trolle 2005 - camera trap study of ocelot and other secretive mammals in the northern pantanal

8/14/2019 Trolle 2005 - Camera Trap Study of Ocelot and Other Secretive Mammals in the Northern Pantanal

1/8

405MAMMALIA 2005 69 (3-4) Publications Scientifiques du Musum national dHistoire naturelle, Paris.

Camera-trap study of ocelot

and other secretive mammals

in the northern Pantanal

M. TROLLENeowildlife Consultancy, Pile All 19A 2tv, 2000 Frederiksberg (Denmark)

[email protected]

Marc KRY

Patuxent Wildlife Research Center, U.S. Geological Survey, 11510 American Holly Drive,

Laurel, MD 20708-4017 (USA); present address: Swiss Ornithological Institute, CH 6204

Sempach (Switzerland)

Trolle M. & Kery M. 2005. Camera-trap study of ocelot and other secretive mammals inthe northern Pantanal. Mammalia 69 (3-4) : 405-412.

ABSTRACT

Reliable information on abundance of the ocelot (Leoparduspardalis) isscarce. We conducted the first camera-trap studyin the northern part of thePantanal wetlands of Brazil, one of the wildlife hotspots of South America.Using capture-recapture analysis, we estimated a densityof 0.112 independ-ent individuals per km2 (SE 0.069). We list other mammals recorded withcamera traps and show that camera-trap placement on roads or on trails hasstriking effects on camera-trapping rates.

RSUM

tude parpige photographique docelotsetautresmammifrescachsdans lePantanal nord. Bien que locelot (Leoparduspardalis) soit relativement com-mun dans les plaines tropicales dAmrique du Sud, peu de donnes sont dis-ponibles sur labondance de ce flid. Nous prsentons ici les rsultats de lapremire tude par pige photographique publie pour la partie nord deszones humides du Pantanal au Brsil, lun des sites les plus importantsdAmrique du sud pour la faune sauvage. En combinant lutilisation depiges photographiques avec des analyses de capture-recapture, nous avonsestim une densit docelots de 0.11 individus indpendants par km2 (SE0.069). Nous donnons aussi une liste dautres mammifres enregistrs parpige photographique. Finalement, nous montrons que le choix des emplace-

ments des piges, sur des routes ou sur des sentiers, a un effet marqu sur letaux de capture photographique.

KEY WORDSBrazil,

camera trapping,density,Felidae,

Leopardus,Neotropics,

ocelot,Pantanal.

MOTS CLSBrsil,

pige photographique,densit,Flidae,

Leopardus,Neotropics,

ocelot,Pantanal.


2/8

Trolle M. & KeryM.

406 MAMMALIA 2005 69 (3-4)

INTRODUCTION

Reliable information on the abundances of theocelot is surprisingly scarce (see Oliveira 1994;Nowel & Jackson 1996; Murray & Gardner1997). Until recently, the very fewpublisheddensityestimates for ocelot from South America

were based on radio-telemetry studies (Ludlow&Sunquist 1987; Emmons 1988). This methodhas a number of limitations for estimating popu-lation density(Karanth 1995, 1999).Trolle & Kry(2003) demonstrated that ocelotsare individually identifiab le from their bodymarkings and estimated density using a capture-

recapture method. This efficient, newmethod ofcombining camera-trapping with capture-recap-ture models has been used increasingly for anumber ofsecretive mammalian species (Karanth1995; Karanth & Nichols 1998; Carbone etal.2001; Maffei etal. 2002, 2004; Noss etal. 2003,2004; Trolle & Kry 2003; Silver etal. 2004).The principal aim ofthis study was to conduct alarger-scale investigation on ocelot densityin thenorthern Pantanal and continue the development

and evaluation ofthe camera-trap methodology.Add itio nally, we presen t he re data on othe rsecretive mammals. Such information is ofgreat importance in planning future studies andconservation measures in the Pantanal region.Finally, we invest igate how the placement oftraps on roads or on trails influences camera-trap-ping rates.

