these, a problem jaguar wounded with buckshot by a local when preying on livestock killed 4 dogs and mauled another 5 during the hunt, and 1 dog was killed in another hunt. The denseness of the vegetation impeded keeping close to the dogs, after they were released on the cat's trail. Once trapped (or treed, in the case of jaguar recaptures) , animals were chemically restrained using a projectile dart shot with a COj pistol (Telinject U.S.A., Inc., Saugus, CA, 91350), or powder rifle (Capchur, Palmer Chemical & Equipment Co., Douglasville, GA 30133). Some were hand-injected using a noose (Ketch-All, Inc., S. Diego, CA 92104) to restrain the animal. Two drugs were used: ketamine hydrochloride (Parke-Davis & Co., Detroit, MI), and Zoletil (same as Tilazol or CI-744; Virbac do Brasil, Sao Paulo, SP 04021). Mean dosage of Ketaset for ocelots was 19.7 mg/kg ± 9.9 mg/kg (range: 10.0 40.4; N= 12;) and for Zoletil, the mean dosage was 6.5 mg/kg ±2.4 mg/kg (range: 3.2 13.3; N= 17). For the jaguar, the average rate for Ketaset was 29.9 ± 9.8 mg/kg (range: 18.2 40.4; N= 4), and for Zoletil, 7.0 + 1.0 mg/kg (range: 5.6 9.1; N= 16). Mean induction time with Zoletil was about 5 minutes, less than half the time required with Ketamine.

PAGE 31

18 Anesthetized animals were examined for general body condition, measured, weighed, ear-notched and/or tattooed, photographed, and fitted with radio-collars (150-152 MHz; Telonics, Mesa, AZ 85204; and Wildlife Materials, Inc., Carbondale, IL 62901) . Relative age was estimated mainly on the basis of presence of milk or permanent dentition, on tooth color and wear (juvenile, subadult, adult) , and these features were used in combination with other physical characteristics, such as weight, size, sign of previous reproduction (for females) , texture and color of the skin of the heel pads (Crawshaw, 1992) . Biometric data on captured carnivores are given in Appendix C. Anesthetized animals were kept under observation until ambulatory. Blood samples were collected from most animals and sent to the biochemical lab at Itaipu Binacional, for clinical and health evaluation of individuals. Blood samples from animals in the latter part of the study were additionally preserved in EZ -Blood, and sent to the National Cancer Institute, Frederick, MD, as part of a collaborative study on the genetics of neotropical felids (O'Brien et al., 1993). Results on these aspects of the study are still pending, and will be published elsewhere. Mean transmitter life for 10 ocelot radiocollars from Wildlife Materials, intended to last an average of 24 months or 720 days, was 165 ± 160 days (range 6 507), or an average of 23.2% ± 21.9% (range: 0.8 70.4%) of the projected operational life. Eight transmitters used in

PAGE 32

19 six ocelots (2 M, 4 FF, including recaptures) failed after < 3 months of use. For comparison, mean operational life for 3 Telonics transmitters was 144.0% ± 15.0% (range: 127% 155%) of the projected life. Another 6 Telonics transmitters still in use have already worked, on average, 73.7% ± 11.6% of their projected life. (Note: only new transmitters were considered in this analysis) . I lost contact with two jaguars wearing Telonics transmitters, but I do not know if the cause was radio malfunction or if the animals were killed and the transmitters destroyed. Radio-equipped animals were searched for from a vehicle, with a mounted omnidirectional antenna, or on foot with a directional antenna. Once a signal was heard, locations were obtained through triangulation, using the hand-held directional antenna. Given the limited ground range of the equipment, about 300m because of the dense vegetation, a light aircraft was used at approximately 14-day intervals to obtain an unbiased sample of locations. The "fly-by" method described by Mech (1983:69) was used to define animal locations from aircraft. Accuracy tests performed by project personnel indicated that locations could be described within a circle of 50 m radius (0.8 ha). One-hundred-eightyfour flights were made, totalling approximately 460 hours (average flight time= 150 minutes, mean of 4.5 animal locations per flight, range 1 16).

PAGE 33

20 Seventy-six percent of the 904 locations of ocelots and 81% of the 236 jaguar locations were aerial. Locations were plotted on a 1:25,000 map of the study area, divided into 1.5 ha cells by a transparent grid overlay. Grid coordinates (vertical and horizontal) were assigned to each location. In the latter period of the study (subsequent to 1992) , geographical coordinates were also obtained with a GPS (Global Positioning System) receiver (Transpak, Trimble Co., and Pronav 100, Garmin Co.). The GPS was particularly useful on jaguar recaptures, using dogs, and on locating transmitters on mortality mode on the ground, after initially locating the signal from the air and marking the coordinates. Home ranges were estimated with the Home-Range (Ackerman et al., 1990) and the Mcpaal (Micro-computer Program for the Analysis of Animal Locations, Stuwe, 1985; National Zoo, Washington, DC) softwares. Results are provided for the Minimum Convex Polygon (Mohr, 1947) , the Harmonic Mean (Dixon and Chapman, 1980) , and the 95% Ellipse (Jenrich and Turner, 1969) methods, for comparison between methods and between studies. However, due to the particular shape of the study area, with sharp boundaries represented by the Igua?u river (Figure 2), most of the estimates resulting from the two latter methods included areas known not to be used by some of the animals. Therefore, comparisons between individuals in this study were made using the Minimum Convex Polygon, corrected upon visual analysis of plotted locations. For

PAGE 34

21 comparative purposes, only the home-ranges of adult animals were used. Outliers were omitted from calculations, including locations of dispersing subadult individuals. A new method was devised to estimate home range size. The arithmetric center (AC) of the locations is determined, expressed by the means of the X and Y coordinates (White and Garrot, 1991) . Then the distance of all locations to the AC is calculated. The home range is defined by a circle of which the radius is the mean distance of locations to the AC, added to 2 standard deviations. A circle thus defined encompasses, on average, 95% of the locations. This method was used herein to compare spatial organization of the home ranges and as a measure of site fidelity between years for animals with over 24 months of monitoring. The coefficient of variation (CV) was used as a measure of the dispersion of locations in relation to the AC (Ott, 1984). Linear distances between simultaneous locations were used to analyze spacing between adult conspecifics with overlapping home-ranges. Only those locations obtained within a 40-min. period of each other were used for this analysis. All transmitters contained two types of motion sensors (Telonics, Mesa, AZ 85204, and Wildlife Materials, Inc., Carbondale, IL 62901) . One (mortality sensor) increased the signal pulse rate when the collar remained immobile for > 2 h, indicating the animal had died or dropped the collar. The other (instantaneous activity sensor) changed the pulse of the

PAGE 35

22 signal depending on the angle of the collar (head-up, headdown) . Signal pulse rate (active or inactive mode) was recorded at 15 minute intervals (Quigley et al., 1979; Sweanor, 1990) . A 45 m-high fire-tower close to my house and a 25 m tower at the hotel by the falls were used to monitor activity of animals within range (< 4 km) , for periods of up to 12 hours. On several occasions, the signal of animals radiocollared in Argentina were picked up from these towers. Too few activity records were obtained from the radio-tagged jaguars, however, because animals were rarely within range of these towers. A photographic camera remote system, activated by an infrared beam (Trail Master, Shawnee, KS) was used at the beginning of the study to document the presence of the different species of carnivores, and to aid in density estimates. However, with a few remarkable exceptions, the system was too often triggered by the abundant nocturnal insect life in Iguagu, and its use was discontinued. All sightings of carnivores and potential prey species were recorded (Appendix B) , together with location and time of day. Likewise, all roadkilled animals found in the Park and periphery were collected, and pertinent information on species, sex, age, health condition, reproductive data, date, and location was recorded. Whenever possible, the stomach and intestines were collected for analysis of contents.

PAGE 36

23 Carnivore fecal samples were routinely collected along roads and trails, dried in large cardboard boxes with lamps permanently lighted, and stored until analysis. Contents were separated into the various components in running water over a wire mesh sieve, and identified to the lowest taxon possible, using a local reference collection of hair, bones, reptile scales, and fruit. Some items were identified at the Capao da Imbuia Museum, in Curitiba, PR. The predator species that left the scat was identified at the collection site, whenever possible, by association with tracks and other circumstantial evidence. When this was not possible, hairs ingested in autogrooming were identified by comparison of cuticule and medulary patterns with hairs of known samples (Pereira Leite, pers. comm.). For this analysis, hairs were selected from samples by macroscopic characteristics, clarified with xylene, and compared with known samples at 400x magnifications. Regurgitated matter (vomits) and stomach and intestinal contents were analyzed in a similar fashion. As an additional attempt to identify predator scats, thin layer chromatography (TLC) was used (Johnson et al., 1984; Watt, 1987; Jimenez, 1993). Approximately 150 samples collected from captive animals of 14 species of carnivores were used to determine patterns (if any) of presence and behavior of 20 bile acids (12 identified, 8 unidentified). The tests were conducted at the College of Pharmacy lab, Universidade Federal do Parana, by a veterinarian with

PAGE 37

24 extensive laboratory experience, under direct supervision from Dr. Cid Aimbire, head of the Chromatography lab. By comparing the results of the control runs and field collected scats, we concluded that the inherent variation within and between groups was too great to facilitate reliable identification of unknown predator scats. Similar results were reached by Watt (1987) and Jimenez (1993). In an attempt to correct for differences in prey size in the diet of these predators, I multiplied the mean weight of each food item by the number of times it was found in the sample of scats, and by the mean number of individuals per scat, thus obtaining an estimate of biomass consumed for each taxa (B^j) . When available, I used weights from local animals; if unavailable, weights from the literature were used. The relative importance of each prey item was then expressed as the percentage of that item in relation to the combined weight of all items.

PAGE 38

CHAPTER 4 RESULTS: OCELOTS Radio-Telemetry A total of 21 ocelots (11 MM, 10 FF) was captured and radio-tagged, 15 (9 MM, 6 FF) in Iguaq:u, Brazil, and 6 (2 MM, 4 FF) in Iguazu, Argentina. Information on trapping results and measurements are given in Appendix C. The cimulative monitoring period for the 21 ocelots was 7,401 days, 3,871 (52%) for males, and 3,530 (48%) for females (Table 4-1). The total number of locations was 904, of which 695 (77%) were aerial. Mean interval between consecutive locations was 8.9 ± 10.1 days (range= 1-85 days). Thus, there was no risk of incurring problems with serial correlation or restricted access to the animals (Ackerman et al., 1991). Locations for some of the study animals are depicted in figures 4-1 through 4-15. Home Ranges Male ocelots used home range areas that were about twice as large as those of females. Estimates (Minimum Convex Polygon) of home range size were calculated only for animals with over 30 locations or which were monitored for more than 25

PAGE 39

26 3 months (Table 4-2) . Outliers, or atypical movements, were eliminated from estimates of home range size, but in a few cases some distant locations were included because of repeated use of these areas at 3to 4-month intervals. The mean home range size for 6 adult males was 38.8 ± 11.8 km^ (range= 20.4 50.9 km^) and for 5 adult females 17.4 ± 16.7 km^ (range= 3.8 40.4). For some animals, the cumulative area curves (Odum and Kunzler, 1955) reached an asymptotic limit after only 10 or 20 locations, whereas for others the home range sizes were still increasing at 50 locations (Figure 4-16, 4-17), suggesting that the full extent of their ranges had not been reached. This continual increase in home range size, especially for males, is likely a reflection of their occasional visits to distant parts of their range. It is noteworthy that the three females for which the curves reached an asymptote with less than 30 locations (F03, F09, and F40) , were all presumed to have small young (see Reproduction) during the period of monitoring. Even though monitored for short periods, F16 and F39 were included to show trend of large home ranges. Home range estimates using the Harmonic Mean (HM) and 95% Ellipse methods (ELL) are included in Table 4-2, for comparison (see also Figure 4-18, 4-19) . The MCP results were consistently more conservative than the other methods. The mean difference between the estimates obtained with the MCP and HM methods was 380% ± 126% (range= 212% 595%) for

PAGE 40

27 females, and 398% ± 84.4% (range= 295% 559%) for the males. Between the 95% Ellipse and the MCP, the difference was 227% ± 123% (range= 99% 464%) for females, and 196% ± 26.6% (range= 172% 242%) for males. Two adult females, F06 and F19, were captured early in the study and were monitored for 507 days (N= 61 locations) and 586 days (N= 68 locations) , respectively, before their transmitters failed. After intervals of 572 days for F06 and 349 days for F19 without contact, they were recaptured and reequipped with new transmitters (Table 4-1) . Since recapture, they have been monitored for 448 days (N= 30 locations) and 424 days (N= 42 locations) , respectively. In the first periods of monitoring, F06 used an area of 39.2 W and F19 used 22.6 km^. After recapture, they used areas measuring 24.0 km^ and 17.0 km^, a reduction of 38% for F06 and 25% for F19. Home range sizes calculated using the Circle method for 50% and 95% of locations are given in Table 4-3. The mean home range size (95%) for adult males using this method was 46.7 km^, and that for females was 18.4 km^. Males and females showed the same increase in size between the 50% and 95% estimates (z= -0.447, P= 0.65; Mann-Whitney two-sample test) ; the 50% estimate was, on average, 4.3 times smaller than the 95% estimate. This indicates a more concentrated use of the area closer to the center of the home range than toward the periphery, for both males and females. The difference between

PAGE 41

28 estimates obtained with this method and those with the MCP was 120% for males and 106% for females. The mean distance of locations to the arithmetic center was significantly greater for adult males (2.4 ± 0.6 km; N= 7; 352 locations) than for adult females (1.5 ± 0.4 km; N= 7; 346 locations; Z= -2.49, P= 0.013, Mann-Whitney test; see Figure 4-20) . Adult males seemed to use their home ranges in a more uniform pattern, over 50% of their locations were at distances between 1.0 and 3.0 km from the AC (Figure 4-21). Females, on the other hand, concentrated their movements in areas less than 2.0 km from the AC. Subadult females show an apparent tendency to remain in a small area, with over 60% of the locations within less than 1 km from the AC. The greater distances shown for the subadult males denotes the exploratory nature of their movements. The arithmetic centers of home ranges of same sex ocelots were widely spaced. Among all resident males, the mean distance between the AC of locations was 4.3 ± 2.2 km (range= 1.4 8.5), and for adult females it was 4.6 ± 2.2 km (range= 1.1 7.5; Table 4-4). A comparison of the distances between the AC of locations of radio-tagged ocelots show some potentially meaningful associations between animals. Noteworthy are the short distances between adult male M04 and M26 and M27. Among the females, the data support the assumption that F03 was the mother of F39, and F09, that of

PAGE 42

29 F40 (see Reproduction) . The distances shown between males and females confirm the association between M05 and F03 (see Reproduction) , while this male was still alive (see Mortality) . He had already died by the time F39 became an adult, and I believe he may have been her father. M35 was believed to associate to F09, likely being the father of F40 (see Reproduction) . Other possible associations are between M45 and F06, and M27 and F03 (although from their ages, it is possible that he is F03's son with M04) . Movements The denseness of the forest and the lack of trails precluded regular ground following of radio-tagged animals. However, on 23 occasions I obtained more than 1 location in the same day for 3 males and for 5 females, with a mean interval between locations of 7.3 ± 2.3 hours. In two instances, no movement was detected. On the remaining 21 occasions, the mean linear distance between these locations for the males was 1.9 ± 0.5 km (range= 1.2 2.6 km; N= 4), and for the females, 0.6 ± 0.5 km (range= 0.1 1.5 km; N= 17). The overall mean was 0.9 ± 0.7 (0.1 2.6) km. A one-way analysis of variance on the linear distance (km) between consecutive locations with a one-day interval for the various age/sex groups (see Table 4.5 for descriptive statistics) failed to show significant differences (F= 2.90, P > 0.05). However, Fisher's Least Significant Difference

PAGE 43

30 Test indicated that adult males travelled significantly greater distances than adult females and subadult females. Furthermore, pair-wise comparisons using Mann Whitney's two sample test indicated significant differences between the distances travelled by adult males x adult females (Z= 1.603, P= 0.108), adult males x subadult females (Z= 2.201, P= 0.028), subadult males x sxibadult females (Z= 2.417, P= 0.016), and adult females x subadult males (Z= 1.467, P= 0.143). Figure 4-22 shows the distribution of mean linear distances between consecutive locations for the different age/sex classes. Males tend to have an unpredictable movement pattern, in that they can be at any distance from the previous location. All other age/sex categories seem to have more predictable patterns. Activity A total of 1,115 activity readings was obtained on the study animals, with a mean of 46 ± 19 (range= 4-78) readings per hourly interval. Overall, ocelots were more active during nighttime (1800 h to 0559 h: 41% of 627 readings) than during daytime (0600 h to 1759 h: 34% of 488 readings; Z = -2.505, P = 0.012; Two-Sample Proportion Test, Hintze, 1987). After a brief peak of activity at about 1400 h (based on 34 readings) , there was a steady increase in activity level from 1600 h to 2000 h, then a decrease followed by two other peaks, a lesser one at 0200 h and a higher one at 0600 h (Figure 4-23) .

PAGE 44

31 There was no statistical difference between levels of activity of males x females. However, males were significantly more active at night than during daytime (43.6% of 289 readings and 31.4% of 210 readings, respectively; Z= 2.758, P= 0.0058, Two-Sample Proportion Test, Hintze, 1987). Social Interactions and Reproduction Four of the seven adult females (F03, F06, F09, F19) were at different stages of pregnancy when captured, and F40 had recently given birth (Figure 4-24; see Appendix C, Table C-1 for dates) . Birth dates were estimated from monitoring pregnant females and by back-dating from the estimated ages of young animals. When captured on 23 August 1990, the mammae of F03 were swollen. On 9 September, I walked in to check a mortality signal from her transmitter, in very dense undergrowth, and approached the transmitter to within < 15 m. As I searched for her presumed carcass, the signal suddenly returned to the active mode and the animal rapidly left the area. On investigating the site, I found where she had been lying, in a den under a fallen log. Since the mortality mode of the transmitter is only activated after 2 hours of immobility, she must have remained that way for at least 4 hours (including the time since I first heard the mortality signal) . A similar instance had once occurred with a radio-collared female bobcat fFelis rufus) that gave birth during a study in the Big

PAGE 45

32 Cypress National Preserve, FL (Crawshaw, 1988) . On 11 September, I once again followed the mortality signal from her transmitter, 0.9 km from the previous site. However, when the signal resumed the active mode, I left without further disturbing her. On 16 September, when she was recaptured, she had already given birth. I also discovered that her transmitter had failed, and I replaced it with a new one. Until 2 February 1991, when the new transmitter failed also, she was located 19 times, with a mean distance between locations of 1.7 ± 0.9 (0.0 3.1) km. On 8 occasions she was found close to her supposed den site, and I strongly suspect she had young there. It is interesting that M05, whose home range encompassed all of the known area of F03, was located on two occasions in the den area. On one of the locations, on 24 September, he was apparently with the female. Despite intensive trapping efforts in that general area, this female was not recaptured. Three and a half years later, on 13 and 19 February 1994, F39 and M37 were captured at the same site, within F03's former range. Their ages were estimated at approximately 4 and 1.5 years, respectively. Given that the locations of both animals (before M37's dispersal and F39's death) were concentrated over F03's former range (Figures 4-14, 4-10, and 4-7, respectively), it is tempting to suggest that F39 may have been F03's offspring, from the 1990 litter, and that one of these females was M37's mother.