STUDY SITEThe study took place in the 106,000 ha privatereserve Estncia Ecolgica SESC Pantanal situ-ated between the Rio Cuiab river and the tribu-taryRio So Loureno in the Baro de Melgaoregion, the north-eastern Pantanal, Mato Grosso,Brazil. Astudyarea ofapproximately54 km2waschosen in the north-eastern corner of the reservebetween the field stations of Santa Maria by theRio So Loureno (1642.655S; 5601.648W)and So Luiz (1641.201S; 5610.486W). Thisparticular area was chosen partly because the

hab itats were represen tative for the area andpartlyfor logistical reasons.

A mosaic of closed and open as well as perma-nentlydryand seasonallyinundated habitat types

was found in the studyarea , including galleryforest, semideciduous forest with the understoreydominated byacur palms (Scheelea phalerata),Cerrado woodland, scrubland , and grassland.During the rainy season (usuallya three-monthperiod around December through February),large areas flood. However, by the end ofthe dryseason, when this study took place (Oct-Dec2002), no permanent water occurred in the studyarea, except for a number ofartificial water holes.Typical of the Pantanal (Trolle 2003), whichconsists mainly of private catt le ranches, the

study area has been affected considerab ly byvarious ranching-related activities. However,since the establishment of the reserve in 1998these influences have been diminished and subse-quently the natural succession has begun, giving ahigh coverage of grass, dense scrub, and thick for-est undergrowth.

MATERIALS AND METHODS

SAMPLINGA standardized camera-trapping protocol appro-priate for the ocelot was used (i.e. ensuring that atleast one trapping station was present in eachhome range). The general framework of ourstudydesign followed Karanth & Nichols (1998)and Nichols & Karanth (2002). The studyarea

was divided into four sub-areas. In each sub-area,14 trapping stations were camera trapped fornine consecutive nights. Nine capture occasions

were defined by the first to ninth dayoftrappingin each ofthe four subareas.We covered the studyarea with a grid configura-tion, with trapping stations spaced approximatelyone km apart. We selected sites for the camerasthat seemed promising, using animal signs asindicators. The trapping stations were placedmainly by dirt roads (car tracks) and ani maltrails. In total, 29 of the trapping stations wereplaced along roads and 27 on animal trails.

Fourteen passive infrared trail monitors wereused (TrailMaster model TM550; Goodson &


3/8

Associates, Inc., Lenexa, Kansas) with adapted,automatic, weatherproof 35-mm Ya shica or32-mm Canon cameras with automatic flash(TrailMaster model TM35-1). In order to gainphotographic records of both flanks of animals,each trap was equipped with two cameras. Thetraps were set up and programmed as describedin Trolle (2003). Sardines in oil were applied asbait. The units were programmed to take photosbetween 1800 h0500 h (approximatelybetweensunset and sunrise). In the Pantanal, this avoids

wasting the film within a fewdays on agoutis(Dasyprocta), birds, and sun-triggered photos(Trolle 2003).

DATA ANALYSISWe used capture-recapture analysis to estimatepopulation size for ocelots. The temporal patternof sighting/non-sighting of individual ocelotscontains information on the population size. Keyto inference about population size is detectionprobability, i.e. the probability with which anocelot present in the studyarea is photographeddur ing one ca pture occasion. We assume d a

closed population, i.e. that there were no numeri-cal changes in the population during the studyperiod, and used CAPTURE (Otis etal. 1978) toestimate population size.Although the closure assumption appeared rea-sonable with a studyduration of only1.5 month,

we tested this by the closure test in programCAPTURE and by a l ikelihood ratio testbetween a constrained and an unconstrainedCormack-Jolly-Seber model in program MARK(White & Burnham 1999). The latter proceduretests ifa model assuming an open population fitssignificantlybetter than one where the popula-tion is assumed to be closed, i.e. where survivalrate is constrained to be equal to 1. If it does not,this is taken as evidence that the closure assump-tion is tenable.Program CAPTURE provides est imators forseven models that make different assumptionsabout sources of variation of detection probabil-ity: M0, Mb, Mt, Mh, Mth, Mbh and Mtb. See