PAGE 46

33 Between 18 October and 20 December 1990, F09 was located 5 times, with a mean linear distance of 0.8 ± 0.6 km (range= 0.0 1.8 km) between her locations, suggesting she may have young. On 19 February 1992, 15 months later, M25 was captured in the same area, at an estimated age of 1.5 years and weighing 10 kg. Unfortunately, his transmitter failed soon after capture, only 3 locations having been obtained, all of which were within F09's home range. Neither of the two were recaptured. However, on 21 September 1993, an adult male (M35, estimated age of 6 years) was captured within F09's range, and part of his home range encompassed all of that female's known range (Figures 4-5 and 4-9) . In addition, F40 (estimated age of 2.0 years) was also captured in the same area, and her range overlapped entirely the northern portion of F09's former range (Figures 4-11 and 4-9, respectively). Again, judging by the combination of data on site of capture, estimated age, plotted locations, and sex, it is tempting to suggest that M35 and F09 are the parents of M25 and F40, from subsequent litters. If this is the case, it is likely that M35 is also the father of M40's litter. Hopefully, genetic analysis of blood samples from the two latter captures will shed some light on the relationship between them.

PAGE 47

34 Dispersal A total of 7 ocelots (4 MM, 3 FF) was captured as subadults, 3 males (M25, M37, M41) and 1 female (F02) in Iguagu, Brazil, and 1 male (M36) and 2 females (F23, F42) in Iguazu, Argentina. The transmitter of M25 apparently failed soon after capture, after only 3 locations. M36 was killed about 13 km from his capture site, about three weeks after capture, and the collar was destroyed (see Mortality) . Three ocelots (2 MM, 1 F) crossed the Iguagu river from one Park to the other. The two subadult males (M37 and M41) crossed from Iguagu to Iguazu, and the female (F23) , who was initially captured in Iguazu, was later recaptured in Iguagu. M37 was captured on 13 February 1994, and he was monitored for 183 days in the Brazilian park. During this time he was located 15 times at intervals of 11.2 ± 6.6 days (range= 3-32 days) and he moved an average of 1.3 ± 0.7 km (range= 0.3 2.5 km) between consecutive locations. On 15 August 1994, he was located in the Argentinean park, 0.6 km from his Brazilian location on 31 July. On 16 August, he was again back on the Brazilian side, 0.9 km from the previous location. He then remained in Iguagu for another 107 days, with a mean linear distance of 1.0 ± 0.6 km (range= 0.1 2.3 km) between consecutive locations (N= 9) . Between 30 November and 15 December, he returned to the Argentinean side, moving an average of 7.6 ± 4.9 km (range= 2.0 14.0 km), between 3 locations. On 15 December, he was once again located in

PAGE 48

35 Iguagu, and remained there until 17 February 1995, during which time the mean distance between 5 consecutive locations was 2.8 ± 3.0 km (range= 0.4 8.7 km). On 17 February, he had crossed the river one more time and has remained in Argentina since, with a mean distance between 4 locations of 3.4 ± 1.6 km (0.9 5.4 km). M41 was captured in Iguagu NP on 4 April 1994. He remained there for 35 days, during which time he was located 12 times with a mean distance between locations of 4.0 ± 4.0 km (range= 0.3 15.6 km). On 9 May, he was found in Iguazu NP, Argentina, 3.1 km from the location on the previous day. When located again on 29 May, he had returned to Brazil, 4.3 km from the previous location. On 29 November, he was again located in Argentina. Over the next 16 days, he was located 5 times, at a mean interval of 5.8 ± 4.2 days (range= 1-13 days), with a mean distance of 5.4 ± 3.5 km (range= 1.0 9.6 km) between consecutive locations. On 24 January 1995, he was once again found in Iguagu, where he has remained there since. During this period, the mean distance between 5 locations was 1.5 ± 0.6 km (range= 0.6 2.2 km). F23 was first captured as a subadult in Iguazu NP on 25 January 1992, but her transmitter failed soon after capture. She was recaptured in the Brazilian park on 28 August, after being sighted feeding on the carcass of a roadkilled crabeating fox f Pus icy on thous) . She was very lean (7.9 kg against 10.0 kg on the first capture), and had been recently

PAGE 49

36 involved in a fight, presumably with a resident animal (intraspecif ic?) . Her right ear was missing and she had open gashes on the chest and limbs. She was taken into captivity for 55 days, for treatment, and was released, after recovery, close to the recapture site on 22 October, at a weight of 12.2 kg. The mean linear distance between 21 locations during the next 217 days was 3.4 ± 2.4 km (range= 0.1 7.7 km). After 3 April 1993, her signal was not heard despite extensive aerial searches. On 27 May 1993, the mortality signal from her collar was heard, and a location indicated the collar was in the Iguagu river, 30.5 km from her previous location. Although the collar could not be recovered, its location suggested she may have been poached for if she had drowned trying to cross back to Argentina, her carcass would have been washed to the shore. I later confirmed this theory (see Mortality) . F02 may have been another case of female dispersal. She had been captured in July 1990 on a property adjacent to Iguaqsu, and was kept in captivity for about 2 weeks before I was informed. She was then radio-collared and released in the Park, at a straight-line distance of 6.0 km from her capture site. By the following day she had returned to that same area, 5.0 km from the release site. After 101 days, her collar failed. During this period the mean linear distance between 36 locations was 0.8 ± 0.9 (0.0 5.0) km. On 24 April 1991, she was recaptured and her transmitter was

PAGE 50

37 changed. In the 71 days this collar transmitted, before also failing, the average distance between 21 locations was 1.1 ± 0.7 km (range= 0.0 2.2 km). On her last location, on 9 July, she had apparently followed the remnants of gallery forest along the Iguaqru river outside the Park towards the town of Foz do Iguagu, and was found 7.0 km from her previous location, on 26 June, 1991. .. , ^ Mortal itv One female (F39) was killed by a bus outside of Iguagu, on the night of 15 August 1994. The driver delivered the carcass, still with the radio-collar, at the' entrance of the Park. The radiocollar of another female, F16, was found by the main road in Iguazu NP, Argentina. Because it had been cut open, I assumed the animal was likely killed by a vehicle or shot, her whole carcass being taken. I recently confirmed the death of two other study animals. F23 was killed on the Brazilian side by an Argentinean poacher and the collar was thrown in the Iguagu river (see section on Dispersal) , as the poacher returned to Argentina. M36 was killed by a farmer while the cat was raiding a chicken coup by the Parana river, where it divides Argentina and Paraguay, about 13 km from his capture site. It is possible that some of the animals (F02, M25) , whose signals disappeared were poached also, having had their collars destroyed. This may have been a negative by-product of local

PAGE 51

38 piiblicity on the project and on radio-telemetry, since in the early days of the study I followed the signals of two collars (from a fox and a jaguar) to the poachers' homes, who had taken them as souvenirs. The carcass of M05 was found, by following the mortality signal, in a dense bamboo undergrowth. His already decomposing body showed evidence of a fight, although the surrounding vegetation was undisturbed. After being badly wounded, the animal had likely searched for a secluded place to die. Considering the overlap between the ranges occupied by M05 and M04 (older and larger than M05) , and the fact that immediately after M05's death, M04 extended his home range over all of M05's area, it seems likely that M05 died from injuries sustained in an intraspecif ic dispute with M04. Following Trent and Rongstad (1974) and Heisey and Fuller (1985) , I calculated survival and mortality rates for the ocelots in Igua9u (Table 4-6) using the formula: s,-^_j£i^j^ and mj= 1-s., where s, is the daily survival rate, x. is the total number of transmitter-days during interval y,. is the total number of mortalities occurring during interval j, and mj is the daily mortality rate in interval .. Overall daily survival rate (s,.) and daily mortality rate (m.) for the radio-tagged ocelots were 0.9944 and 0.0056, respectively. Considering only full years of monitoring (1991-1994), the mean daily survival rate was 0.9914 ± 0.0114

PAGE 52

(Table 4-6) . The cumulative mortality reached an asymptote at 24% after 600 days of monitoring (Figure 4-25) , with no more Icnown mortalities for the remainder of the study (May 1995) . Density Six adult males were radiocollared in the study area in INPBr. One of them, M05, was killed, likely by M04, whose home range thereof encompassed the previous range of M05. The home ranges of these five males covered the entire study area. Five adult females were radiocollared and monitored. Two of the females (F06, F39) included in their ranges some small forest fragments outside of the Park boundary. Based on an empty space between the ranges of these females, I assumed there was at least one (possibly two) adult female that was not captured. Therefore, I estimated a density of 13.7 adult ocelots/100 km^ in the study area (or 1 adult ocelot/8 km^; or 0.12 adult ocelot/km^). In addition, four subadults (3 MM, 1 F) were captured in the same area during the study period. Assuming a mean of 1 young per adult female/two years, this would give a total of 14 animals (5 adults males, 6 adult females, 3 subadults) using the 80 km^ study area, or a density of 17.5 ocelots/100 km^ (or 1 ocelot/5.7 km^) . Assuming a homogeneous distribution over the 1,750 km^ of the Park, the total estimated population would be approximately 300 ocelots.

PAGE 53

40 Diet A total of 362 carnivore scats were found in Iguagu National Park during the study period. Fifty-six (15.4%) were positively identified as ocelots scats. Eighty food items of 18 different taxa were recorded (Table 4-7), with an average of 1.4 items per scat. Mammals comprised 80% of the total number of items (Figure 4-26) , with small rodents accounting for almost half of the mammals consumed in terms of frequency of occurrence (Figure 4-27) , followed by opossum and armadillo (11% each), and agouti (9.4%). However, if ranked according to the estimated biomass consumed (mean weight of prey x number of individuals per scat x number of occurrences) , small rodents are preceded by another eight prey items (Figure 428) . The fox was known to have been taken as carrion, when F23 was seen feeding on a roadkill (see Dispersal) . It is possible (and likely) that animals such as adult Procyon and Mazama and other large species are also consumed as carrion, since they are sometimes killed by cars in the park, and may at times become important alternative food sources. However, since these large prey were rarely found in ocelot scats, their overall contribution is likely to be negligible. In Figure 4-29, I considered only the most frequently taken taxa (> twice in my small sample) . The prominence of armadillos, opossum, agoutis, and lizard in ocelot diet are in agreement with data on relative abundance, based on sightings, of these species in the Park (see Appendix B) . All these species are

PAGE 54

41 < 1/3 of the mean ocelot body weight, as compared to the other large items that appeared only once in the sample.

PAGE 55

1 42 Table 4-1 — Monitoring period for 21 ocelots in Iguagu and Iguazu National Parks, Brazil and Argentina. Letter with the animal number indicate the sex; C= capture; R= recapture (a number indicates number of recaptures during that month); Ks animal killed; F= transmitter failure; D= animal dropped collar. 1990 1991 1992 1993 1994 JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND F02 C2 — F...R — F F03 CFR F F06 C F R*'''. > F09 C F ' •' F16 C K F19 C F r'' > F23 CF R'''' K F39 C K F40 C > F42 C > 1990 1991 1992 1993 1994 JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND M04 C R** D M05 C3K M20 C P M25 CF N26 C F M27 CR-3R F R > M35 C > M36 CK M37 C > M41 C R > M45 C > Animal taken into captivity for treatment Collar changed

PAGE 56

43 Table 4-2 — Home range estimates for 21 radio-collared ocelots in Igua^u and Iguazu National Parks, Brazil and Argentina. N= number of locations; MC= Minimum Convex Polygon method; HM= Harmonic Mean method; ELL= 95% Ellipse method . AN # N MC HM ELL F02 59 15.6 15.6 27.8 F03 33 5.2 15.8 12.4 F06 91 40.4 136.0 44.0 F09 36 7.6 37.0 20.2 F16 10 12.1 25.7 56.2 F19 110 30.0 102.5 29.6 F23 19 27.8 127.3 112.7 F39 23 13.1 78.0 31.0 F40 40 3.4 15.2 6.9 F42 21 5.1 40.3 13.7 MO 4 60 20.9 59.2 47.5 M05 27 8.3 18.2 15.3 M20 38 41.5 99.7 75.2 M26 57 46.2 119.5 92.5 M27 78 50.9 119.2 87.0 M35 34 40.4 134.8 60.7 M37 32 73.1 94.1 M41 40 121.1 441.9 189.2 M45 26 27.7 67.4 52.6 Although area estimates are given for subadult animals, they cannot be considered actual home ranges.

PAGE 57

Table 4-3 — Mean distance (km) of locations to the arithmetic center (AC) for radio-tagged ocelots in Igua^u and Iguazu National Parks, and home ranges estimates (km2) using the mean distance to AC (50% HR) and mean distance + 2 SD (95% HR; see Methods) . No outliers were excluded. AN #/AGE N MEAN 4 SD RANGE 50% HR 95% HR F02 (SA) 61 0.9 + 1.1 0.1 7.1 2.5 33.3 F03 (AD) 33 1.2 + 0.5 0.4 2.0 4.5 17.4 F06 (AD) 91 1.8 + 1.1 0.1 5.0 10.2 54.2 F09 (AD) 36 1.7 + 0.6 0.5 3.4 9.1 29.3 F16 (AD) 10 1.8 + 0.7 0.9 3.2 10.2 35.4 F19 (AD) 110 1.5 + 0.9 0.1 4.2 7.1 37.5 F23 (SA) 25 4.2 + 5.8 0.8 • 32.1 55.4 800.1 F39 (AD) 23 1.6 + 0.8 0.7 4.3 8.0 35.3 F40 (AD) 42 0.7 + 0.3 0.1 1.8 1.5 6.6 F42 (SA) 32 1.4 + 0.9 0.3 3.9 6.1 35.4 M04 (AD) 53 2.6 + 1.1 0.5 4.9 21.2 77.2 M05 (AD) 27 1.3 + 0.7 0.4 3.6 5.3 25.6 M20 (AD) 63 3.2 + 1.5 0.1 6.6 32 . 1 126. 9 M26 (AD) 58 2.8 + 1.6 0.5 6.2 24.6 119.1 M27 (AD) 83 2.8 + 1.4 0.3 5.9 24.6 104.1 M35 (AD) 42 2.5 + 1.5 0.5 5.8 19.6 100.5 M37 (SA) 43 3.0 + 3.7 0.3 15.3 28.3 350.2 M41 (SA) 42 3.5 + 3.5 0.3 •19.1 38.5 356.8 M45 (AD) 26 1.9 + 1.0 0.5 4.3 11.3 51.7 Note : Although area estimates of subadult animals are included, they cannot be considered as actual home ranges.

PAGE 58

45 Table 4-4 — Distance (km) between the arithmetic centers (AC) of home ranges of adult ocelots in Iguagu National Park, Brazil. M05 was not included in other comparisons because he was killed (by M04?, see text) before other animals were radio-collared. MALES M05 M26 M27 M35 M45 M04 4.4 2.0 1.4 6.9 4.2 M26 2.6 8.5 5.4 M27 5.9 3.0 M35 3.5 FEMALES F06 F09 F39 F40 F03 4.5 6.2 1.1 6.8 F06 3.6 4.9 3.0 F09 7.0 1.8 F39 7.5 MALES/ FEMALES F03 F06 F09 F39 F40 M04 2.8 5.5 5.4 3.8 6.7 M05 0.1 4.6 6.3 1.0 7.0 M26 2.7 6.8 7.3 3.3 8.5 M27 1.7 4.3 4.9 2.8 5.9 M35 6.7 2.5 2.4 7.2 0.8 M45 3.2 1.4 3.7 3.7 3.8

PAGE 59

46 Table 4-5 — Mean linear distances moved between locations with a 1-day interval for 20 ocelots in Iguaqru and Iguazu National Parks, Brazil and Argentina. N= number of locations. AGE/SEX N MEAN ± SD RANGE Adult males (n= 6) 34 2.1 ± 1.7 0.0 7.3 Adult females (n= 7) 30 1.3 ± 1.1 0.0 4.0 Subadult males (n= 4) 15 1.9 ± 1.3 0.3 5.8 Subadult females (n= 3) 20 1.1 ± 1.0 0.0 5.0 All animals (n= 20) 99 1.6 ± 1.4 0.0 7.3

PAGE 60

Table 4-6 — Daily survival and mortality rates of radiotagged ocelots in Igua^u and Iguazu National Parks, Brazil and Argentina. N,-= number of animals monitored; # LOC= number of locations in the period; D^= number of animals that died; LC= transmitter failure; s,-= daily survival rate; m,= daily mortality rate. Adapted from Trent and Rongstad (1974) and Heisey and Fuller (1985) . YEAR # MONTHS # LOG LC m, 1990 5 6 120 1 0 0.9916 0.0084 1991 12 8 189 1 4 0.9947 0.0053 1992 12 7 156 0 1 1.0000 0.0000 1993 12 8 79 2 2 0.9746 0.0254 1994 12 10 282 1 0 0.9964 0.0036 1995 5 9 78 0 0 1.0000 0.0000

PAGE 61

48 Table 4-7 — Food items found in 56 ocelot scats in Iguagu National Park, Brazil. WGT= mean live weight (kg) of prey item; N= number of scats the item was found in; #IND= number of individuals per scat; ITEMS= percentage of total number of items found; % SCATS= percentage of scats containing that food item; ; B = estimated biomass of that prey item. PREY I*EM WGT N #IND ITEMS SCATS B . est small rodents 0.05 31 2 41.3 55.3 3.1 Opossum 1.50 7 1 9.3 12.5 10.5 (Didelohis aurita) Armadillo 3.30 7 1 9.3 12.5 23.1 (Dasvpus novemcinctus ) Agouti 3.20 6 1 8.0 10.7 19 .2 (Dasvprocta azarae) Invertebrates 6 — (insects= 5; land snail= 1) Rabbit U . / o 1 1 A 0 5 . 3 2 . 34 (Svlvilaqus brasiliensis ) Small marsupial 0.01 3 1 A n 0 • -J Bird 0.65 3 1 4.0 5.3 1.95 (Cracidae, Tinamidae) snake 1.00 3 1 4.0 5.3 3.0 Lizard 1.60 2 1 2.7 3.6 3.2 (Tupinambis tequixin) Unidentified fruit 2 2.7 3.6 Cavy 0.40 1 1 1.3 1.8 0.40 (Cavia aperea) Porcupine 3.50 I I 1.3 1.8 3.50 (Coendou prehensilis) squirrel 0.80 1.3 1.8 0.80 (Sciurus aestuans) Deer 22.5' 1.3 1.8 22.5 (Mazama sp. ) Fox 6.0 1.3 1.8 6.0 (Dusieyon thous)'' Raccoon 10.0 1.3 1.8 10.0 (Procyon cancrivorus) Grison 1.1 1.3 1.8 1.1 (Galictis cuia) TOTAL 75 99.7 133.8" ' Average of the weights of the two species (M. americana and M. nana ; ^ Taken as carrion (see text); " Does not add to 100% because of more than 1 item per scat.

PAGE 62

49 Figure 4-1 — Radio-locations (•) and capture locations () of an adult male ocelot (M04) , Iguagu National Park, Brazil. Monitoring period: 30/08/90 14/10/91.

PAGE 63

50 Figure 4-2 — Radio-locations (•) and capture locations () of an adult male ocelot (M05) , Igua^u National Park, Brazil. Monitoring period: 01/09/90 03/11/90.

PAGE 64

51 Figure 4-3 — Radio-locations (•) and capture locations () of an adult male ocelot (M26) , Igua9u National Park, Brazil. Monitoring period: 23/05/92 06/10/93.

PAGE 65

52 Figure 4-4 — Radio-locations (•) and capture locations () of an adult male ocelot (M27) , Igua9u National Park, Brazil. Monitoring period: 27/05/92 present.

PAGE 66

53 Figure 4-5 — Radio-locations (•) and capture locations () of an adult male ocelot (M35) , Iguagu National Park, Brazil. Monitoring period: 21/09/93 present.

PAGE 67

54 Figure 4-6 — Radio-locations (•) and capture locations () of an adult male ocelot (M45) , Igua^u National Park, Brazil. Monitoring period: 11/05/94 present.

PAGE 68

55 Figure 4-7 — Radio-locations (•) and capture locations () of an adult female ocelot (F03) , Igua?u National Park, Brazil. Monitoring period: 23/08/90 02/02/91.

PAGE 69

56 Figure 4-8 — Radio-locations (•) and capture locations () of an adult female ocelot (F06) , Igua^u National Park, Brazil. Monitoring period: 15/09/90 present.

PAGE 70

57 Figure 4-9 — Radio-locations (•) and capture locations () of an adult female ocelot (F09) , Igua^u National Park, Brazil. Monitoring period: 18/10/90 02/07/91.

PAGE 71

58 Figure 4-10 — Radio-locations (•) and capture locations () of an adult female ocelot (F39) , Iguagu National Park, Brazil. Monitoring period: 19/02/94 15/08/94.

PAGE 72

Figure 4-11 — Radio-locations (•) and capture locations (} of an adult female ocelot (F40) , Igua^u National Park, Brazil. Monitoring period: 18/03/94 present.

PAGE 73

60 Figure 4-12 — Radio-locations (•) and capture locations () of an adult female ocelot (F19) , Iguazu National Park, Argentina. Monitoring period: 03/08/91 present.

PAGE 74

61 Figure 4-13 — Radio-locations (•) and capture locations () of a subadult female ocelot (F42) , Iguazu National Park, Argentina. Monitoring period: 05/04/94 present.

PAGE 75

62 Figure 4-14 — Radio-locations (•) and capture locations () of a subadult male ocelot (M37) , Iguagu National Park, Brazil. Monitoring period: 13/02/94 present.

PAGE 76

63 Figure 4-15 — Radio-locations (•) and capture locations () of a subadult male ocelot (M41) , Iguagu National Park, Brazil. Monitoring period: 04/04/94 present.

PAGE 77

64 Figure 4-16 — Cumulative area curve for home ranges (Minimum Convex Polygon) of 7 adult female ocelots in Igua
PAGE 78

65 Area (Sq. km) 0 10 20 30 40 50 60 70 80 90 Number of Locations — M04 -^M05 M20 -BM26 -XM27 M35 M45 Figure 4-17 — Cumulative area curve for home ranges (Minimum Convex Polygon) for 7 adult male ocelots in Iguagu and Iguazu National Parks, Brazil and Argentina.

PAGE 79

66 AREA (knn2) HO I F03 FOB F09 F16 F19 F39 F40 ANIMAL # MCP MhM ^95% ell ^Circle Figure 4-18 — Comparison of home range estimates (km^) for 7 adult female ocelots in Iguagu and Iguazu National Parks, using different estimation methods. MCP= Minimvim Convex Polygon; HM= harmonic Mean; 95% ELL= 95% Ellipse.

PAGE 80

Figure 4-19 — Comparison between home range estimates for 7 male ocelots from Iguagu and Iguazu National Parks, Brazil and Argentina, using different estimation methods.

PAGE 81

68 Km 5 4 3 2 1 0 ADM IN7) ADF N7) SAM IN21 SAF IN31 AGE/SEX Figure 4-20 — Distance (mean ± 1 SD; km) from each ocelot's locations to the arithmetic center of their home ranges (see methods) . ADM: adult males; ADF: adult females; SAM: subadult males; SAF: subadult females.

PAGE 82

69 PERCENT OP LOCATIONS ADM ADF SAM SAP AGE/SEX Distance (Km) <1 ^1/2 m2/3 ^3/4 ^)5 Figure 4-21 — Percent of locations at various distances from the arithmetic center of different age/sex classes of 21 ocelots in igua9u and Iguazu National Parks, Brazil and Argentina.

PAGE 83

70 PERCENT OF LOCATIONS 6 0 1 ADM ADF SAM SAF AGE/SEX ~ ' Distance (Km) 1(1 ^1/2 Ha2/3 ^3/4 B)5 Figure 4-22 — Distances (km) between consecutive locations for different age/sex classes of ocelots in Iguagu and Iguazu National Parks, Brazil and Argentina.

PAGE 84

71 PERCENT ACTIVE 100 I 80 Ol I I I I — I — I I I I I — I I I I — I — I — I — I — \ — I I I 0600 0800 1000 1200 MOO 1600 1800 2000 2200 240G G200 0400 TIME OF DAY Figure 4-23 — Hourly percent activity for 1,115 readings on 21 ocelots in Igua^u and Iguazu National Parks, Brazil and Argentina.

PAGE 85

72 NUMBER OP OCCURRENCES Figure 4-24 — Temporal pattern of reproductive events of female ocelots in Igua^u and Iguazu National Parks, Brazil and Argentina.

PAGE 86

73 Percent 0 60 120 180 240 300 360 420 480 540 500 660 Days Cumulative Mortality Animals monitored Figure 4-25 — Cxmulative mortality of radio-tagged ocelots in Iguagu and Iguazu National Parks, Brazil and Argentina. Percent mortality and number of animals remained stable after 660 days of monitoring for the remainder of the study.

PAGE 87

74 Figure 4-26 — Percentage of the different taxa found in 56 ocelot scats in Iguagu National Park, Brazil.

PAGE 88

75 Small lodentf Agouti Figure 4-27 — Frequency of occurrence (%) of mammal species found in 56 ocelot scats, Igua^u National Park, Brazil. Other carnivores include the crab-eatingfox, crab-eating raccoon, and grison (one occurrence each= 1.6%).

PAGE 89

76 % Estimated Biomass [kgl 2 5 -I Airoadillo D*«i Agouti Opoiium Raccoon Raptilea Fox Poicupina Small todents Figure 4-28 — Prey items found in 56 ocelot scats in Iguagu National Park, ranked by estimated biomass represented in scats (%) . Snake and lizard were combined as reptiles.

PAGE 90

77 Figure 4-29 — Percent estimated biomass (kg) of taxa found > twice in a sample of 56 ocelot scats in Igua
PAGE 91

CHAPTER 5 RESULTS: JAGUAR Radio-Telemetry Seven jaguars were captured and radio-collared, including two adult males, one adult female, three subadult males, and one subadult female. One of the adult males (M48) was captured in Iguazu National Park, Argentina; all other animals were captured in Iguagu, Brazil. Another adult male was captured on the edge of Iguazu and translocated as a problemV animal to a provincial park south of the Argentine park. This animal was not radio-collared. Appendix C summarizes information on trapping results and morphometries of individuals. Five of the jaguars in Iguagu were probably related. F17, the adult female, was presumed to be the mother of M13 and F21 (littermates) and M32 (from a subsequent litter) . As the resident adult, M33 was probably the father of both litters. A total of 236 locations was obtained on the study animals, 142 (60%) for males, and 94 (40%) for females. Adult and subadult animals were monitored about equally, with 49% 78

PAGE 92

79 and 51% of locations, respectively. The cvimulative monitoring period was 2,292 days, 1,764 (77%) for males and 528 (23%) for females. Overall, the mean interval between locations was 10.1 ± 15.2 days (range: 0 102 days). Eighty percent of the locations were aerial. Tables 5-1 and 5-2 provide information on the monitoring of the study animals and Figures 5-1 through 5-6 show the individual maps with plotted locations. Home Ranges and Movements Home range estimates (Minimum Convex Polygon method) varied considerably for the study animals (Table 5-3) , ranging from 8.8 km^ (F21) to 138 km^ (M13) . Undoubtedly, some of this variation can be accounted for by differences between sex and age classes. However, the cumulative area curves (Odum and Kunzler, 1955; Figures 5-7, 5-8), show the ranges of most individuals are underestimated, due to small sample sizes and to reasons discussed below. The only exception is F17, for which an asymptote was reached at 50 locations. This female utilized a total area of 70.0 km^ during 14.5 months of monitoring. In addition to the short monitoring period, the movements of most of the radio-tagged jaguars were influenced by human activities. Adult male M48 was translocated 8 km from his first capture site, as a problem-animal. His movements were further affected by a shotgun wound, when he resumed predation on livestock (see Mortality and Relationships with Humans) .

PAGE 93

80 The movements of M33, a male well past his prime, were likely influenced by an acquired dependency on the live baits in my traps (8 recaptures; see Appendix C) . This may have been due to increased difficulty in his hunting ability, since his condition deteriorated steadily between captures. Similarly, the movements of F21, a subadult female, were influenced by a dependency on human-originated food, at the garbage dump at the hotel by the falls, and then by a short translocation following capture (see Mortality) . Captured as subadults, M13 and M32 were in the process of establishing their adult home ranges when they were killed by poachers, in very similar situations (see Mortality). The signal of subadult M34's transmitter disappeared after only 21 locations, while he was still using his natal area. As with the ocelots (see previous chapter) , the dense vegetation and lack of an adequate trail system precluded routine ground monitoring of radio-tagged animals. Due to the small sample sizes, no statistics could be appropriately applied to the data. Six of the study animals were located with a 1-day interval on 24 occasions. For all animals, the mean linear distance moved on these locations was 1.3 ± 1.2 km (range: 0.2 5.4 km). Overall, males tended to be found at greater distances than females on consecutive locations at increasing intervals (Figure 5-9) . Table 5-2 shows the mean linear

PAGE 94

81 distances between consecutive locations for each of the radiotagged jaguars. Likewise, males were found more often at grjeater distances from the arithmetic center (AC) of their ranges than females (Figure 5-10) . The distribution of locations in relation to the AC shows a tendency of greater use of the periphery of the home ranges for adult males. The adult female, on the other hand, displayed a more regular use pattern of her area (Figure 5-10) . Reproduction On 26 September 1990, a cub was seen at Km 4.5 of the Pogo Preto road (PPR) . His track width was 7 cm, and his weight was estimated at 30 kg (see Appendix C) . On 19 February 1991, an adult female with two large cubs were sighted on Km 3.7 of this same road. On 19 April, M13 was captured at Km 2.5, at an estimated age between 12/14 months. On 31 July, F17 was captured at the same site, her age estimated at about 8/9 years old. On 12 September, F21 was captured behind the hotel near the falls, and her age was about the same as that estimated for M13. The distribution of their locations strongly suggests that they formed a family unit, very likely the same animals seen in February. Assuming the age estimates for the subadults are correct, their birth would have occurred between February and April of 1990, and

PAGE 95

82 mating (110 days of the gestation period) from late October to December. When captured, F17 was in about mid-term pregnancy. On 7 October 1991, 68 days later, the mortality mode of her transmitter was activated but soon resumed to the active mode. As mentioned in the previous chapter, similar instances occurred with bobcat and ocelot females prior to giving birth. The mean distance between consecutive locations for this female increased from 2.3 ± 1.1 km to an asymptote of about 3.5 ± 2.5 km, between October 1991 and October 1992 (Figure 511) , likely a reflection of mobility of young. On 17 November 1992, M32 was captured at the same site as M13 and F17, with an estimated age of about 12 months. Further indication for the assumption that he was F17's son is provided by the fact that when this female was recaptured on 9 October (39 days earlier) , she showed signs of a light mange, which was also present on M32 (no other animals were found with this disease) . [Note: Due to a history of aggressiveness and a tendency to take domestic animals, in her old age, combined with bad physical condition when recaptured (both her upper canines had broken at the base and were badly infected) , this female was removed to a zoo] . M32 was also very lean, and was locally medicated for anemia and dehydration, before release at the capture site. M34 was captured on 14 August 1993, also at Km 2.5 of the P090 Preto road, at an estimated age of 10-12 months (he still

PAGE 96

83 had his milk dentition) . Therefore, his birth must have occurred between August and October 1992. This fact poses an interesting question since F17, then the established female, was still accompanied by M32. Thus, she could not have been the mother of M34. And, since she was removed from the Park on 9 October, the mother of M34 must have been established in the same area occupied by F17, which leads to further the assumption that she may be F17's daughter. Tracks of an adult female have been regularly found along the PPR, and one uncollared female was darted free-ranging, from a blind, on the night of 26 July 1993, when raiding a sheep corral just outside the Park. However, the drug was not injected and the animal escaped. Intensive efforts to trap this female have so far been fruitless. Judging by tracks found on the PPR during early 1995, there is another subadult animal using the area, likely a subsequent offspring of this same female. Dispersal Two subadult males, M13 and M32, were monitored during dispersal. After a few long-range movements (> 5 km) outside of his natal area, M13 crossed the Iguagu river into Iguazu National Park, in Argentina, on 24 September 1991. Assuming his birth was approximately in March 1990 (see section on Reproduction), his age would then be about 18 months. He remained in Argentina until 9 November, during which time he was located 5 times, with a mean linear distance between

PAGE 97

84 locations (MLDBL) of 2.4 ± 2.7 km. On 10 November, he was back in Iguaqiu, where he remained until 13 November, with a MLDBL of 2.0 ± 1.4 (0.6 3.8; N= 4) km. On 21 November, he was found again in Argentina, where he traversed the whole peninsula, and on 1 December, was found in Paraguay, having crossed the Parana river at a point where it is > 400 m wide. He was confined to a small island of forest, surrounded by farmland. On 17 December, he was back once again in his natal area in Igua^u, a linear distance of 16.2 km from his previous location. On 18 January 1992, he was located 33 km northeast of where he was located on 9 January. When he was recaptured on 13 March, to change his collar, he weighed 88 kg. He had several open cuts, which he likely incurred from intraspecif ic fights, and his upper right canine was broken off almost at the base. Between 18 January and 16 May, the mean distance between 13 locations was 12.9 ± 11.1 km (range: 1.0 33.0 km) . On 26 May, the mortality signal was received and I located the collar at a house by the Igua^u river, on the opposite side from the Park. After landing the helicopter at the house, I discovered M13 had been killed by the owner and his 15 year-old son while poaching white-lipped peccaries ( Tayassu pecari ) in the Park. The cumulative mean distance (± 1 SD; km) of locations to the arithmetic center of M13's natal area increased steadily and seemed to be levelling off when he was killed (Figure 512). The coefficient of variation (CV; Ott, 1984; see

PAGE 98

85 methods) , plotted on the axis of the figure, indicates the random nature of his movements. During his first 9 locations following capture on 17 November 1992, M32 covered an area of 5.4 km^, presumably still in his natal area. After 21 January 1993, the signal of his transmitter could not be found within the study area. By gradually increasing the area searched during flights, I found him on 22 February on the eastern sector of the Park, 64 km east of his last location. Assuming he was born in September 1991 (see Reproduction), his age then was 16 months. His dispersal may have been precipitated by the removal of his mother, F17, from the Park in October. His next 10 locations encompassed an area of 308 km^, with a mean linear distance between locations (MLDBL) of 9.0 ± 3.6 km (range: 4.4 16.7 km) . His movements for the next 10 locations decreased to a mean of 3.4 ± 1.2 km (range: 1.8 4.9 km), and were confined to an area of 17 km^. In his next 10 locations, he increased the area used, as a result of two long range movements (17.5 and 10.2 km) to an area where he was later seen twice (during recapture attempts) with an adult female. Therefore, he was already becoming an established adult, as indicated by the asymptote reached in the mean distance from the AC of his natal area and on the CV of his locations (Figure 5-13) . He was killed by poachers in June 1994. As a comparison to the dispersal patterns shown in Figures 5-6 and 5-7 for the Igua^u males, I applied the same

PAGE 99

86 method to a dispersing svibadult male from another study (Figure 5-14; data from Crawshaw and Quigley, 1991). The greater number of locations for this male shows the expected pattern of increasing mean distances and SD from the AC of the natal area until an asymptote is reached when the animal becomes established in an area. Conversely, the CV shows an initial increase in the dispersion of locations as the animal disperses, then gradually decreasing as it becomes attached to a specific area. Subadult male M34 would have provided another example of dispersal, but I lost contact with his transmitter while he was still in his natal area (see Mortality) . Mortality Three of the 7 radio-collared jaguars were killed by poachers within the limits of Igua^u National Park. Two males (M13 and M32, brothers of subsequent litters of F17) were shot by hunters waiting at blinds for other game species (peccaries, Tayassu pecari and T. taiacu . and paca. Agouti paca ) . The female, F21, was shot apparently as a trophy in a contracted hunt in the most densely visited part of the Park. Being habituated to humans, she regularly used a rustic shed of a concessionaire as an off -hours refuge, and was killed a short distance from it. An uncollared adult male (Appendix C, Table C-2) was blinded by a small-grain cartridge shot on September 1993, and

PAGE 100

87 was killed by a truck on the highway that borders the northeastern section of the Park. A large jaguar cub (c?iyca 20 kg) was killed by a vehicle on this same stretch of highway on 30 March 1995. In addition, an uncollared large male (Appendix C, Table C-2), captured as a problem-animal in a private property neighboring Iguazu NP and translocated to Urugua-i Provincial Park, to the south of Iguazu, was shot by locals within one week of his release. The transmitter signals of M33 and M34 disappeared prematurely, and they were not found, despite extensive aerial searches. It is possible that M34, a subadult male, dispersed out of our searching range. However, this is unlikely since very little undisturbed habitat remains outside the Park on the Brazilian side. All these areas and those in Argentina were thoroughly searched from the air. M33 was an old, established animal, and his transmitter had still another full year of battery life at the time he disappeared. The low probability of Telonics transmitters failure (see Material and Methods) and the history of high mortality of jaguar in the area suggests that these animals were likely killed and their collars destroyed. The overall daily survival (Sj) and mortality rates (m,) for the radio-tagged jaguars were 0.9788 and 0.0212, respectively (see Mortality section in the previous chapter, for explanation of method) . If we assume that both animals

PAGE 101

88 whose signals disapeared (M33 and M34) were poached, then the cumulative mortality curve reached 100% after 600 days of monitoring (Figure 5-15) . Density Despite heavy human disturbance within and around the study area, jaguar sign (mainly tracks) was frequently encountered along roads and trails. Six jaguars were captured and radio-tagged in the 80 km^ in a 4 -year period. Radiotelemetry data and indirect evidence indicate that at least one adult male and two adult females used the entire area. On 19 April 1990, an adult male other than the one radiocollared, was photographed by an infrared sensored remote camera. Since he was not captured, I do not know if it was a resident or transient animal. The estimated minimum density of adult animals, therefore, was 3.7 jaguar/100 km^. Assuming the adult females had, on average, 1.5 young per litter in alternate years, the total estimate for the area would be 6 animals in the 80 km^, or 7.5 jaguar/100 km^. If applied to the 1,750 km^ of Park, the total density would be approximately 64 adults, or 134 animals, including all ages. Diet A total of 362 carnivore scats was found in Iguagu, 73 (20.2%) of which were positively identified as having been produced by jaguar (Table 5-4) . Mammals comprised 80% of the