Otis etal. (1978) and Karanth & Nichols (1998)for full details on these models. We used good-

ness-of-fit tests, between-model tests and theCAPTURE model-selection rout ine to selectthe appropriate model for population-size estima-tion.We estimated densityof ocelots in the studyareabydividing the estimated population size by theestimated effective trapping area as described byKaranth & Nichols (1998). However, here weused both the full mean maximum di stancemoved (MMDM) and half the MMDM todefine boundary strip width. MDM is the maxi-mum distance between two trap locations for ananimal trapped at more than one location. It isviewed as an approxim ation to home range

diameter. The boundary strip is added to thecore trapping-area pe rimeter (the minimumconvex polygon defined by trap locations) toyield the total, effective trapping area from whichthe captured ani mals are likely to have beendrawn from. Arecent field test of density-estima-tion methods for small mammals has shown thefull MMDM to be superior to the halfMMDMmethod (Parmenter etal. 2003). However, deter-mination of boundary-strip width is a thorny

issue, so for better comparison among studies, weprovide three densityestimates pertaining to (1)the core area only, and adding a strip of (2) halfand (3) the fu l l MMDM . See Karanth &Nichols (1998) for formulae of estimators andtheir variances.We used log-linear models to test if camera-trap rates were the same on roads and on trails.For each species with sufficient sample size, weused a Chisquare test to see if the numbers ofcaptures on roads reflected expectations basedon the number of trap-nights on roads (261)versus on trails (243). These analyses wereconducted in the programme GenStat (Payneetal. 1993).

RESULTS

The 56 camera-trapping stations covered a mini-mum-convex polygon (MCP) area of 53.72 km2.

The total camera-trapping effort was 504 camera-trapping nights.

Running head: Ocelot densityin the northern Pantanal

407MAMMALIA 2005 69 (3-4)


4/8

OCELOT DENSITYSixteen captures of ocelot were obtained, out of

which 14 captures rendered photos of a qualitythat allowed individual recognition. Nine indi-vidual ocelots could be recognized, including sixfemales and two males. There was no evidencefor a violation of the closure assumption (test inCAPTURE: z= - 0.92, P = 0.18); test in MARK:1

2, P = 0.30). As selected byCAPTURE, webased est imation on the constant model M0.Detection probabilityper occasion was estimated

at 0.127 and, hence, over the entire study, at0.706. Populatio n size est im ate was 12 (SE3.351) with a 95% CI ranging from 1026.Only two animals were trapped at more than onetrap station. Maximum distances moved were 0.6and 5.1 km, respectively. This yielded a mean-maximum distance moved (MMDM) of2.85 km(SE 2.25 km). Using half the MMDM to definethe width of the boundary strip around the coreminimum-convex polygon area defined by thetrap locations yielded a de nsity est imate of

Trolle M. & KeryM.

408 MAMMALIA 2005 69 (3-4)

TABLE 1. Estimation of ocelot population size and density.

We use three different methods of obtaining the trapping area, MCP, half and full MMDM. This refers to the minimal convex polygondefined by the trap locations and to the area obtained by adding a boundary strip with width corresponding to half or the entiremean maximum distance moved between different captures of the same individual. The area is given in km2 and density as theestimated number of independent ocelots per km2.

Method Area (km2) (SE) (SE)

MCP (no boundary strip) 53.72 12 (3.35) 0.2234 (0.0624)

plus half MMDM 107.22 12 (3.35) 0.1119 (0.0685)

plus full MMDM 173.23 12 (3.35) 0.0693 (0.0624)

TABLE 2. Species recorded by camera trapping.

Didelphis albiventris White-eared opossumDasypus novemcinctus Nine-banded long-nosed armadilloEuphractus sexcinctus Six-banded armadillo

Priodontes maximus Giant armadilloMyrmecophaga tridactyla Giant anteater

Tamandua tetradactyla Southern tamandua

Cebus apella Brown capuchin monkeyCerdocyon thous Crab-eating fox

Chrysocyon brachyurus Maned wolfNasua nasua South American coatiProcyon cancrivorus Crab-eating raccoon

Leopardus pardalis OcelotPuma concolor Puma

Tapirus terrestris Brazilian tapirPecari tajacu Collared peccaryTayassu pecari White-lipped peccary

Blastoceros dichotomus Marsh deerMazama americana Red brocket deer

Mazama gouazoubira Brown brocket deerHydrochoeris hydrochaeris Capybara

Cuniculus paca Paca

Dasyprocta azarae Azaras agoutiThrichomys apereoides Spiny ratSylvilagus brasiliensis Brazilian forest rabbit

Note: When available, scientific names in this paper follow Voss et al. (2001) and common names Emmons & Feer (1997). Additionalscientific and common names follow Eisenberg & Redford (1999).