PAGE 102

89 106 items recorded (Figure 5-16), followed by birds (8.5%), reptiles (6.6%), fruit (3.8%), and invertebrates (1.9%). Mammals and reptiles accounted for all food items taken in the summer, the diversity of items of items increasing gradually from fall through spring (Figure 5-17) . Considering only the mammals (Figure 5-18) , peccaries were the prey taken most often (45% of all items) , followed by opossum, armadillo, and deer, to mention only those occurring above 10% in the sample. However, if the different food items are corrected by weight, the order of relative importance of these species changes considerably (Figure 5-19) . Peccaries become even more prominent (77%) , followed by deer (14%) , coati (2.2%), armadillo (2.0%), agouti (1.1%), and opossum (1.0%). Therefore, these species alone account for over 97% of the diet derived from the jaguar scat sample. It is noteworthy that these same species are among the most abundant in the study area, judging from my sighting data (Appendix B) . However, the proportion in which they are taken indicates different levels of selectivity by jaguar (Figure 5-20) . Peccaries were taken much more often than the relative abundance estimate would predict, whereas agoutis were taken much less. Deer were the only item taken in egual proportion to its relative abundance. Relationships with Humans

PAGE 103

90 During the period of the study there was a sharp increase in the nvimber of instances involving predation or other problems concerning predators, particularly jaguar, reported to Park authorities. This was likely a result of media releases about the project, which aroused interest at both local and national levels. During the latter two years of the study, I personally attended 10 complaints from neighboring ranchers and residents. Some of these cases involved considerable economic losses. On one occasion, a jaguar killed 22 sheep in a 10-day period; on another, 36 head of cattle (mostly calves 1 to 1^ years old, but also some adult cows) , were killed in a 3-4 month period. Estimated losses in these cases were roughly US$1,100 and US$5,400, respectively. Most people attributed the increase of instances to an increase in the population of jaguars in the Park. However, the fact that two subadult males (M13 and M32) could establish adult home ranges within the Park indicates that there were still open spaces in the population. Other hypotheses can be postulated. Prior to the project, there was little response from Park and wildlife authorities to predation complaints, for lack of technical information and expertise on the subject. Therefore, there was an unspoken policy between ranchers of resolving the situation by locally eliminating the problem-animal. With the knowledge about the project, and of the critical status of the species, some of the more conscientious ranchers tended to report losses rather than

PAGE 104

91 eliminate the animal. The widely publicized arrest of two poachers that killed radio-equipped jaguars also must have acted as a deterrent, with a widespread belief that all jaguars in the Park were being radio-monitored. Several of the radio-tagged jaguars had their ecology and behavior affected by the close contact with humans. On September 1991, I received a complaint from the manager of the hotel at the falls that a jaguar had been approaching the swimming pool area in broad daylight, in spite of the large number of guests and employees. Her behavior, however, showed no sign of aggression. Attempts to scare the animal away with fire-crackers had been unsuccessful. Two nights after setting a large trap baited with a live chicken, at less than 150 m from the swimming pool, a subadult female (F21) was captured. Upon examination, I discovered that her upper right premolar tooth of the milk dentition had been retained, causing infection and swelling of the giam that extended about two thirds of the length of her canine. This apparently prevented her from preying on natural prey, and led to a dependency on food remains from the hotel restaurant, at the garbage dump. With the help of a veterinary, the tooth was extracted and she was treated at the site. She was then released at a site about 4 km away. After that, she did return to the hotel area occasionally, but was not as visible as before, only her radio signal revealing her presence.

PAGE 105

Table 5-1 — Monitoring period of 7 jaguars in Iguagu and Iguazu National Peirks, Brazil and Argentina. Letter with the animal number indicates sex. N= number of locations; c= capture; R= recapture (a number indicates number of recaptures during that month); K= animal killed; L= lost contact. AN* V PERIOD OF MONITORING 1991 1992 1993 1994 JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND F17 69 C R» F21 25 C~K MI3 53 C Rb-K 1132 36 C K M33 17 C2R R-2RL M34 21 C L M48 15 C R' ' — Animals removed from Park (see text). ^ — Collar changed.

PAGE 106

93 Table 5-2~Radio-telemetry information on :aguars ^^^^^^^^
PAGE 107

94 Table 5-3 — comparison of home range estimates (km^) for 5 jaguars in Iguagu and iguazu National Parks, derived from conventional methods and the proposed Circle method. N= number of locations (no outliers were subjectively excluded); MCP= Minimum Convex Polygon method; HM= Harmonic Mean method; ELL= 95% Ellipse method; CM= Circle method (95% of the locations (MLDAC + 2 SD) . AN# N HOP HM ELL CM F17 69 70.0 164.0 113.9 149.6 F21 25 8.8 ? 15.9 6.6 M13* 10 22.6 77.8 131.1 81.7 Mis'* 11 138.6 583 577 123.0 M32'' 9 5.4 35.3 33.9 75.4 M32'' 6 104.0 ? 1,447 98.5 M33 17 25.7 75.7 80.0 78.5 M34 21 16.4 66.7 44.0 60.8 M48 15 86.5 238 378 172.0 • — Natal ''--Adult area home range ( incomplete; see Dispersal)

PAGE 108

95 Table 5-4~Food items found in 73 jaguar scats in Iguagu National Park, Brazil. WGT= mean live weight of prey item (kg); N= number of scats the item was present; #IND= number of individuals per scat; % scats= percentage of scats containing that food item; % items= percentage of total number of items found; B^^^= estimated total weight of that prey item ( kg ) . PWPY TTEM WGT N #IND ITEMS SCATS Peccary 30.0 Jo 1 Q 1 J3 . 0 /Tavassu SD.\ Opossum 1.50 1 A 10 1 Q yl J. ^ (Didelphis aurita) Armadillo Q 1 8 5 12 . 3 30 fDasypus novemcinctus ) Deer 22.5 9 1 8.5 12.3 203 (Mazama sp. ) Bird 0.65 9 1 8.5 12,3 5.9 (Cracidae, Tinamidae) Lizard 1.60 6 1 5.7 8.2 9.6 (Tupinambis tequixin) Coati 5.50 6 1 5.7 8.2 33.0 (Nasua nasua) Agouti 3.2 5 1 4.7 6.8 16.0 (Dasvprocta azarae) Unidentified fruit — 4 3.8 5.5 Scruirrel 0.80 2 1 1.9 2.7 1.6 (Sciurus aestuans) Invertebrates 2 1 1.9 2.7 ReQjbit 0.8 1 1 0.9 1.4 0.8 (Svlvilaqus brasiliensis ) Paca 10 1 0.9 1.4 10.0 Capuchin monkey 3.0 1 0.9 1.4 3.0 (Cebus apella) Tayra 5.5 1 0.9 1.4 5.5 (Eira barbara) Margay 3.0 1 0.9 1.4 3.0 (Felis wiedii) Unidentified snake 1.00 1 0.9 1.4 1.0 TOTAL 106 99.7 145.1= 1,478 ' Average of the weights of the two species (M. americana and M. nana; Does not add to 100% because of more than 1 item per scat.

PAGE 109

96 Figure 5-1 — Radio-locations (•) and capture locations () of an adult male jaguar (M33) , Iguagu National Park, Brazil. Animal was recaptured a total of 9 times at the two sites of the PPR. Monitoring period: 11/08/93 29/05/94.

PAGE 110

97 Figure 5-2 — Radio-locations (•) and capture locations () of an adult female jaguar (F17) , Igua^u National Park, Brazil. Two captures at same site. Monitoring period: 31/07/91 09/10/92.

PAGE 111

98 Figure 5-3 — Radio-locations (•) and capture locations () of a subadult female jaguar (F21) , Iguagu National Park, Brazil. Monitoring period: 12/09/91 13/12/91.

PAGE 112

99 Figure 5-4 — Radio-locations (•) and capture locations () of a subadult male jaguar (M13) , Iguagu National Park, Brazil. Note locations in Argentina and Paraguay. Monitoring period: 19/04/91 26/05/92.

PAGE 113

100 Figure 5-5 — Radio-locations (•) and capture locations () of a subadult male jaguar (M32) , Iguagu National Park, Brazil. Monitoring period: 17/11/92 11/06/94.

PAGE 114

Figure 5-6 — Radio-locations (•) and capture locations () of a subadult male jaguar (M34) , presumably still in natal area, Igua?u National Park, Brazil. Monitoring period: 14/08/93 21/09/94.

PAGE 115

i 102 Area (Sq. Km) 140| Number of Locations — ^F17 -4-F21 -**-M33 -d-M34 -Am48 Figure 5-7 — Cumulative area curve (MCP) for 2 female (F17, adult, and F21, subadult) and 3 male (M33, M48, adults, and M34, subadult) jaguars in Iguagu and Iguazu National Parks, Brazil and Argentina.

PAGE 116

Figure 5-8 — Ciomulative area curves (MCP method) for two subadult males in Igua^u National Park, Brazil.

PAGE 117

104 Figure 5-9 — Comparison of the mean linear distances between consecutive locations of males and females at increasing intervals (days) .

PAGE 118

105 PERCENT OF LOCATIONS 80 -I ADM ADF SAM SAP (N= 61) (N^ 69) (N= 82) (N: 25) AGE/SEX Distance (km) ^1-2 W2-3 ^3-4 E ^ >4 Figure 5-10 — Percent of locations at increasing distances from the arithmetic center for different age/sex classes for 7 jaguars in Igua9u and Iguazu National Parks, Brazil and Argentina. N= number of locations.

PAGE 119

106 KILOMETERS 7 I — 6 _ T " 9 4 3 2 1 1 q\ Oct/Nov Oct/Jan Oct/Mar Oot/May Oct/ Jul Oct/Oct (N10) (N16) (N27) (N29) (N37) (N49) Figure 5-11 — Cumulative mean distances (± 1 SD; km) between consecutive locations for an adult female jaguar (F17) after the presumed birth of a litter, in October 1991, in Iguagu National Park, Brazil.

PAGE 120

107 35 KM PERCENT X ^ ^ I I -L. 20 3D 40 # OF LOCATIONS 50 -t Mean +1 SD CV 140 120 100 80 60 40 20 60 Figure 5-12 — Cumulative mean linear distance of locations to the AC of the natal area of male M13, in Iguagu National Park, Brazil. The coefficient of variation (CV, in %) is plotted on the Yj axis.

PAGE 121

108 80 KM 10 15 20 25 30 tt OF LOCATIONS -t Mean +1 SD CV PERCENT 35 250 200 150 100 50 40 Figure 5-13 — Cumulative mean distance of locations (± 1 SD, in km) to the AC of the natal area of jaguar M32, in Igua^u National Park, Brazil. The coefficient of variation (CV, in %) is plotted on the axis.

PAGE 122

109 AO 50 60 70 # OF LOCATIONS -iMean +I SD CV PERCENT 100 120 100 80 60 40 20 0 10 Figure 5-14 — Cimulative mean distances (± 1 SD; km) of locations to the AC of the natal area of male jaguar M04, in Miranda, Pantanal of Mato Grosso, Brazil; CV plotted on the Yj axis (data from Crawshaw and Quigley, 1991) .

PAGE 123

110 100 80 60 40 20 0 Percent Mortality 1 0 60 120 180 240 300 360 420 480 540 600 Days Cumulative Mortality Poaching Removed Un k nown Figure 5-15 — Mortality of radio-tagged jaguar (N= 7) in Iguagu and Iguazu National Parks, Brazil and Argentina. Animals in the unknown category were presumed killed (see section on Mortality) .

PAGE 124

Ill Figure 5-16 — Percentage of the different types of food items (N= 106) found in 73 jaguar scats in Iguaqju National Park, Brazil.

PAGE 125

112 Figure 5-17 — Percentage of the different types of food items taken by season in 73 jaguar scats from Igua^u National Park, Brazil.

PAGE 126

113 Figure 5-18 — Percentage of the different species of mammals found in 73 jaguar scats in Iguagu National Park, Brazil.

PAGE 127

114 Figure 5-19 — Percentage of the corrected estimated biomass of the different items found in 73 jaguar scats from Iguagu National Park, Brazil.

PAGE 128

115 PERCENT 80 I Peccary Deer Coati Armadillo Agouti Opossum PREY Scats Relative Abundance Figure 5-20 — Percentage of corrected prey taken by jaguar in Igua^u National Park, compared with relative abundance derived from sighting data (Appendix B) .

PAGE 129

CHAPTER 6 COMPARATIVE ECOLOGY OF OCELOT AND JAGUAR, WITH IMPLICATIONS FOR MANAGEMENT AND CONSERVATION In this chapter I will discuss differences and similarities between ocelots and jaguars in Iguagu and among various study sites, with a final discussion of the human impact on felid management and prospects for jaguar and ocelot conservation, including cultural differences in the way both local people and tourists perceive coexistence with these two predators . As pointed out by Emmons (1987) based on her studies of ocelot and jaguar in Peru, most of the ecological differences between these species relate directly or indirectly to body size and food availability. In Iguagu, mean jaguar weight is about eight times greater than that of ocelots (80 kg and 11 kg, respectively) . Home ranges of ocelots in Igua^u were considerably larger than those reported from other areas (Table 6-1) . In four studies (Ludlow and Sunquist, 1987; Emmons, 1988; Tewes, 1986; Laack, 1991) , the average home range size for adult males was 8.6 ± 3.2 km^ (range: 5.7 12.3 km^) and for adult females, 3.8 ± 2.2 km^ (range: 1.8 7.0 km^; Table 6-1). Only in Belize (Konecny, 1989) were home range sizes comparable to 116

PAGE 130

117 those in Igua?u, although sample sizes in Belize were limited to one male and one female. Given the smaller home range sizes in the Venezuela and Peru studies, it is not surprising that estimated densities for these sites were much higher than that estimated for Iguagu (2.8and 6-fold differences, respectively) . Apart from differences in the methods used among studies, I believe that the variation shown is related to edge/habitat effects, the availability of standing water, and the resulting variation in densities of the main prey species. The study sites in the llanos of Venezuela, the Pantanal, the floodplain forest in Manu, and Laguna Atascosa in Texas all seem to have permanent sources of water and associated forest cover. In Igua^u, there was a network of small water courses that were dry during much of the year, with permanent water restricted to a few larger rivers. The movements of some study animals in Igua^u consistently followed some of these water courses. Furthermore, the homogeneous habitat and reduced availability of water may have resulted in a lower overall density of the main prey species (or decreased vulnerability) . This, in turn, may have changed ocelot spacing patterns, resulting in larger individual areas with a high degree of home range overlap. However, if I use only the 50% HR (Table 4-3, mean of 14.9 ± 9.1 km^, range: 1.5 32.1 km^; N= 12), there was very little overlap, indicating that areas toward the center of ocelot home ranges were more exclusive.

PAGE 131

118 Some study animals showed a reduction in home range size, which may also be an indication of competition for space (and access to food) in Igua^u. The home ranges of two female ocelots, F06 and F19, the longest-monitored females, showed reductions in size of 38% and 25%, respectively. The reduction for F06 was likely due to her removal from the Park for a period of four months for treatment in captivity and, when she was released back in the Park, she was not able to acquire the full extent of her former range. The reduction in size of F19's area was probably associated with her advanced age and that part of her range was acquired by her presumed daughter, F42 (see Figures 4-12 and 4-13) . A similar situation may have occurred with F03 and F39, and F09 and F40, both presumed mother-daughter pairs (Figures 4-7 and 4-10, and 4-9 and 4-11, respectively) . By contrast, Laack (1991) noted no reduction in area used by five ocelots (two males and 3 females) monitored for extended periods of time in Texas. Reduction in range size of ocelots in Venezuela was temporary, being smaller during the wet season (Ludlow and Sunquist, 1987) . The large ocelot home ranges in Iguagiu do not appear to be unusual, since several other species of carnivores in the Park also had relatively large home ranges. One 3.5 kg adult male margay (Felis wiedii; M15) occupied an area of 15.9 km^ during an 18-month monitoring period (Appendix D) . A female jaguarundi (F. vagouaroundi ; F18) in Iguazu, Argentina, used

PAGE 132

119 an area of 6.8 km^ during the 18 days she was monitored (8 locations), and a subadult male (M47) in Iguagu, Brazil, used 17.7 km^ during a 4-month period (17 locations. Appendix D) . One adult male crab-eating fox ( Dusicvon thous. Mil) ranged over an area of 6.7 km^ in a 6-month period (28 locations). Another adult male fox used intensively an area of 3.4 km^, near the Park's entrance, but made three separate trips to the falls area by the hotel, increasing his total area to 15.2 km^ (7-month monitoring period, 91 locations) . In his trips, he appeared to follow the main highway; he was killed by a vehicle in September 1991. One adult male coati (M24) used an area of 6.2 km^ in a 7-month period (21 locations), and one pregnant female used 1.1 km^ in a 1-month period (9 locations) . As far as I could tell, both of these animals were traveling alone during the monitoring period although they may have associated with bands occasionally. Large ranges for some of these species were also reported by Konecny (1989) . At his study site in Belize, the range of one male margay was 11 km^; two male jaguarundis had ranges of 88 and 100 km^, and one female 13 km^. By contrast, the mean home range for three pairs of foxes in the Venezuelan llanos was only 0.74 km^ (range 0.45 1.0 km^; Sunquist et al. 1989) . Brady (1979) , at the same study site, reported ranges of 0.96, 0.6, and 0.54 km^ for another three pairs of foxes. These differences in home range size once again point out a

PAGE 133

120 greater productivity and consequent prey availability of the llanos, as compared to Iguagu and Belize. Evidence of competition for space in Iguagu was also supported by intraspecif ic disputes such as that involving the dispersal of F23 (see Dispersal) and the one resulting in the death of adult male M05 (see Mortality) . Tewes (1986) reported one instance of an aggressive encounter between two male ocelots, and Laack and Tewes (1988) mention one young male that was likely killed by an adult male. Emmons (1988) also mentions the death of one male ocelot in an intraspecif ic fight and the probable displacement of an old male by another male. She also recorded one antagonistic encounter between one adult and one subadult female and, subsequently, the displacement of this same adult female from her home range due to deteriorating condition related to old age. The distances covered and time taken by dispersing ocelots until establishment of adult home ranges are likely a function of available space in a given population. In Igua^u, dispersing subadults (M37, M41, F02, F23) were unable to settle in an area and become residents within the periods they were monitored (14, 11, 11, and 16 months, respectively). Furthermore, some adult males seemed to be "drifters", roaming within the ranges of probably related animals, as suggested by the extensive overlap between the ranges of M26 and M27 and the known range of M04 (Figures 4-3, 4-4, and 4-1, respectively) . Since the transmitter of F03 failed early in

PAGE 134

121 the study and I did not capture the other resident female in the area these males occupied, I do not know how these animals associated. Dispersal was also recorded for two subadult female ocelots (F02 and F23) in Igua?u. F23 was killed by a poacher before establishing a home range, and the fate of F02 is unknown because of transmitter failure. Laack (1991) also reported female dispersal in four instances. In her study, the distances between natal areas and established home ranges were slightly shorter for females than for males (8.0 and 10.8 km, respectively) . Results on the dispersal of ocelots and jaguar in Iguagu seem to indicate some important differences. Departure from the natal area appears to be more drastic and faster for jaguars (Quigley and Crawshaw, in prep.; this study); only in one instance did M13 return briefly to his natal range before heading east in the Park. Ocelot dispersal, on the other hand, was more extended and involved several exploratory movements after which animals returned to the natal area for varying lengths of time, before again leaving. Laack (1991) also mentions extended dispersal periods (of up to 8.5 moths) before some s\ibadult ocelots became established. Dispersing ocelots appear to use corridors of dense cover, avoiding open areas (Navarro, 1985; Tewes, 1986; Ludlow and Sunquist, 1987; Laack, 1991). Similarly, in this study ocelots also followed forest corridors during dispersal.