5/8

ocelots in the studyarea of 0.112 (SE 0.069)animals per 5 km2 (Table 1). The boundary-strip width chosen stronglyaffects the densityestimate. Therefore, we also present densityesti-mates that pertain to just the core MCP area (noboundary strip) and to an effective area with thefull MMDM added.

DATA ON ADDITIONAL MAMMAL SPECIESA total of 24 large to medium-sized mammalspec ies were recorded by the camera traps(Table 2), manyof which are secretive and littleknown from the Pantanal. Worthwhile mention-ing is also the fact that burrows and digging

activity of Priodontesmaximuswere frequentlyfound all over the studyarea.

EFFECT OF TRAP LOCATION ON OR OFF ROADS ONCAMERA-TRAPPING RATESAbout half of the trapping stations were placedon dirt roads (51.8%) and about half on animaltrails (48.2%). Camera-trapping rates on roads

were significantlydifferent from those on trails(Table 3; log-linear model: 11

2, P < 0.001). Forexample, 96% of202 captures ofcarnivores wereon roads. In sharp contrast, only 26% of 46 cap-tures oftapirs were on roads. Indeed, none ofthespecies with sufficient sample size were caught inproport ional numbers on roads and on trails(Table 4).

DISCUSSION

BIOLOGICAL CONSIDERATIONSThe ocelot density found in our north-eastern

Pantanal study site, 0.112 (SE 0.069) animalsper km2, was much lower than in the south-eastern Pantanal (0.564 (SE 0.201); Trolle &Kry 2003). Longer-term studies are needed toconfirm whether this large difference is perma-ne nt or maybe a resu lt of ocel ot po pulatio nfluctuations (e.g., due to fluctuating preyavail-ability). However, the results indicate that there


409MAMMALIA 2005 69 (3-4)

TABLE 3. Camera-trapping rates (ctr) from two Pantanal sites.

NE Pantanal (this study) SE Pantanal (Trolle 2003)

Ctn 261/243 412Trapping stations 29/27 (roads/trails) 33Trapping hours dusk-dawn dusk-dawn

ctr ctr

roads trails roads/cattle trails

Didelphis 0.41 Dasypus 0.77 1.23 2.91Priodontes maximus 0.77 0.82 Myrmecophaga tridactyla 0.77 0.41 0.73Cerdocyon thous 38.70 2.47 7.52Chrysocyon brachyurus 3.83 Dusicyon vetulus Procyon cancrivorus 22.22 0.41 8.50Leopardus pardalis 5.75 0.41 6.80Leopardus tigrinus 0.24Leopardus wiedii 0.49Panthera onca 0.24Puma concolor 3.45 0.41 1.45Tapirus terrestris 4.60 13.99 2.91Cuniculus paca 1.65 Sylvilagus brasiliensis 0.77 1.65


6/8

maybe large differences in ocelot density from

area to area in the Pantanal.Tapirus terrestris, Cerdocyon thous, and Procyoncancrivorusall had comparativelyhigh trappingrates, and are undoubtedly relativelyabundant inthe studyarea (Table 3; MT, pers. obs. from theBrazilian and PeruvianAmazon). Priodontesmax-imus seems much more common in the SESCarea than in our study site in the south-easternPantanal (Trolle 2003). Finally, neither Didelphisalbiventris, Chrysocyon brachyurus, nor Sylvilagusbrasiliensiswere found to occur in the southernstudy site (Trolle 2003).