PAGE 135

122 Beier (1995) also showed the importance of corridors for dispersing pvimas in California. The probability of securing an adult home range within the Park appeared to be greater for jaguars than for ocelots. Following initial dispersal, ocelots M37 and M41 crossed back and forth between Iguagu and Iguazu, and had not become siteattached by the end of the study, after periods of 6 and 9 months, respectively. It is possible that the great extent of overlap between the ranges of M26, M27, and M04 (assuming the latter was still occupying his known range after he lost his transmitter) further indicate a difficulty in finding an unoccupied area in the population. This apparent difficulty could be a reflection of saturation density associated with a low turnover rate in the population. That some subadult female ocelots also dispersed (F02, F23) lends support to this assumption that the habitat was at carrying capacity. Malebiased dispersal is expected in polygynous species with malemale competition for females (Greenwood, 1980; Cockburn, 1992) . On the other hand, two of the subadult jaguars, M13 and M32, were becoming attached to specific areas within the Park when they were killed by poachers. The apparently greater probability for subadult jaguar to secure adult ranges within the Park could be a result of a higher turnover rate of the population due to high mortality. Mortality rates of ocelots and jaguars are guite different in Igua9u (Figure 6-1; data from Figures 4-25 and 5-

PAGE 136

123 15) . Five out of 21 ocelots (24%) died during the study period. Two of these were killed during dispersal (M36 and F23), 2 were roadkilled (F16 and F39) , and 1 was apparently killed in a intraspecif ic dispute. By contrast, I had lost all of my 7 radio-tagged jaguars in a 3 -year period. Three were poached inside the Park (M13, M32, and F21) , 2 had to be removed from the Park because of livestock depredation habits (F17 and M48; the latter had already been wounded), and 2 disappeared prematurely and were suspected of having been killed (M33 and M34) . In addition, 3 other, uncollared jaguars were known to have been killed during the study period, one for livestock depredation outside the Argentine park, one shot by a poacher, and one cub killed by a vehicle at the Park boundary. It is uncertain if the reliance on alternative food sources, such as baits in traps and garbage dumps, was related to competition but a large number of animals were captured or located near these sources. Ocelots were captured at garbage dumps 15 times in Iguagu, as were 10 coatis and 2 foxes. Another 2 ocelots, 4 foxes and 3 jaguar were captured as they attempted predation on livestock at human settlements. A conspicuous example of this reliance on alternative food sources in Igua^u is the great numbers of coatis concentrated in the falls area, where they beg for food from tourists. There was also interspecific predation among carnivores in Iguagu. Two tayras, two jaguarundis, and one fox (all

PAGE 137

124 radio-tagged) were killed by predators. One of the jaguarundis (F18) was apparently killed by a pxma. Evidence from scats also showed predation on grison ( Galictis cuja) , tayra ( Eira barbara) , and raccoon ( Procyon cancrivorus ) . Differences in the levels of daytime and nighttime activity of ocelots and jaguars in Iguaqru and at other study sites (Table 6-1) are probably related to activity patterns of the main prey species, as suggested by other authors (Ludlow and Sunquist, 1987; Emmons, 1987, 1988; Crawshaw and Quigley, 1989) . The high level of daytime activity displayed by the two felid species in Iguaqu can probably be explained by the prominence of the diurnal agouti and tegu lizard in their diets (Figure 4-27 and 5-19) . In an attempt to standardize the food habits data, I compared the mean prey-predator ratios (MPPR) for ocelot and jaguar, and considered each item found in the scat samples (Tables 4-6 and 5-4) as one instance of predation. Using a mean weight of 11 kg for ocelot and 80 kg for jaguar, I employed the following formula: MPPR= S(prey weight/predator weight x N)/2(N) Because of discrepancies between the results from the mean weight of vertebrate prey (MWVP; Iriarte, 1988) calculations obtained for the same data sets by different authors (Jorgenson and Redford, 1993; Oliveira, 1994), I derived the mean prey weight (MPW) from the formula above, by multiplying the mean prey-predator ratio by the mean weight of

PAGE 138

125 the predator. I also recalculated MPW for the fore-mentioned ocelot and jaguar studies (Table 6-2) . The resulting prey-predator ratio for ocelots was 0.13, indicating that in Iguagu the MPW for ocelots was 1.4 kg. Data on feeding habits of ocelots in Belize show the same trend to those from Igua?u. The calculated MPPR for that study was 0.24, with a MPW of 2.4 kg (Konecny, 1989). By contrast, the MPPR for ocelots in Venezuela was 0.05, with a resulting MPW of 0.52 kg (Ludlow and Sunquist, 1987). Similarly, in Manu National Park, Peru (Emmons, 1987), the MPPR was 0.08 and the MPW was 0.61 kg. These differences likely reflect ocelot adaptability to site-specific differences in prey density and vulnerability. Food habits data for jaguar show different trends (Table 6-2). In Igua^u, the MPPR was 0.18 with a resulting MPW of 14.4 kg, which is closer to the results from Manu (Emmons, 1987) . However, in Emmons' study large chelonians were an important component in the jaguar's diet. This likely reflects the adaptability of the species to using this abundant resource in the Amazonian floodplain. In Belize, jaguars were preying on smaller prey (MPW= 6.7 kg), which again was the most available resource (Rabinowitz and Nottingham, 1986) . Densities of the larger species such as peccaries, tapir and deer were likely reduced by subsistence hunting by humans in the Belize study area (Rabinowitz, 1986) . The MPW from these three studies, all in continuous-forest

PAGE 139

126 habitats, show a sharp contrast to those from the Pantanal of Brazil and the llanos of Venezuela (Crawshaw and Quigley, 1984; in prep.; Hoogesteijn et al. 1993), where cattle was also available (Table 6-2). Even if only native prey are included, the MPW (28.5 kg) of jaguars in the Pantanal is still almost two-fold that for Igua
PAGE 140

127 prey, of which 81,000 kg would be presumably provided by peccaries alone. My information on peccaries is scanty. Because of the much larger group size, white-lipped peccaries tend do be more conspicuous than collared peccaries. Furthermore, their habit of raiding crops adjacent to the Park, lead people to grossly overestimate their density. My density estimate for whitelipped peccaries in the study area was about 70 individuals (in one herd that apparently split in subgroups occasionally) in the 80 km^, or 0.9 individual/km^. However, I knew of the recent killing of at least 8 members of this herd, as they invaded crops at a neighboring property. Fragoso (1994) found home range sizes of 21.8 and 109.6 km^ in Maraca island, Amazonia, with densities ranging from 1.2 to 8.1 peccaries/km^. Assuming a conservative estimate of 1.5 peccaries/km^ for the 1,750 km^ park, the resulting estimate would be 2,625 animals, or a total biomass of 78,750 kg (weight estimated as 30 kg, as opposed to 35 kg of adults, to account for young and subadult animals) . Collared peccary densities ranged from 1.0 to 3.3/km^ in Maraca (Fragoso, 1994) . For Manu National Park, in the Peruvian Amazon, Emmons (1987) estimated a density of 5.6 individuals/km^ for the two species combined. Based on my sighting data in Iguagu, collared peccaries were either scarcer than white-lipped or less conspicuous, or both (36% for collared peccaries and 64% for white-lipped; N= 33

PAGE 141

128 sightings) . Mean number of individuals per sighting was 2 for collared and 12 for white-lipped. Using this sighting ratio' to estimate collared peccary density, as compared to whitelipped, the total estimate for the Park would be roughly 1,575 animals, or a biomass of 34,650 kg (mean weight of 22 kg). Therefore, the estimated combined adult biomass of the two species would be approximately 113,400 kg for the entire Park, or about 65 kg/km^. To this figure should be added the potential yearly productivity, estimated at about 50,750 kg, for a total of 164,150 kg. The 81,000 kg required by the total jaguar population would thus comprise roughly half of the available peccary biomass. This level of predation would be unsustainable per se, even in the short run. However, to this figure should be added the competition from poachers, that hunt selectively for these two species also (Becker, 1981; Jorgenson and Redford, 1993; Townsend, 1995), and the control exerted by neighboring farmers when peccaries raid their crops. Given the pronounced selection for white-lipped peccaries by jaguars (Crawshaw and Quigley, 1984; Fragoso, 1994; this study. Figure 5-20) and even allowing for some error in my estimates for this species in Igua^u, my prediction is that they will not withstand current predation levels (both from jaguar and from humans) , and will become extinct in the Park within the foreseeable future. In this event, jaguars, as an adaptable species, will likely gradually

PAGE 142

129 switch to smaller, abundant prey, as was probably the case in Belize (Rabinowitz and Nottingham, 1986) . In an attempt to provide a better understanding of the relationship between the larger species of f el ids in Igua^u, I included in this analysis a sample of 28 puma scats (7.7% of the 362 scats found) . Therefore, I had a total sample of 157 scats for these three species, thus distributed: jaguar= 73 (46.5%), ocelot= 56 (35.7%), and puma= 28 (17.8%). To measure the relative importance of each main prey item within and among these three species of f el ids, I used two approaches. For among-species comparison, I considered the percentage of the total estimated weight of each prey species (estimated weight times frequency of ocurrence in all scats) in the scat samples for each of the predator species (Table 63, Figure 6-2). For within-species comparison, I calculated the percentage of the total weight of each prey species as a function of the total weight of food items for each species of predator (corrected by estimated biomass of each prey species for each predator; Figure 6-3) . Only those items that occurred more than once in the samples were included. The mean prey-predator ratio for puma was 0.27, with a mean prey weight of 10.8 kg (see Table 6-2 for ocelot and jaguar values) . Interestingly, the MPPR for pumas in Iguagu is higher than that for jaguar, even though MPW is smaller. This is due to the smaller weight of pumas (mean weight= 40 kg) . Puma diet included all prey items taken by ocelot and

PAGE 143

130 jaguar, whereas small rodents were absent from jaguar scats and ocelots scats did not show peccary nor coati remains. On the other hand, armadillo, agouti, opossum, and lizard were present in the scat samples of all three felids (Table 6-3, Figure 6-2) . The relative importance of each prey species corrected by biomass showed extensive overlap between the diets of jaguar and puma (Figure 6-3) . Estimated biomass of peccaries and deer together comprised 93% of the jaguar diet and 82% of the puma scat sample. Armadillos brings the total up to 91% for pumas. Ocelots, by contrast, concentrated on armadillo, agouti, and opossum (total of 91.6%). Because no puma was radio-monitored in this study, I do not know the spatial relationships among these three species of felids. Pumas are apparently less prone than ocelots and jaguars to entering live-traps (McBride, 1976; this study). Only one subadult male puma was captured (Appendix C) . Overall, pumas were sighted more frequently than jaguars in the study area (Appendix B) but left less indirect signs of their presence, in the form of tracks and scats. Sighting data indicated a resident population in the study area. One large male was sighted at least three times, one adult female with small young was sighted on the PPR in October 1990, and a subadult was photographed in the same area in March 1991 (same animals seen with the female?) . From these data, there was no evidence for spatial or temporal separation between

PAGE 144

131 puma and jaguar in Igua^u, as suggested by Schaller and Crawshaw (1980) in the Pantanal of Mato Grosso, and by Emmons (1987) in the Peruvian Amazon. It is interesting to note that during May-June 1995 there was a dramatic increase of the frequency of puma sign on the PPR, despite the presence of an uncollared adult female jaguar and one or two subadult animals in the same area. A total of 57 puma scrapes was recorded in a one-month period, 3 of which were associated with feces. No puma scrapes had been recorded previously in the area. One wonders if this more conspicuous presence of puma in the study area was related to a decreased jaguar density, since the incidence in marking coincided with the demise of M33 and M34. Continued research in Igua?u will hopefully reveal further information on the relationships between these species. Management and Conservation When one first looks at Igua^u from the air and realizes its condition as an island (Plate I, Chapter 1), thoughts of inbreeding depression and genetic deterioration (Ralls and Ballou, 1983; Ralls et al., 1986; Lacy, 1992) come to mind, especially applied to species such as the jaguar. However, data presented herein suggest this may not be a serious consideration in jaguar conservation in Igua^u, as long as a corridor with undisturbed areas is maintained in Argentina. Dispersing subadults of both jaguar and ocelot can, and do.

PAGE 145

132 eventually cross between countries and ensure gene flow between populations. Some 12,000 km^ of subtropical forest remain in the province of Missiones, in Argentina (J. C. Chebez, pers. comm.)A total of 4,140 km^ (34.5%) of this area is protected to some degree, forming an almost continuous corridor, about 200 km in length, linking Iguagu NP, in Brazil, to Turvo State Park in the state of Rio Grande do Sul, also in Brazil (Figure 6-4; Ministerio de Ecologia y Recursos Naturales Renovables, 1993). Turvo State Park protects the last known population of jaguar in that state, and represents the southern limit of the species in Brazil (Crawshaw and Pilla, 1994) . Jaguar populations in the province of Missiones, Argentina, may well represent the southern limit of the species in South America. All aspects considered, in the long run ocelots have a much better chance of surviving in Igua9u than jaguars. Ocelots use smaller areas (and, hence, occur at a higher density) and rely on more abundant and productive food sources and, more important, humans are more prone to accept living in close contact with them than they are with jaguars. Even though an alternative, staple, prey base and the habitat for jaguar may be preserved, problems with their coexistence with humans will likely tend to escalate if present conditions persist. Direct competition from poachers taking its main prey (peccaries, deer, and paca) will tend to increase the

PAGE 146

133 frequency of livestock depredation around the Park. Jorgenson and Redford (1993) showed extensive overlap in the prey taken by big cats and man in several areas in the Neotropics, which resulted in the decline of populations of the former. Results from Townsend (1995) for lowland Bolivia also show the negative impact of human hunting on the same species of prey taken by jaguar in Igua^u. The large areas occupied by individuals, the narrow shape of the Park, and the lack of effective deterrents will always lead jaguars to the periphery in search of food outside it. Once habituated to feeding on livestock, it is difficult to stop them. This is especially critical for females with cubs, which tend to pass, through learning, this behavior to their offspring (Ewer, 1973; Crawshaw and Quigley, 1984). The considerable economic impact that this livestock depredation imposes (see page 92) is amplified for low-income land-owners, and it is unreasonable to expect these small farmers and ranchers to accept jaguar predation without retaliation. Indeed, jaguar killing for livestock depredation control is presently one of the main mortality factors for the species throughout its remaining range (Rabinowitz, 1986; Crawshaw and Quigley, 1991; Quigley and Crawshaw, 1992; Hoogesteijn et al.l993). In this study, two adult males from Argentina were shot because of livestock depredation. In addition to animals killed by ranch owners, an unknown nximber of jaguars will continue to be killed by poachers

PAGE 147

! . ] i 134 hunting for other species in the Park. Such was the case with the two radio-tagged subadult males, M13 and M32. These mortality factors, although also affecting ocelots (such as F23) , are not as critical as they are for jaguar. The picture from which I gleaned some hard-won facts on the ecology of jaguar in the present study is not a very encouraging one. There were problems with almost every individual that was radio-monitored: cats that were shot because they killed animals outside the Park, or because they came too close to a poacher who was killing their prey inside the Park; animals that became dependent on food scraps from humans; and animals that had to be removed because of acquired bad habits towards humans or their property. However, given the sheer size of Igua^u and the fact that the habitat and some very productive prey species will remain, I would like to believe that jaguars just may have a chance to endure there. For this to succeed, three main conditions have to be implemented shortly: a) completing and maintaining a perimeter road around the entire perimeter of the Park, which would facilitate item "b" below; b) a more efficient control of poaching by Park police, and the maintenance of a fence (combining conventional wire-netting and electric) that would decrease the chances of animals (both carnivores and herbivores) from searching for food outside of Park boundaries.

PAGE 148

135 c) helping to secure the presence of a permanent corridor with undisturbed habitat in the province of Missiones, Argentina (Figure 6-4) ; In March 1994, a workshop took place in Foz do Iguagu to discuss an emergency management plan for the Park. Some 80 people participated, representing all governmental and nongovernmental agencies which, directly or indirectly, related to the use, protection, management, research, and conservation of the Park. Among other important issues discussed, the three fore-mentioned management recommendations were considered priorities. Funds to implement the resulting recommendations are secured through the National Program for the Environment, managed by IBAMA, Brasilia, aided by outside consultants . Reinforced by this project and a similar study in Turvo State Park (Crawshaw and Pilla, 1994; Figure 6-4), a close tie between IBAMA, Argentine environmental authorities and some NGOs (such as WWF-US and Sociedade de Preservagao da Vida Silvestre SPVS) is beginning to develop. These organizations are working towards the maintenance of a corridor of subtropical forest in Missiones, linking Iguagu and the Pargue Estadual do Turvo, in Brazil (WWF-US, 1991). A small step further in cross-country colaboration may allow for a metapopulation approach (Gilpin, 1987; Guerrant, Jr., 1992; Padua, 1993; Forys, 1995) for the management and conservation of jaguar populations in the area.

PAGE 149

136 As pointed out by Mech (1995) for the wolf recovery program in the United States, I also consider essential the maintenance of a permanent local program of environmental education. This program should emphasize the importance of Iguagu not only for the conservation of natural resources, including the large predators, but also to the local economy, in terms of tourism. Through this program, close contact should be kept with neighboring land-owners, to help solve problems when they arise. One of the objectives of the recently created National Predator Center (IBAMA, 1994) is to promote some form of compensation for losses inflicted by predators. This compensation may be accomplished by cuts in federal, state, or municipal taxes, or by direct reimbursement using funds raised by NGOs. A similar attempt is underway in northwestern Argentina (Perovic, 1993) . As a consequence of the reported jaguar depredation problems in the vicinity of Iguazu National Park and other areas, a managementconservation action plan is beginning to be implemented in the province of Missiones (Chebez, 1995) . Ultimately, the fate of the jaguar in Iguagu and, for that matter, of any large predators constrained within relatively small, isolated parks and reserves, remain in the hands of the people that live around these areas and coexist with these species. However, it is the responsibility of managing agencies to resolve local conflicts that are inevitable in the interface between the natural and the man-

PAGE 150

137 modified worlds. Only through the integration of applied research, implementation of management recommendations derived from these findings, involvement of NGOs, and participation of local communities through environmental education programs, will these species have a chance to survive.

PAGE 151

138 Table 6-1 — Comparison of results from different jaguar. HR= home range estimates (MCP); DENS expressed as % active day, % active night, sample studies on ocelot and .= density; ACTIVITY= sizes in parenthesis. SPECIES/AREA OCELOT : Iguagu* Belize'^ Venezuela" Pantanal"^ Peru' Texas* Texas' 66 38.8 (6) 31.0 (1) 10.2 (2) NA 5.7 (3) 12.3 (5) 6.3 (3) HR DENS. $9 #ad/km2 17.4 (5) 14.6 (1) 3.4 (6) 1.3 (3) 1.8 (3) 7.0 (3) 2.9 (3) 0.14 NA 0.40 0.76 0.80 NA NA ACTIVITY HABITAT D! N: 34% 41% D: 25% N: 44% D: N: D: N: D: N: 29% 75% 43% 51% 27% 66% NA D: 19% N: 69% subtropical forest tropical forest gallery forest, f loodplain deciduous and gallery forest, f loodplain seasonally flooded forest thorn forest thorn forest JAGUAR: Iguagu' 88.7 (4) Belize** 33.4 (4) Pantanal^ 152.0 (1) P. Primavera^ 259.0'' (1) 70.0 0.037 D; 73% (1) N: 65% 10.0 0.066 (1) 140.0 0.032 D: 52% (4) N: 38% 99.0 0.033 NA (5) subtropical forest tropical forest deciduous and gallery forest, floodplain deciduous and gallery forest, floodplain ' This study; Konecny, 1989; ° Ludlow and Sunquist, 1987 (includes data from Sunquist et al., 1989); crawshaw and Quigley, 1989; * Emmons, 1988; * Tewes, 1986; ' Laack, 1991; ^ Rabinowitz and Nottingham, 1986 and RaJainowitz, 1991; ^ Crawshaw and Quigley, 1991; ^ Crawshaw, 1995 and Crawshaw et al., 1993; ^ this male was translocated a distance of 35 km, but has since commuted back and forth between the old and new areas; NA: not available

PAGE 152

139 Table 6-2 — Comparison between mean prey-predator ratio (MPPR) and mean prey weight (MPW, kg) among different studies of ocelot and jaguar. OCELOT JAGUAR MPPR MPW MPPR MPW Iguagu* 0.13 1.4 0.18 14.4 Belize'''^ 0.24 2.4 0.14 6.7 Venezuela*^ 0.052 0.52 Peru* 0.076 0.608 0.34 10.8 Pantanal* 0.31 28.5 Pantanal^ 0.70 63.0 Venezuela** 0.54 49.7 ' This study; Konecny, 1989; "= Rabinowitz and Nottingham, 1986; Ludlow and sunquist, 1987; * Emmons, 1987; * crawshaw And Quigley, 1984, in prep.= only native prey considered; ' same reference but including cattle; Hoogesteijn et al., 1993.

PAGE 153

140 Table 6-3 — Percentages of the total estimated weight of the different prey species in the scats of ocelot (N= 56), puma (N= 28), and jaguar (N= 73), Iguagu National Park, Brazil. TEWS= total estimated weight (kg) in scats (N= 157). Only those prey species that ocurred more than twice in the sample were included. SPECIES TEWS OCELOT PUMA JAGUAR TOTAL Peccary 1,320.0 0 13.6 86.4 100 Deer 338.0 6.6 33.4 60.0 100 Armadillo 86.0 26.9 38.6 34.5 100 Coati 44.0 0.0 25.0 75.0 100 Agouti 41.6 46.2 15.4 38.4 100 Opossum 34.5 30.4 26.1 43.5 100 Lizard 19.2 16.7 33.3 50.0 100 Small Rodents 1.7 93.9 6.1 0.0 100

PAGE 154

141 Cumulative Mortality {%] at 1 uu 80 60 4 U 20 0 / .*•*** 1 1 1 1 1 1 1 1 0 60 120 180 240 300 360 420 Monitoring Period (days) 480 540 600 Ocelot — *— Jaguar Figure 6-1 — Mortality rates of ocelot and jaguar in Igua
PAGE 155

142 100 PERCENT 80 60 40 20 OCELOT PUMA JAGUAR Peccary Agouti Deer Opossum Armadillo Coati Lizard \M\ Small Rodents Figure 6-2 — Comparison of the main prey species in the diets of jaguar, puma, and ocelot in Igua^u National Park, Brazil.

PAGE 156

143 PERCENT 80 60 40 20 OCELOT Peccary Agouti PUMA JAGUAR Y//A Deer tttttj Armadillo I I Opossum Lizard Coati Small Rodents Figure 6-3 — Percentage of each main prey species in the diets of ocelot, puma, and jaguar, corrected by estimated biomass of each prey species in the scat samples (ocelot= 56; puma= 28; jaguar= 73) .

PAGE 157

144 Parana River Iguazu National Parle (56.000 ha) BRAZIL kRGENTINA^ Iguazu National Park (17S,000 ha) llguacu River BRAZIL Urugua-i Provincial Reserve (84,000 ha) u 0 STATE OF SANTA CATARINA BRAZIL MISSIONES >ROVINCEl: Yaboti Biosphere Reserve (223,000 ha) -Peperiguagu River ,Mocona Provincial Park (1,000 ha) Turvo State Park (17,500 ha) STATE OF RIO GRANDE DO SUL BRAZIL Uruguai River 1111 REMAINING SUBTROPICAL FOREST Figure 6-4 — Map of the province of Missiones, Argentina, showing remaining forest cover and some of the protected areas. Iguazu National Park and Turvo State Park, both in Brazil, are also shown. Scale: 2.5 cm = 52 km.

PAGE 158

APPENDIX A MAMMALS RECORDED IN IGUAQU Mammal species (bats excluded) recorded in Igua9U National Park (or immediate vicinity), Brazil, between March 1990 and December 1994. MARSUPIALIA Black-eared opossum ( Didelphis aurita) White-eared opossum (D. albiventris ) Four-eyed-opossum ( Philander opossum ) Whooly opossum ( Caluromys lanatus ) EDENTATA Nine-banded armadillo ( Dasypus novemcinctus ) Six-banded armadillo ( Euphractus sexcintus ) Collared anteater ( Tamandua tetradacty la ) RODENTIA Akodon montensis A. niqrita A» cf . cursor A. cf . serrensis A. cf . azarae Orvzomys f lavescens 0. niqripes Nectomys squamipes Holochilus brasiliensis . Tree-squirrel (Sciurus aestuans) Rabbit ( Silvilaqus brasiliensis ) Cavy (Cavia sp.) Coendou ( Sphyqurus prehensilis ) Agouti ( Dasyprocta azarae) Paca (Agouti paca ) Capybara ( Hydrochoerus hvdrochaeris ) ARTIODACTYLA Red brocket deer (Mazama americana) Rufous brocket deer (M^ nana) Collared peccary ( Tayassy tajacu ) White-lipped peccary ( Tayassu pecari ) 145

PAGE 159

146 PERISSODACTYLA Tapir ( Tapirus terrestris ) PRIMATES Capuchin monkey (Cebus apella ) Howler monkey ( Alouatta fusca) CARNIVORA Crab-eating fox ( Dusicyon thous) Crab-eating raccoon ( Procyon cancrivorus ) Bush dog ( Speothos venaticus ) ?^ Coati (Nasua nasua) Grison ( Galictis cuja ) Tayra (Eira barbara) River otter (Lutra lonqicaudis ) Giant river otter (Pteronura brasiliensis ) ?^ Ocelot (Felis pardalis ) Margay ( Felis wiedii) Puma (Puma concolor) Jaguarundi ( Felis vagouaroundi ) Jaguar (Panthera onca ) ^ One sighting reported during the study period; evidence from chromatography results; evidence from scat analysis (hair identified by distinctive cuticule and medulary patterns ) . ^ Recorded in Iguazu NP, Argentina; reported by Forest Police in Iguagu NP, Brazil.

PAGE 160

APPENDIX B SIGHTING RECORDS IN IGUAQU Sighting records of mairanals in Iguagu National Park (or immediate vicinity), Brazil, from March 1990 through May 1995. SPECIES # SIGHTINGS % TOTAL N % NON-CARNIVORES < 10 KG OF BODY WEIGHT Aqouti (Dasvprocta azarae) 658 57.7 45.8 Rabbit (Silvilaqus brasiliensis ) 215 18.9 15.0 Capuchin monkey (Cebus apella) 153 13.4 10.6 Opossum (Didelphis sp.) 61 5.3 4.2 Armadillo (Dasypus novemcinctus ) 33 2.9 2.3 Tree-squirrel (Sciurus aestuans) 11 1.0 0.8 Howler monkey (Alouatta fusca) 2 0.2 0.1 Coendou (Coendou prehensilis) 2 0.2 0.1 Caw jcavia sp. ) 2 0.1 0.1 Four-eyed-opossum (Philander opossum) 1 0.1 Whoolv opossum (Caluromys lanatus) 1 0.1 1, 139 99.9 79.0 NON-CARNIVORES > 10.1 KG OF BODY WEIGHT Red brocket deer (Mazama americana ) White-lipped peccary ( Tayassu pecari ) Collared peccary ( Tayassu tajacu ) Rufous brocket deer (M^ rufina) Unidentified deer ( Mazama sp.) Tapir ( Tapirus terrestris) Capybara ( Hydrochoerus hydrochaeris ) Paca ( Aqouti paca ) Collared anteater ( Tamandua tetradactyla ) 110 60.8 7.7 23 12.7 1.6 13 7.2 0.9 11 6.1 0.8 9 5.0 0.6 5 2.7 0.3 4 2.2 0.3 4 2.2 0.3 2 1.1 0.1 TsT 100.0 12.6 CARNIVORES Coati (Nasua nasua) 36 31.0 2.5 Crab-eating fox ( Dusicvon thous ) 27 23.3 1.9 Puma ( Puma concolor) 14 12.1 1.0 Continued — sighting records of mammals in Igua9u National Park (or immediate vicinity), Brazil, from March 1990 through May 1995. SPECIES Tayra ( Tayra barbara) # SIGHTINGS % TOTAL N % 12 10.3 0.8 147

PAGE 161

Jaguarundi (Felis yaqouaroundi ) Jaguar (Panthera onca) Grison (Galictis cuja ) Ocelot ( Felis pardalis ) River otter (Lutra platensis ) 148 8 6.9 0.5 7 6.0 0.5 4 3.4 0.3 4 3.4 0.3 4 3.4 0.3 116 99.8 8TT TOTAL SIGHTINGS = 1,436

PAGE 162

APPENDIX C CARNIVORE TRAPPING RESULTS IGUAgU PROJECT, BRAZIL Ocelots A total of 21 ocelots were captured in Iguagu and Iguazu National Parks, and equipped with radiocollars (Table C-1). Although the overall sex ratio of captured animals in both parks was about equal (11 males/ 10 females, or 1.0:0.9; Figure C-1), it differed in opposite direction between Parks. In INPB, 9 males and 6 females were captured (1.0:0.6), whereas in INPA, it was two males and four females (1.0 M:2.0 F) . Forty-seven percent of the animals were more than 5 years old, indicating a stabilized population. Ocelot capture rate for 4,700 trap-nights in Iguagu (March 1990/October 1994; 797 trapping days, mean number of traps per day = 5.9) was 0.93% (106 trap-nights/ocelot capture), including recaptures (N = 44). (Note: trapping results from Iguazu NP were not included because data on trapnight numbers were incomplete; 9 captures of 6 individuals). Overall, highest monthly capture success, including recaptures, was 2.45%, in September (Figure C-2), followed by May (2.41%) and August (2.15%). Figure C-3 shows monthly 149

PAGE 163

150 capture success considering only first captures. There was a sharp peak in capture success for males in May, whereas females showed a peak in August, coinciding apparently with reproductive activities (see section on Reproduction). However, M05 was captured a total of 5 times (including first capture) between 1 September and 25 October 1990, with a mean interval of 13.5 ± 13 days (range= 1-31) between captures. Of 7 recaptures of M27, 3 were in August and 1 in September 1992 (4, 23, and 20 days interval between recaptures), 1 in September 1993, and 1 in September 1994. Although the difference was not statistically significant, males were apparently more prone to entering traps, comprising 60% of 15 first captures of individuals, and 63% of 38 captures including recaptures (Test of Proportions, P > 0.05; Hintze, 1987). Similarly, there was no difference between the means of the interval ( in days ) between recaptures for males and females (Mann-Whitney two-sample test, z= 1.22, P= 0.22) . Traps baited with dead chickens (N = 371) were more successful (6 captures, capture rate = 1.6%) than those baited with live chickens (N = 2,503; 22 captures; capture rate = 0.9%), or with live rats (N = 925; 3 captures; capture rate = 0.3%). However, dead chicken also attracted several nontarget species, such as opossum ( Didelphis marsupialis , 145 captures, and D. albiventris . 4 captures) and tegu lizards ( Tupinambis tequixin . 8 captures).

PAGE 164

151 Mean body length of males was 80.1 cm ± 3.3 cm (range = 76.0 85.5, N= 11), and of females was 74.6 cm ± 3.9 cm (range= 65.0 79.5, N= 15, Table C-1). The mean weight of males was 12.8 kg ± 2.0 kg (range= 9.0 15.7, N= 11), and of females was 9.9 kg ± 1.3 kg (range= 6.7 12.4, N= 14). Males were larger ( z= 3.39, P < 0.002) and heavier ( z= 3.08, P < 0.002) than females (Mann-Whitney unpaired test). Weight showed a significant correlation both with body length and with chest girth, for males and females (Figures C-4 and C5)). Jaguar Capture rate for 472 trap-nights (April 1990 April 1994) in Iguagu NP was 2.7%, or 1 capture for 36.3 trapnights (13 captures, including 8 recaptures of M33). Capture rate was much higher for the two males trapped in Iguazu (about 22%, or 1 capture for 4.7 trap-nights), but these data are biased since the two animals had habituated to feeding on domestic animals (3 captures, including 1 recapture of M48). Trained dogs were used on three occasions when attempting to recapture animals to change the collar (M13, M32, M48). Measurements of the study animals are presented in Table C-2. Weight showed a high correlation with chest girth (r^= 0.8841, P < 0.001; N= 37; Figure C-6), thus allowing for weight estimation when a large scale is not available in the

PAGE 165

152 field (y= -63.9 + 1.59x). There was no inter-sex difference in these measurements for this sample (18 males, 19 females, Mann-Whitney unpaired two-sample test) . There was a positive correlation between weight and estimated age for males 0.6402, P= 0.031; N= 7) and females (r^= 0.8309, P= 0.002; N= 8). Only animals with < 36 months were included in the analysis, since past that age, weight tends to reach an asymptote. Animals from other localities were also included (total of 15 animals: 7 MM, 8 FF) . The resulting regression lines (males: y= 3.96 + 0.17x; females: y= -13.3 + 0.61x) may be used as a predictor in estimating age of subadult jaguars [of the paraquavensis (Seymour, 1989) subspecies] of known weight (Figure C-7). I also ran a multiple regression analysis of weight against track measurements (front paw measured on the animals) to aid in information collected from tracks measured in the field. Paw measurements used were: total paw length (TPL) , total paw width (TPW) , heel pad length and width (HPL and HPW) , and toe length and width (TOL and TOW), measured on the 3rd toe (Becker and Dalponte, 1991; Crawshaw, 1991). For this analysis, I also included animals from other areas of Brazil (total of 28 animals: 12 MM, 16 FF) . Of the six measurements, only TPL and TOL showed significant correlation with weight (r^= 0.5719, P = 0.000, N= 28, and r^= 0.6776, P= 0.000, N= 23, respectively; Figures C8, C-9). The regression equations (y= -75 + 16. Ix and y= -70

PAGE 166

153 + 43. 5x, respectively) should prove useful in estimating weight (and, indirectly, age) from jaguar tracks measured in the field.

PAGE 167

154 Table C-1 — Measurements of ocelots from Iguagu (BR) and Iguazu (AR) National Parks. Weight ^n kg, measurements in cm; SA= subadult; AD= adult; Drugs: K= Ketaset, z= Zoletil; rate in mg/kg. One recapture of F02 was included because of significant change in measurements. AN* SITE DATE SEX AGE WGT BODY TAIL CHEST F02 Iguagu, BR 08/21/90 F SA 6.7 65.0 30.0 33.0 F02 Iguagu, BR 4/24/91 F SA 11.3 75.0 33.0 42.0 F03 Iguagu, BR 08/23/90 F AD 9.0 76.0 30.0 45.5 F06 Iguagu, BR 09/15/90 F AD 9.9 76.0 30.0 39.0 F09 Iguagu, BR 10/18/90 F AD 11.3 79.0 33.0 40.0 F16 Iguazu, AR 07/30/91 F AD 9.0 74.0 33.0 43.0 F19 Iguazu, AR 08/03/91 F AD 9.5 75.5 37.0 40.0 F23 Iguazu, AR 01/25/92 F SA 10.0 71.0 34.5 38.0 F39 Iguagu, BR 02/19/94 F AD 10.5 79.0 35.0 38.0 F40 Iguagu, BR 03/18/94 F AD 10.1 79.5 33.5 37.0 F42 Iguazu, AR 04/05/94 F SA 9.6 73.0 34.0 37.5 M04 Iguagu, BR 08/30/90 M AD 12.5 83.0 36.0 48.5 NO 5 Iguagu, BR 09/01/90 M AD 12.0 77.0 35.0 47.5 M20 Iguazu, AR 08/14/91 M AD 15.5 84.0 35.0 46.0 M25 Iguagu, BR 02/19/92 M SA 10.0 77.0 33.0 39.0 M26 Iguagu, BR 05/23/92 M AD 12.0 79.0 35.5 44.0 H27 Iguagu, BR 05/27/92 M AD 14.0 79.0 39.0 45.0 H35 Iguagu, BR 09/21/93 M AD 15.1 85.5 37.0 43.5 M36 Iguazu, AR 10/04/93 M SA 9.0 76.0 41.0 34.0 M37 Iguagu, BR 02/13/94 M SA 12.0 77.0 35.5 41.0 M41 Iguagu, BR 04/04/94 M SA 13.0 83.0 36.0 44.0 M45 Iguagu, BR 05/11/94 M AD 14.0 79.5 35.0 49.0

PAGE 168

155 Table C-2 — Measurements of jaguars from Iguagu (BR) and Iguazu (AR) National Parks. Weight in kg, measurements in cm. Same animal numbers indicate recaptures of animals first captured as subadults. BL= body length. AN# SITE DATE SEX AGE BL TAIL CHEST Wol 17 Igua<;u, BR 07/31/90 F AO 141 . 0 54 . 0 91.0 /O . 0 21 Iguagu, BR 09/12/91 F SA 113 . 0 54 . 0 70.0 47 . 0 13 Iguagu, BR 04/19/91 H SA 118 . 0 60 . 0 88.0 60 . 0 13 Iguagu, BR 03/13/92 M AD 136.0 60.0 87.0 88.0 32^ Iguazu, BR 11/17/92 M SA 122.0 57.0 63.0 37.0 33 Iguagu, BR 08/11/93 M AD 120.0 64.0 85.0 80.0 34 Xguagu, BR 08/14/93 M SA 106.0 54.0 70.0 45.0 48 Iguazu, AR 09/22/94 M AO 124.5 62.0 86.0 77.0 UN' Iguazu, AR 12/29/92 M AD 164.0 66.0 93.0 92.0 UN* Iguagu, BR 09/09/93 H AO 135.0 65.0 96.0 92.0 ^ — Approximate weight, for only an inaccurate scale was available. 2 — When found dead, on 11 June 1993, his weight was estimated at about 90 kg (see section on Mortality). ^ — Animal captured in Iguazu National Park, AR, by Park personnel, translocated to Parque Provincial Urugua-i, and released without a radiocollar . * — Uncollared animal wounded by poacher and killed by truck in Iguagu National Park, BR.