METHODOLOGICAL CONSIDERATIONSDetermination of effectivetrappingareaDetermination of strip width and its effect ondensityestimates needs further study. The sug-gestion of Karanth & Nichols (1998), to definestrip width byan estimate of the radius of thehome range (i.e. half the MMDM), is based onsmall mammals. It is unclear ifthis is appropriate

for large territorial carnivores. In a recent study,adding a full MMDM strip width performed bet-

ter at estimating density(Parmenter etal. 2003).

Hence, we present three different estimates ofdensity. Data to determine strip width mayalsocome from outside a study, such as from a sepa-rate radio-telemetry study. Given the results byParmenter etal. (2003) we would perhaps preferto calculate density with the ful l MMDMboundary strip width added. See also the discus-sion in Noss etal. (2003).

Size ofstudyareaOur density-est imation method breaks down

when there are no animals captured at multiplelocations. Both our ocelot data sets were smalland just reached the minimal required samplesize for computing an MMDM plus associatedmeasure of imprecision (SE). Our small samplesizes are reflected in large confidence intervals.While these are undesirable, they still showoneadvantage ofthis method, viz. that imprecision inthe component estimates of density is properlyaccounted for.

Obviously, trapping in a large area helps withregard to both of these issues. With increasing

Trolle M. & KeryM.

410 MAMMALIA 2005 69 (3-4)

TABLE 4. Comparison of trapping rates on roads and on trails for selected species. For each species is given the number ofcaptures (cap), the camera-trapping rates (ctr = number of captures per 100 trapping nights), the percentage of trapping stationsthat recorded the species (% trap stations), and, for each species, the chisquare test statistic and p-value for a test of associationbetween observed and expected numbers of captures on and off-roads. The Null hypothesis assumes that species are captured onroads and on trails according to the proportion of trap-nights (261/243). All tests have one d.f.

Overall Traps on roads Traps on trailsTest for

association

Trapping stations 56 29 27Camera-trapping 504 261 243 Chisquare p-nights value

cap ctr% trap

cap ctr% trap

cap ctr% trap

stations stations stations

P. maximus 4 0.79 7% 2 0.77 7% 2 0.82 7%

C. thous 107 21.2 50% 101 38.7 86% 6 2.47 11% 102.10 < 0.001

C. brachyurus 10 1.98 18% 10 3.83 34% 0 0 0% 13.86 < 0.001

P. cancrivorus 59 11.7 45% 58 22.2 83% 1 0.41 4% 71.65 < 0.001

L. pardalis 16 3.17 18% 15 5.75 31% 1 0.41 4% 14.70 < 0.001

P. concolor 10 1.98 18% 9 3.45 31% 1 0.41 4% 7.36 0.007

T. terrestris 46 9.13 39% 12 4.60 24% 34 14.0 56% 10.96 < 0.001

For Priodontes maximus, no test of association is shown for small sample size.


7/8

core/boundary area ratio , the choice of stripwidth becomes less influential for densityestima-tion. In addition, larger areas are likely to containmore animals, so that strip width can be esti-mated with greater precision.Finally, we would recommend that sub-areas arecamera trapped for longer than the nine days weused in this study. This will heighten the chanceofrecaptures, giving better estimations.

Effectoftrap location on oroffroadson camera-trappingratesThese results have significant methodologicalimplications. There is a controversyamong lead-

ing camera-trapping researchers over the use ofcamera-trapping rates to assess densities of secre-tive mammals that are not individually identi-fiable (Carbone etal. 2001, 2002; Jennelle etal.2002). This debate is important, because cameratrapping is likely to be used increasingly to pro-vide data for conservation planning and assess-ment. Our results show that the placement ofcamera traps in relation to roads in a studyarea(something that will often varybetween studies)

is an essential covariate to take into considerationwhen trying to compare camera-trapping ratesbetween studies and sites.