PAGE 169

156 Table C-3 — Measurements of other carnivores from Iguagu (BR) and Iguazu (AR) National Parks. WGT: mass (kg); BL: body length (cm); CHEST: chest girth. Raccon: Procyon cancrivorus ; coati: Nasua nasua ; tayra: Eira barbara; fox: Cerdocyon thous ; little-spotted cat Felis tiqrina ; margay: F. wiedii ; jaguarundi: F. yaqouaroundi ; Puma: Puma concolor . SPECIES AGE SEX DATE BL TAIL CHEST WGT Raccoon AD F 07/24/90 55.5 31.5 40.0 7.3 Raccoon AD M 12/02/92 67.8 33.0 40.0 8.5 Coati YG F 02/20/91 37.5 34.5 20.5 1.2 Coati YG F 04/06/91 44.5 43.5 26.0 1.9 Coati AD F 04/24/91 56.0 51.0 37.5 4.8 Coati AO F 10/13/92 60.0 48.0 33.5 6.0 Coati AD F 05/26/94 58.0 46.0 37.0 6.0 Coati' AD F 04/09/94 55.0 50.0 37.0 5.3 Coati* AD M 02/05/91 61.0 51.0 38.0 7.0 Coati AD M 08/12/93 60.0 50.0 40.0 6,0 Coati* AD M 04/09/94 61.0 54.0 34.5 6.7 Tayra* SA M 09/16/92 65.5 25.5 31.5 5.0 Tayra* AD M 09/23/92 66.5 30.5 33.5 6.0 Tayra** AD M 11/04/93 70.5 36.5 35.0 7.7 Fox AD F 05/03/91 68.0 29.0 42.0 8.2 Fox AD F 05/24/91 72.0 29.0 49.0 6.8 Fox AD F 09/23/91 67.0 25.0 37.5 5.5 Fox AD F 07/09/92 68.0 31.0 _ 6.0 Fox SA F 09/22/93 65,0 26.0 35.2 5.1 Fox AD F 11/03/93 64.0 36.6 36.0 5.5 Fox" AD F 07/23/90 62.5 21.0 37.0 5.7 Fox* r m/1 1 /Qi uo / XI. / y i. 40 . 5 7 , 3 Fox* AD F 05/13/91 64.5 29.0 38.0 6,3 Fox* AD F 09/19/92 64.0 29.0 37.5 5,7 Fox AD M 09/07/92 70.0 33.0 7.0 Fox* SA M 02/03/91 68.0 28.0 34.0 4.8 Fox* AD M 03/01/91 74.0 26.5 38.0 6.5 Fox* AD M 09/21/91 66.5 37.0 5.7 Fox* AD M 06/25/92 72.0 28.0 37.0 6.5 Fox AD ? 06/12/92 74.5 30.0 35.5 5.0 Little-spotted A M 10/15/90 53.0 28.5 25.5 3.0 cat*

PAGE 170

157 Table C-3 — Measurements of other carnivores from Iguaqu (BR) and Iguazu (AR) National Parks. WGT: mass (kg); BL: body length (cm); CHEST: chest girth. Racoon: Procyon cancrivorus ; coati: Nasua nasua ; tayra: Eira barbara; fox: cerdocycn thous; little-spotted cat Felis tigrina; margay: F. wiedii; jaguarundi: F, vaqouaroundi ; Puma: Puma concolor . T ^% ^ SPECIES C IT V n ATT? U Ax sis RT. TATT. CHEST WGT Lxttle-spotted YG u iU/ i7 / y o ^ / • D 1 S ^ 0 . 7 cat Margay' A M 07/13/92 51.0 33.5 27.0 3.5 Margay*" 1 F 12/16/94 48.0 30.0 22.5 2.3 Margay° 2 F 07/30/92 30.5 26.5 2.0 Jaguarundi* F 08/01/91 60.0 38.5 29.0 3.5 Jaguarundi"* e F 07/31/87 59.0 42.0 27.5 3.9 Jaguarundi** 1 F 08/18/92 59.5 38.0 32.0 4.9 Jaguarundi' 47 M 09/17/94 66.5 43.0 26.5 5.0 Jaguarundi* 49* M 09/27/94 55.5 36.5 26.0 3.0 Jaguarundi** 1 M 07/10/91 67.0 42.0 34.0 5.5 Jaguarundi** 1 M 08/27/92 66.5 38.5 34.0 6.0 Jaguarundi** 1 M 10/25/93 73.5 31.0 20.0 1.5 Puma^ AD F 08/26/92 113.5 66.5 61.0 35.4 Puma SA M 08/07/94 84.0 51.0 41.5 13.0 * — Captured in the Park and radiocollared. ** — Roadkills, Iguagu. = — Animals measured at Itaipu Zoo, Foz do Iguagu, PR. — Animal captured in Iguazu NP, Argentina. • — Animal measured at Ilha Solteira Zoo, SP, 07/31/1987. * — No radio-collar available at the time of capture. 9 — Animal captured in a wire-snare by locals, while preying on livestock south of Iguazu NP, Argentina.

PAGE 171

158 Figure C-1 — Sex/age structure for 21 ocelots captured in Iguagu and Iguazu National Parks, Brazil and Argentina, between August 1990 and May 1994.

PAGE 172

159 1.6 Capture Success (%1 Jan 172 Feb 709 Mai 506 Api 259 May 207 Jun Jul 265 345 Month #TN Aug 372 Sep 489 Oct 410 Nov 259 Dec 88 Y//A Maifta ^3 Females Total Figure C-2 — Monthly capture success for male and female ocelots in Iguacu National Park, including recaptures (N= 38). Number below month indicate number of trap-nights.

PAGE 173

160 Capture Success (%) 3 I Jan Feb Mai Apr May Jun Jul Aug Sep Oct Nov Dec 172 709 506 259 207 265 345 372 489 410 259 88 Month #TN Y//A Males 1^^^^ Females -*Total Figure C-3 — Monthly capture success for male and female ocelots in Iguagu National Park, Brazil. Only first captures included (N= 15). Number below month indicates number of trap-nights .

PAGE 174

161 Figure C-4 — Relationship between body length and weight in ocelots (measurements of animals from elsewhere in Brazil were included) .

PAGE 175

162 Figure C-5 — Relationship between chest girth and weight of ocelots (measurements of animals from elsewhere in Brazil were included).

PAGE 176

163 Figure C-6 — Correlation between chest girth (cm) and weight (kg) for jaguar (N= 37). Animals from other parts of Brazil were included in the analysis .

PAGE 177

164 MONTHS 0 20 40 60 80 100 WEIGHT (Kg) Figure C-7 — Correlation between weight and estimated age for jaguars with < 36 months (after this age, an asymptote is reached on the weight axis). Animals from other areas of Brazil were included.

PAGE 178

Figure C-8jaguars (N= -Correlation between total paw length (cm) and weight (kg) of 28) .

PAGE 179

166 120 WEIGHT (kg) |Y» -70 43.5x1 Eye a 0 DD 1 100 80 60 40 20 1.5 2.5 3 3.5 Toe Length (cm) 4.5 Figure C-9 — correlation between toe length (cm) and weight (kg) of jaguar (N= 26) .

PAGE 180

APPENDIX D NOTES ON MARGAY (Felis wiedii ) AND JAGUARUNDI (E\ yaqouaroundi ) IN IGUAgU NATIONAL PARK, BRAZIL Introduction As for most of the neotropical small cats, little is known about the biology of the margay and jaguarundi in the wild. In the first intensive field study on these species, Konecny (1989) radiocollared one margay and three jaguarundis in Belize, providing information on individual movement and activity patterns. Bisbal (1986, 1987, 1989) presented information on the food habits, distribution and habitat association from specimens in museum collections. Mondolfi (1986) summarized anecdotal information on margay, with data on stomach contents from a few specimens. Manzani and Monteiro Filho (1989) and McCarthy (1992) added information on food habits and activity and behavior. In the course of a carnivore ecology study in Iguagu National Park, in southwestern Brazil, one margay and two jaguarundis were captured and equipped with radiocollars . In addition, information from roadkilled animals was collected. Results of the study provide insights into the ecology of 167

PAGE 181

168 these species in subtropical rainforest, allowing for comparisons with that from different environments. Results and Discussion Three jaguarundis (one subadult female and two subadult males) , and one adult male margay were captured in 3,599 trapnights (672 trapping days, mean number of traps per day =5.3) in Iguagu and Iguazu National Parks, respectivelly . Two jaguarundis (1 male and 1 female) and the margay were equipped with radiocollars (Appendix C, Table C-1). The third jaguarundi was not radio-collared for lack of a transmitter at the time of capture. I believe margay capture rate would have been better if traps had been set in trees, given the more arboreal habits of this species (citation) . However, I do not know the reason why jaguarundis were not trapped more frequently. Unfortunately, Konecny (1990) does not mention the capture rate for his study site in Belize. Jaguarundis were relatively common in Iguagu, comprising 7% of 103 sightings of carnivores between March 1990 and November of 1993 (Appendix B) . This percentage is twice as large as that for ocelots, of which there were 40 captures for the same number of trapnights. In addition, 4 jaguarundis were found dead (3 roadkills', 1 drowned in canal) in the Park or in the immediate vicinity (Appendix C, Table C-3).

PAGE 182

169 Radiotelemetry : Jaquarundi ; The female jaguarundi was located 8 times after capture, in a period of 18 days. During this period, she occupied an area of 19.6 km^. On 20 August, the mortality mode of the transmitter was heard, and on field checking, her remains were found covered by litter, with circumstantial evidence of predation by puma. The male jaguarundi (M47) was captured on 17 September 1994 and has been monitored to date (17 locations, at a mean interval of 9.2 ± 10.6 days, range= 1-35 days). An analysis of his locations plotted on a map of the study area reveals one of the problems in estimating home ranges. His locations are either concentrated in the vicinity of the hotel by the falls, or they follow the course of the Iguagu river, before and after the falls. If all his locations are included in the calculations, the area encompassed by the circle includes a large (unused) portion of the Iguagu river, for a total of 46.7 km^. To circumvent this problem, his locations were divided as forming two polygons that added gives a total of 7.2 km^. Marqay ; During the 18 months of monitoring (13 July 1991 14 January 1993, 89 locations, at a mean interval of 6.2 days ± 6.1, range= 0 30), the margay ranged over an area of 15.9 km^ (MCP) or 43.1 km^ (HM). A cummulative area curve (Odum and Kunzler, 1955) indicated that his home range was still

PAGE 183

170 increasing by the time the battery in his collar failed. With the margay, the problem of home range shape did not occur (Figure D-1 ) . A seasonal analysis revealed no differences in mean linear distance between consecutive locations. Similarly, there was no difference in the linear distances between locations obtained at one, two, three, four, and five days interval (Table D-2). Considering the total number of activity readings (N= 530), there was no difference between daytime (0600 h to 1759 h: 51% of 356 readings) and night-time (1800 h to 0559 h: 57% of 174 readings) readings (Z = 1.311, P = 0.1898; Two-Sample Proportion Test, Hintze, 1987). There was no difference in activity levels between winter (N= 247 readings) and other seasons combined (N= 144). The limited access to the radioequipped animals, due to the dense vegetation and lack of trails, precluded consistent monitoring of diel activity patterns, with too few activity readings for some of the hourly intervals. Mortality One margay and five jaguarundis were killed by vehicles in or near Iguagu NP (Table 6.1). One subadult male jaguarundi drowned in a human-made canal in the Park. As already mentioned, F18 was apparently killed by a puma, 20

PAGE 184

171 days after being radio-collared. I do not know if the collar in anyway influenced her death.

PAGE 185

172 Table D-1 — Comparison of home range estimates (in km^) for 1 margay (Felis wieddi) and 2 jaguarundis (F. yaqouaroundi ) equipped with radio-collars in Iguagu and Iguazu National Parks, Brazil and Argentina, derived from traditional methods and the proposed Circle Method. N= number of locations (no outliers were excluded); MC= Minimum Convex Polygon method; HM= Harmonic Mean method; ELL= 95% Ellypse method; CM= Circle method (see text). AN# N MCP HM ELL CM Margay (Ml 5) 89 15.9 43.1 24.4 15.9 Jaguarundi (F18) 8 6.8 55.4 46.7 22.1 Jaguarundi (M47) 17 17.6 72.7 69.0 46.7

PAGE 186

173 Table D-2 — Mean linear distances moved between locations with 1-5 days interval for one adult male margay (Felis wieddii) in Iguagu National Park, Brazil. INTERVAL NUMBER OF MEAN ± SD RANGE ( DAYS ) LOCATIONS 1 14 1.3 ±0.8 0 3.0 2 10 1.5 ± 0.9 0.1 3.4 3 7 1.9 ± 1.0 0.5 3.2 4 12 1.7 + 1.1 0.3 3.5 5 9 1.8±1.1 0.7-3.9

PAGE 187

174 PERCENT OF LOCATIONS • 50 1 (1 1-2 2-3 3-4 4-5 )5 DISTANCE (in kml Y//A Margay M15 IssSSl Jaguarundi F18 tttttj Jaguarundi M47 IN' 891 IN' 81 (N* 17) Figure D-1 — Movements of one male margay and two jaguarundis (1 M, IF) in Iguagu and Iguazu National Parks, Brazil and Argentina.

PAGE 188

APPENDIX E SUMMARY OF ENVIRONMENTAL EDUCATION ACTIVITIES PROJECT "CARNIVORES OF IGUAgU" Between 1990 and 1993, considerable effort was spent on several environmental education activities within and at the periphery of Iguagu National Park, Brazil, as part of "Projeto Carnivores de Iguagu" (Carnivores of Iguagu) . The objectives of the program were twofold: first, create an environmental awareness to the situation of carnivores, emphasizing the need to proper management and conservation measures; secondly, to provide a better understanding of the Park and of the ecosystem it protects. Most of the over 800,000 annual visitors to the Park associate the national park just with the Iguagu Falls, ignoring its size and importance, as the largest sanctuary of wildlife in southern Brazil. The environmental education program (EEP) was designed to attend 6 different target-groups (or public): a) students of the school inside the Park, Escola Municipal Jose Acelino de Carvalho; b) students from schools in and nearby Foz do Iguagu; c) students from schools in the towns of Medianeira, Ceu Azul, and Capanema; d) tourists (both national and international); 175

PAGE 189

176 e) park wardens of the Battalion of Military Police and IBAMA employees living inside the Park; f) other segments of the community, including ranchers and farmers and their employees, and neighboring residents . In addition, 44 students and professionals of Biology, Ecology, and related fields passed though Iguagu as trainees, for periods that ranged from 3 to 90 days. Training involved field experience on research techniques, with special emphasis on radio-telemetry. Some of the trainees have assumed important positions in their institutions, have initiated their own research on carnivores, or have started working on graduate studies, with an emphasis on carnivores. Listed below are some of the highlights of the training program in Iguagu (in chronological order): Leandro Silveira, biologist, Universidade Catolica de Goias; presently initiating a carnivore ecology study, coordinated by PC, in Emas National Park, in Goias; Lucila Manzatti, ecologist, Centre de Monitoramento Ambiental da Serra do Itapeti, Mogi das Cruzes, SP; coordinator of the Environmental Education component of the Iguagu project, expanding to other projects in association with PC; Sandra Maria Cintra Cavalcanti, agronomist, Faculdade de Agronomia e Zootecnia Manoel Carlos Gongalves, SP. Worked for one year as a field assistant in the Iguagu

PAGE 190

177 project, and is now enrolled in a MS program at Utah State University, Logan, USA, focusing on management problems of large carnivores; Ricardo Luiz Pires Boulhosa, biologist, Universidade de Sao Paulo USP; presently conducting a 3-year study on livestock predation by jaguar and puma in Pocone, northern Pantanal of Mato Grosso, together with PC; Cibele Barros Indrusiak and Jan Mahler Jr., biologists, Universidade Federal do Rio Grande do Sul. They worked as a field assistants from March 1992 through November 93. Presently enrolled in the MS progrcim at the Universidad de Cordoba, Argentina; their thesis focus on carnivores and herbivores in Turvo State Park, Rio Grande do Sul; Ronaldo Gon^alves Morato, veterinarian, Universidade de Vigosa, Minas Gerais. Presently in the M. S. program at the Universidade de Sao Paulo, SP. His thesis focuses on reproductive ecology of felids. Maria Renata Pereira Leite, veterinarian, Universidade Federal do Parana; presently in the M. S. program at the same University. Her thesis will focus on status and feeding ecology of felids in the Atlantic forest in Parana state, including Superagui National Park. Pablo Perovic, Argentine Biologist, FUCEMA, Buenos Aires; presently conducting research on jaguar predation on livestock in Jujui, Salta, using compensation to ranchers

PAGE 191

178 for their losses as a strategy for jaguar conservation in the area; Fernando Dutra Lima, veterinarian, Companhia Vale do Rio Doce, Projeto Carajas. Director of the Parque Zoobotanico, Projeto Carajas, presently conducting research on ecology of large felids, in southern Para, in association with PC; Carlos Alberto T. De Lucca, ecologist, CESP; working in a large cat study (coordinated by PC) in Porto Primavera, MS, at the proposed site of a hydroelectric project, with the Companhia Energetica de Sao Paulo (CESP); Listed below are some of the Environmental Education activities conducted by project personnel: one full year (1993) working with the 43 first grade students in the school inside Iguagu National Park, with weekly activities (Cibele Indrusiak and Jan Karel Mahler Jr.); a survey on Park/neighbors relationships in 7 schools in the periphery of the park, involving 104 1st. grade students (Lucila Manzatti); a total of 23 talks about the project and Iguagu National Park to 1stand 2nd-grade students, teachers, and supervisors in county and state schoolsaround the Park, Park employees, guards (Policia Florestal), Antioch

PAGE 192

179 college students, tourist guides, for a total of 912 people (PC and co-workers); 14 talks on behavior and ecology of felids, as applied to coexistence with people (and how to avoid dangerous situations), in Carajas (900 people attended 6 talks), and in Trombetas (8 talks for 1,600 people), in Para state, Amazonia. These talks were requested by two big mining companies that run large projects (Companhia Vale do Rio Doce and Mineragao Rio do Norte) in those localities after an 8-year-old boy was killed by a puma in Carajas. TV programs were made for display at the local stations, and a special in the Fantastico series (for the national TV Globo) was made in Carajas (PC and Lucila Manzatti) ; one 4-hour talk on field research in Iguagu to biology students, Universidade do Vale do Rio dos Sinos, Sao Leopoldo, RS, at the 30th anniversary of the Biology Department (attended by about 90 students; PC); one talk and discussions on problems in reintroduction and translocation of wildlife, with emphasis on carnivores, at the II Encontro Nacional de Centres de Triagem de Animais Silvestres/IBAMA, ES (attended by about 50 biologists and veterinarians; PC); an itinerant poster about the Park and the project was exhibited for varying periods at the Bank of Brazil and

PAGE 193

municipal schools in Foz do Iguagu, Medianeira, and Ceu Azul, and at the hotel by the falls; between June 1992 and May 1993, about 14,000 people (mostly schools) visited the Museum at the Visitor Center in the Park, with displays on the wildlife of the Park and Pantanal of Mato Grosso (restructured by Project personnel ) ; two folders, one about the project and one alerting to problems of artificial feeding of coatis, have been made and distributed. Another folder is being made, on visitor procedures while in the Park, in conjunction with the Polxcia Florestal; a 7-minute special on the project was aired, at the national level (Fantastico, TV-Globo) , in October 1993. several news articles were published about the Iguagu project at the local level (Brazil and Argentina) and national level. Articles were published also in the US, Canada, Italy, and Japan.