Acknowledgements

The study was a collaborat ion betweenZoologica l Museum , Univers ity ofCopenhagen, Museu Nacional, UniversidadeFederal do Rio de Janeiro, and SESC Pantanal.The project was made possible by the generoussupport of SESC Pantana l , WWF-Denmark/Novo Nordisk (Biodiversity Grant),Dronning Margrethe og Prins Henriks Fond,Beckett-Fonden, Torben og Ali ce FrimodtsFond, Inge og Skjold Burnes Fond, IngenirK.A. Rohde og hustrus legat, Fonden Kjebi, andFujifilm Denmark. Many thanks are due to JooAlves de Oliveira, Luiz Flamarion de Oliveira,Jos Luiz Cordeiro, Hans J. Baage, MogensAndersen, and MaryWisz . We apprec iate

helpful advice and comments on earlier draftsof thi s pape r byJ. D. Nicho ls, U. Ka ranth,

M. C. Runge, C. Peres, and two anonymous ref-erees. We also thank N. Espuno for providingthe French summary. MK was funded bygrant81ZH-64044 from the Swiss National ScienceFoundation.

REFERENCES

CARBONE S., CRISTIE S., CONFORTI K., COULSON T.,FRANKLIN N., GINSBERG J. R., GRIFFITHS M.,HOLDENJ., KAWANISHI K., KINNAIRD M., LAIDLAWR., LYNAM A., MACDONALD D. W., MARTYR D.,MCDOUGAL C. , NATH L., OBRIEN T.,SEIDENSTICKER J., SMITH D. J. L., SUNQUIST M.,TILSON R. & WAN SHAHRUDDIN W. N. 2001. The use of photographic rates to estimate densitiesof tige rs and othe r cryptic mamm als. Ani malConservation 4: 75-79.

CARBONE S., CRISTIE S., CONFORTI K., COULSON T.,FRANKLIN N., GINSBERG J. R., GRIFFITHS M.,HOLDEN J., KINNAIRD M., LAIDLAWR., LYNAM A.,MACDONALD D. W., MARTYR D., MCDOUGAL C.,NATH L., OBRIEN T., SEIDENSTICKERJ., SMITH D.

J. L., TILSON R. & WAN SHAHRUDDINW. N. 2002. The use of photographic rates to estimate densi-ties ofcryptic mammals: a response toJennelle etal.AnimalConservation 5: 121-123.

EISENBERG J . F. & REDFORD K. H. 1 9 9 9 . Mammalsofthe Neotropics, Vol. 3, The CentralNeotropics: Ecuador, Peru, Bolivia, Brazil. UniversityofChicago Press, Chicago : pp

EMMONS L. H. 1988. Afield studyof ocelots (Felispardalis) in Peru. Revue dEcologie(Terre etVie) 43:133-157.

EMMONS L. H. & FEE R F. 1997. Neo trop ica lRainforestMammals: A FieldGuide. 2nd edition ,UniversityofChicago Press, Chicago: pp.

JENNELLE C. S., RUNGE M. R. & MACKENZIE D. I.2002. The use of photographic rates to estimatedensities of tigers and other cryptic mammals: acomment on misle ading con clus ion s. Ani malConservation 5: 119-120.

KARANTH K U. 1995. Estimating tiger Pantheratigrispopulations from camera trap data using cap-ture-recapture models. Biological Conservation 71:333-338.

KARANTH K.U. 1999. Counting tigers, with confi-dence , in SEIDENSTICKER J. , CHRISTIE S. &

JACKSON P. (Eds): Riding the Tiger: TigerCon servation in Human-dominated Landscapes.Cambridge UniversityPress, Cambridge: 350-53.

KARANTH K. U. & NICHOLS J. D. 1 9 9 8 . E

stimation of tiger densities in

Indi

ausing pho-tograph ic captures and recaptures. Ecology79:

2852-2862.


411MAMMALIA 2005 69 (3-4)


8/8

KARANTH K. U. & NICHOLS J. D. 2002. Monitoring Tigers and their Prey: a Manual forResearchers, Ma nagers and Conservationist s inTropica l Asia. Centr e for Wi ldlife Studies ,Bangalore: pp.

LUDLOWM. E. & SUNQUIST M. E. 1987. Ecologyand behavior of ocelots in Venezuela. NationalGeographic Research 3: 447-461.