PAGE 194

LIST OF REFERENCES Ackerman, B. B., Lindzey, F. G., and Hemker, T. P. 1984. Cougar food habits in southern Utah. J. Wildl. Manage. 48 (1): 147-155. Ackerman, B. B., Samuel, M. D., and Leban, F. A. 1991. User's manual for program HOME RANGE Version 2. Forestry, Wildl. and Range Experiment Station, University of Idaho, Moscow, 81 pp. Becker, M. 1981. Aspectos da caga em algumas regioes de cerrado de Mato Grosso. Brasil Florestal 11 (47): 51-63. Becker, M. and Dalponte, J. 1991. Rastros de mamiferos silvestres brasileiros. Editora da Universidade de Brasilia, Brasilia, DF, 180 pp. Beier, P. 1995. Dispersal of juvenile cougars in fragmented habitat. J. Wildl. Manage., 59 (2): 228-237. Bisbal, F. J. 1986. Food habits of some Neotropical carnivores in Venezuela (Mammalia, Carnivora) . Mammalia, 50 (3): 329-339. Bisbal, F. J. 1987. The carnivores of Venezuela: Their distribution and the ways they have been affected by human activities. M. S. Thesis, Univ. of Florida, Gainesville. Bisbal, F. J. 1989. Distribution and habitat association of the carnivores in Venezuela. In: Advances in Neotropical Mammalogy, Redford, K. H. and Eisenberg, J. F. (eds.). Sandhill Crane Press, Gainesville, Florida: 339-362. Brady, C. A. 1979. Observations on the behavior and ecology of the crab-eating fox ( Cerdocyon thous) . In: Vertebrate Ecology in the Northern Neotropics, Eisenberg, J. F., (ed.). The National Zoological Park, Smithsonian Institution, Washington, DC: 161-171. Chebez, J. C. 1995. Plan de Accion: Conservacion de las poblaciones de carnivores. Typed Report to the 181

PAGE 195

182 Administracion de Parques Nacionales, Buenos Aires, Argentina: 3 pp. Cockburn, A. 1992. Habitat heterogeneinity and dispersal: environmental and genetic patchyness. In: Animal Dispersal: Small Mammals as a Model. Stenseth, N. Chr. and Lidicker, Jr., W. Z. (eds.), Chapman and Hall, London, pages 65-95. Crawshaw, P. G., Jr. 1987. Preliminary survey on the carnivores in Iguagu National Park. Report to Program for Studies in Tropical Conservation, University of Florida, 8 pp. Crawshaw, P. G. , Jr. 1988. Ecology of bobcats in south Florida: Effects of flooding on diet, habitat use, and social structure. Project proposal to Ph. D. Graduate Committee, Department of Wildlife and Range Sciences, University of Florida, Gainesville, FL, 11 pp. Crawshaw, P. G., Jr. 1989. Responses of mammalian carnivores to seasonal resources in Iguagu National Park, Brazil. Research proposal to WWF-US, 16 pp. Crawshaw, P. G., Jr. 1990. Responses of mammalian carnivores to seasonal resources in Iguagu National Park, Brazil. Interim Report to WWF-US, 17 pp. Crawshaw, P. G., Jr. 1992. Recommendations for study design on research projects on neotropical felids. In: "Felinos de Venezuela: Biologia, Ecologia e Conservacion" , FUDECI, Caracas: 187-222. Crawshaw, P. G. , Jr. 1995. Relatorio sobre o monitoramento de grandes felinos na area de influencia da futura UHE de Porto Primavera, SP E MS Sub-Pro jeto Carnivores. CESP/THEMAG, Sao Paulo, SP, janeiro de 1995, 16 pp. (In Portuguese) Crawshaw, P. G., Jr. and Pilla, J. 1994. Ecologia e conservagao das populagoes de felinos do "Parque Florestal Estadual do Turvo", Rio Grande do Sul, Brasil. Report to World Wildlife Fund-US, 20 pp. (In Portuguese) Crawshaw, P. G., Jr. and Quigley, H. B. 1984. A ecologia do jaguar ou onga pintada (Panthera onca) no Pantanal Matogrossense. Final report to the Institute Brasileiro de Desenvolvimento Florestal IBDF, Brasilia, DF: 110 pp. (In Portuguese)

PAGE 196

183 Crawshaw, P. G., Jr. and Quigley, H. B. 1989. Notes on ocelot movement and activity in the Pantanal region, Brazil. Biotropica 21 (4): 377-379. Crawshaw, P. G., Jr. and Quigley, H.B. 1991. Jaguar spacing, activity, and habitat use in a seasonally flooded environment in Brazil. J. Zool. (London), 223: 357-370. Crawshaw, P. G., Jr. and Quigley, H. B. In prep. Jaguar and puma feeding habits in the Pantanal. Typed manuscript, 16 pp. Crespo, J. A. 1982. Ecologia de la comunidade de mamiferos del Parque Nacional Iguazu, Missiones. Rev. Mus. Argentine Ciencias Naturales "Bernardino Rivadavia", Ecologia, 3: 45-162. Dixon, K. R. and Chapman, J. A. 1980. Harmonic mean measure of animal activity areas. Ecology 61: 1040-1044. Eisenberg, J. F. 1981. The Mammalian Radiations: An Analysis of Trends in Evolution, Adaptation, and Behavior. University of Chicago Press, Chicago, 610 pp. Eisenberg, J. F. 1986. Life history strategies of the Felidae: variations on a common theme. In: Cats of the World: Biology, Conservation, and Management, (Miller, S. D. and Everett, D. D., eds.), Washington, DC: Nat. Wildl. Federation: 293-303. Eisenberg. J. F., O'Connell, M. A., and August, P. V. 1979. Density, productivity, and distribution of mammals in two Venezuelan habitats. In: Vertebrate Ecology in the Northern Neotropics, Eisenberg, J. F., (ed.). The National Zoological Park, Smithsonian Institution, Washington, DC: 187-207. Emmons, L. H. 1987. Comparative feeding ecology of felids in a neotropical rainforest. Behav. Ecol. Sociobiol. 20: 271-283. Emmons, L. H. 1988. A field study of ocelots in Peru. Rev. Ecol. Terre Vie 43: 133-157. Ewer, R. F. 1973. The Carnivores. Cornell University Press, Ithaca, N. Y., 494 pp. Fragoso, J. M. 1994. Large mammals and the community dynamics of an Amazonian rain forest. Ph. D. dissertation. University of Florida, Gainvesville, FL, 210 pp.

PAGE 197

184 Forys, E. A. 1995. Metapopulations of marsh rabbits: a population viability analysis of the Lower Keys marsh rabbit ( Svlvilaqus palustris hefneri ) . Ph. D. * dissertation, University of Florida, Gainesville, FL, 244 pp. Gilpin, M. E. 1987. Spatial structure and population vulnerability. In: Viable Populations for Conservation, Soule, M. E. (ed.), Cambridge University Press, Cambridge: 125-139. Greenwood, P. J. 1980, Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour, 68: 1140-1162. Guerrant, Jr., E. 0. 1992. Genetic and demographic considerations in the sampling and reintroduction of rare plants. In: Conservation Biology The Theory and Practice of Nature Conservation, Preservation, and Management, Fiedler, P. L. and Jain, S. K. (eds.). Chapman and Hall, New York and London: 321-344. Hayne, D. W. 1949. Calculation of home range size. J. Mammal. 30: 1-18. Heisey, D. M. and Fuller, T. K. 1985. Evaluation of survival and cause-specific mortality rates using radio-telemetry data. J. Wildl. Manage. 49 (3): 668-674. Hintze, J. L. 1987. Number Cruncher Statistical System. Program manual, Kaysville, UT. Hoogesteijn, R., Hoogesteijn, A., and Mondolfi, E. 1993. Jaguar predation and conservation: cattle mortality caused by felines in three ranches in the Venezuelan Llanos. In: Mammals as Predators, Dunstone, N. and Gorman, M. L. (eds.). Symposia Zool. Soc. London 65: 391407. Iriarte, J. A. 1988. Feeding ecology of the Patagonian puma ( Felis concolor patagonica ) in Torres del Payne National Park, Chile. M.A. Thesis, University of Florida, Gainesville, FL, 80 pp. Jennrich, R. I. and Turner, F. B. 1969. Measurement of noncircular home range. J. Theor. Biol. 22: 227-237. Jimenez, J. 1993. Comparative ecology of Dusicyon foxes at the Chinchilla National Reserve in north-central Chile. M. S. Thesis, University of Florida, Gainesville, FL, 168 pp.

PAGE 198

185 Johnson, M. K., Belden, R. C. and Aldred, D. R. 1984. Differentiating mountain lion and bobcat scats. J. Wildl. Manage. 48: 239-244. Jorgenson, J. P. 1993. Gardens, wildlife densities, and subsistence hunting by Maya indians in Quintana Roo, Mexico. Ph. D. Dissertation, University of Florida, Gainesville, FL, 336 pp. Jorgenson, J. P. and Redford, K. H. 1993. Humans and big cats as predators in the Neotropics. In: Mammals as Predators, Dunstone, N. and Gorman, M. L. (eds.), Symp. Zool. Soc. of London, 65: 367-390. Konecny, M. J. 1989. Movement patterns and food habits of four sympatric carnivore species in Belize, Central America. In: Advances in Neotropical Mammalogy, Redford, K. H. and Eisenberg, J. F. (eds.). The Sandhill Crane Press, Inc., Gainesville, FL: 243-264. Lacy, R. C. 1992. The effects of inbreeding on isolated populations: Are minimum viable population sizes predictable? In: Conservation Biology The Theory and Practice of Nature Conservation, Preservation, and Management, Fiedler, P. L. and Jain, S. K. (eds.). Chapman and Hall, New York and London, pages: 277-296. Laack, L. L. 1991. Ecology of the ocelot ( Felis pardalis ) in south Texas. M. S. Thesis, Texas A & I University: 113 PPLaack, L. L. and Tewes, M. E. 1988. Annual report for the 1988 fiscal year: utility of translocation for ocelot recovery in the United States. Caeser Kleberg Wildlife Research Institute, Kingsville, Texas, 24 pp. Lovejoy, T. E., Bierregaard Jr., R. 0., Rylands, A. B., Malcolm, J. R., Quintela, C. E., Harper, L. H., Brown Jr., K. S., Powell, A. H., Powell, G. V. N., Schubart, H. 0. R., and Hays, M. B. 1986. Edge and other effects of isolation on Amazon forest fragments. In: Conservation Biology: The Science of Scarcity and Diversity, Soule, M. E. (ed.), Sinauer Assoc., Massachusetts: 257-285. Ludlow, M. E. 1986. Home range, activity patterns and food habits of ocelot in Venezuela. M. S. Thesis, University of Florida, Gainesville, FL, 70 pp. Ludlow, M. E. and Sunquist, M. E. 1987. Ecology and behavior of ocelots in Venezuela. Natl. Geogr. Res. Rep. 3 (4): 447-461.

PAGE 199

186 Malcolm, J. 1993. The small mammals of Amazonian forest fragments: Pattern and Process. Ph. D. dissertation, University of Florida, Gainesville, 218 pp. Manzani, P.R. and Monteiro Filho, E. L. A. 1989. Notes on the • food habits of the jaguarundi, Felis yaaouaroundi (Mammalia, Carnivora) . Mammalia 53 (4): 659-660. McBride, R. 1976. Status and ecology of the mountain lion ( Felis concolor stanlevana ) of the Texas-Mexico border. M. S. Thesis, Sul Ross State University, 156 pp. McCarthy, T. J. 1992. Notes concerning the jaguarundi cat ( Herpailurus yaaouaroundi ) in the Caribean lowlands of Belize and Guatemala. Mammalia 56 (2): 302-306. Mech, L.D. 1983. Handbook of Animal Radio-tracking. Minneapolis: University of Minnesota Press. Mech, L. D. 1995. The challenge and opportunity of recovering wolf populations. Conservation Biology 9 (2): 270-278. Ministerio de Ecologia y Recursos Naturales Renovables. 1993. Areas protegidas de la selva paranaense en la provincia de Missiones, Argentina. MERNR, Missiones. Mohr, C. 0. 1947. Table of equivalent populations of North American small mammals. Amer. Midi. Nat. 37: 223-249. Mondolfi, E. 1986. Notes on the biology and status of the small wild cats in Venezuela. In Cats of the World: Biology, Conservation, and Management, eds. S. D. Miller and D. D. Everett, Washington, D.C.: National Wildlife Federation: 125-146. Navarro, D. 1985. Status and distribution of the ocelot ( Felis pardalis ) in south Texas, USA, Kingsville, Texas. M. S. Thesis, Texas A&I Univ., pp. 92 pp. Nowak, R. M. 1991. Walker's Mammals of the World Fifth Edition. John Hopkins University Press, Baltimore, MD, Volumes I and II, 1615 pp. Odum, E.P. and E.J. Kuenzler. 1955. Measurement of territory and home range size in birds. Auk 72: 128-137. O'Brien, S., Johnson, W. E., and Crawshaw, P. G., Jr. 1993. Wild felids of Brazil: A study on genetics, reproduction, and infectious diseases. Research proposal to the Brazilian National Research Council CNPq, 7 pp.

PAGE 200

187 Ott, L. 1984. An Introduction to Statistical Methods and Data Analysis. Duxbury Press, Boston, 775 pp. Padua, C. V. 1993. The ecology, behavior, and conservation of the Black-lion tamarin ( Leontopithecus chrvsopvgus , Mikan, 1823). Ph. D. Dissertation, University of Florida, Gainesville, FL, 182 pp. Perovic, P. G. 1993. Evaluacion del dano sobre la ganaderia, por actividad del overo (Panthera onca) , y propuestas para sua mane jo en un area de Las Yungas, Departamento Palpala, Provincia de Jujuy, Argentina. Unpubl. report, FUCEMA, Buenos Aires, Argentina, 21 pp. Poupard, J. P., Simao, A., Quintao, A. T., Porto, E. R. , and Comastri, E. M. 1981. Piano de Manejo do Parque Nacional do Iguagu. IBDF-DN, Brasilia, 104 pp. Quigley, H. B. and Crawshaw, P. G., Jr. 1992. A conservation plan for the jaguar Panthera onca in the Pantanal region of Brazil. Biol. Conserv. 61: 149-157. Quigley, H. B. and Crawshaw, P. G., Jr. In prep. Reproduction, growth, and dispersal of jaguar in the Pantanal. Typed manuscript, 17 pp. Quigley, H.B., Garshelis, D., Pelton, M. , Taylor, C. & Villlarrubia, C. 1979. Use of activity monitors in telemetry studies. Proc. Second Intl. Conf. Wildl. Biotelem. , Laramie, Wyoming. Rabinowitz, A. R. 1986. Jaguar predation on domestic livestock in Belize. Wildl. Soc. Bull 14: 170-174. Rabinowitz, A. R. and Nottingham, B. G. 1986. Ecology and behaviour of the jaguar ( Panthera onca) in Belize, Central America. J. Zool. London 210: 149-159. Ralls, K. and Ballou, J. 1983. Extinctions: lessons from zoos. In: Genetics and Conservation: A reference for managing wild animal and plant populations, SchonewaldCox, C. M., Chcunbers, S. M. , MacBryde, B., and Thomas, L. (eds.), Benjamin/Cummings, Menlo Park, California: 164184. Ralls, K., Harvey, P. H., and Lyles, A. M. 1986. Inbreeding in natural populations of birds and mammals. In: Conservation Biology: The Science of Scarcity and Diversity, Soule, M. E. (ed.), Sinauer Assoc., Massachusetts: 35-56.

PAGE 201

188 Redford, K. H., and Eisenberg, J. F. 1992. Mammals of the Neotropics The Southern Cone, Volume 2. The University of Chicago Press, Chicago, 430 pp. Robinson, J. G. and Redford, K. H. 1979. Body size, diet, and population variation in neotropical forest mammals species: predictors of local extinction? In: Advances in Neotropical Mammalogy, Redford, K. H. and Eisenberg, J. F. (eds.). Sandhill Crane Press, Gainesville, FL: 567594. Schaller, G. B. and Crawshaw, P. G., Jr. 1980. Movement patterns of jaguar. Biotropica 12 (3): 161-168. Seymour, K. L. 1989. Panthera onca. Mammalian Species, 340: 1-9. Shields, W. M. 1987. Dispersal and mating systems: Investigating their causal connections. In: Mammalian Dispersal Patterns. Chepko-Sade, B. D. and Halpin, Z. T. (eds.). University of Chicago Press, Chicago and London: 3-24. Stviwe, M. 1985. Manual for the Micro-Computer Program for Animal Locations MCPALL. National Zoo, Wahington, DC. Sunquist, M. E. 1992. The ecology of the ocelot: The importance of incorporating life history traits into conservation plans. In: Felinos de Venezuela: biologia, ecologia y conservacion, edited by Fudeci, Caracas. Sunquist, M. E., Sunquist, F., and Daneke, D. E. 1989. Ecological separation in a Venezuelan llanos carnivore community. In: Advances in Neotropical Mammalogy, Redford, K. H. and J. F. Eisenberg (eds.). The Sandhill Crane Press, Inc., Gainesville, FL: 197-232. Tewes, M. E. 1986. Ecological and behavioral correlates of ocelot spatial movements. Ph. D. Dissertation, University of Idaho, Moscow, ID, 128 pp. Trent, T. T. and Rongstad, 0. J. 1974. Home range and survival of cotton-tail rabbits in southwestern Wisconsin. J. Wildl. Manage. 48 (3): 459-472. Townsend, W. 1995. Living on the edge: Siriono hunting and fishing in lowland Bolivia. Ph. D. Dissertation, University of Florida, Gainesville, FL, 169 pp. Watt, E. M. 1987. A scatological analysis of parasites and food habits of jaguar (Panthera onca ) in the Cockscomb

PAGE 202

189 Basin of Belize. M. S. Thesis, Univ. Toronto, Toronto, Ontario, 90 pp. White, G. and Garrot, R. 1991. Analysis of wildlife radiolocation data. Academic Press, New York, 383 pp. World Wildlife Fund. 1991. Maintenance of a wildlife corridor of Atlantic forest in the Province of Missiones, Argentina. Project proposal, 16 pp.

PAGE 203

BIOGRAPHICAL SKETCH Peter Gransden Crawshaw, Junior was born on 9 January 1952 in Sao Vicente, Sao Paulo state, Brazil. He graduated in Biological Sciences at the Universidade do Vale do Rio dos Sinos, Sao Leopoldo, Rio Grande do Sul, in December 1977. As his under-graduate thesis, he studied the biology of the white-eared opossum ( Didelphis albiventris ) . In January 1978, he was hired by the Brazilian Institute for Forest Development IBDF (presently the Brazilian Institute for the Environment and Renewable Natural Resources IBAMA) to work as the Brazilian counterpart in a project led by Dr. George Schaller, to study jaguar ecology in the Pantanal of Mato Grosso, Brazil. For the next seven years, he worked on different aspects of the biology of jaguars, pumas, and ocelots, as well as caimans and capybaras. In August 1985, he joined the University of Florida's Wildlife and Range Sciences Department and completed his M. S. in December 1987. His thesis focused on the reproductive biology of the Paraguayan caiman (Caiman yacare) in the Pantanal. In 1988, he started his Ph. D. program. He is now the head of IBAMA 's National Center for the Conservation of Predators (CENAP) , created in July 1994. He is married to Mara Teresa Crawshaw and has three children: Danielle, 20, Beatriz, 17, and David Edward, 11. 190

PAGE 204

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Melvin E. Sunquist, Cl^air Associate Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. George W. Tanner Associate Professor of Forest Resources and Conservation

PAGE 205

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Zoology and Geology 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, 1995 _ Dean, ©blleqe of Agriculture I, febllege of Agricull Dean, Graduate School