MAFFEI L., CULLAR E. & NOSS A. 2002. Uso detrampas-camara para la evaluacion de mamiferos enel ecotono Chaco-Chiquitania. Revista Boliviana deEcologia yConservacion Ambiental11: 55-65.

MAFFEI L., CULLAR E. & Noss A. 2004. Onethousand jaguars (Pa nthera on ca ) in Boli viasChaco? Camera trapping in the Kaa-Iya NationalPark. Journal ofZoology262: 295-304.

MURRAYR. L. & GARDNERG. L. 1997. Leoparduspardalis. Mammalian Species548: 1-10.

NICHOLS J. D. & KARANTH K. U. 2002. Statisticalconcepts: Estimating absolute densities of tigersusing capture-recapture sampling, in KARANTHK. U. & NICHOLS J. D. 2002. Monitoring Tigersand theirPrey: A Manual forResearchers, Managersand Conservationists in Tropical Asia. Centre for

Wildlife Studies, Bangalore, India: 121-137.NOSS A. J., PEAR. & RUMIZ D. I. 2004. Camera

trapping Priodontesmaximus in the dry forests ofSanta Cruz, Bolivia. Endangered SpeciesUpdate21:43-52.

NOSS A. J., CULLARR. L., BARRIENTOS J., MAFFEI L.,CULLAR E., ARISPE R., RUMIZ D. I., & RIVERO K.2003. A camera trapping and radio telemetrystudyof Tapirus terrestris in Bolivian dry forests.Tapirconservation 12: 24-32.

NOWEL K. & JACKSON P. 1996. Wild Cats: StatusSurveyand Conservation Action Plan. IUCN/SSCCat Specialist Group, Gland: pp.

OLIVEIRA. DE T. G 1994. NeotropicalCats: EcologyandConservation. Edufma, So Lus.

OTIS D. L., BURNHAM K. P,. WHITE G. C &ANDERSON D. R. 1978. Statistical inferencefrom capture data on closed animal populations.WildlifeMonograph 62: pp.

PARMENTER R. R., YATES T. L., ANDERSON D. R.,BURNHAM K. P., DUNNUM J. L., FRANKLIN A. B.,FRIGGENS M. T., LUBOWB. C., MILLERM., OLSONG. S., PARMENTER C. A., POLLARD J., REXSTAD E.,SHENK T. M., STANLEY T. R. & WHITE G. C.2003. Small-mammal densityestimation: Afieldcomparison of grid-based vs. web-based densityestimators. EcologicalMonographs73: 1-26.

PAYNE R. W., LANE P. W., DIGBY P. G. N., HARDINGS. A., LEECH P. K., MORGAN G. W.,. TODD A. D,THOMPSON R., TUNNICLIFFE WILSON G., WELHAMS. J. & WHITE R. P. 1993. Genstat5, Release3.ReferenceManual. Clarendon Press, Oxford: pp.

SILVER, C. S., OSTRO L. E. T., MARSH L. K., MAFFEIL., NOSS A. J., KELLY M., WALLACE R. B., GOMEZH. & AYALA G. 2004. The use of camera trapsfor estimating jaguar Panthera onca abundance anddensity using capture/recapture analysis. Oryx38:148-154.

TROLLE M. 2003. Mammal survey in the south-eastern Pantana l , Brazi l . Biodiversity andConservation 12: 823-836.

TROLLE M. & KryM. 2003. Ocelot densityesti-mation in the Pantanal using capture-recaptureanalys is of camera-trapping data . Journa l ofMammalogy84: 607-614.

VOSS R. S., LUNDE D. P. & SIMMONS N. B. 2001. The mamma ls of Paracou, French Guiana: aNeotrop ica l lowland rainforest fauna. Pa rt 2.Nonvolant species . Bul letin of the AmericanMuseum ofNaturalHistory263: 1-236.

WHITE G. C. & BURNHAM K. P. 1999. ProgramMARK: survival est imation from populatio nsof marked animals. Bird Study46 (Supplement):120-138.

Trolle M. & KeryM.

412 MAMMALIA 2005 69 (3-4)

Submitted on 9April2004;Accepted on 25March 2005.

trolle 2005 - camera trap study of ocelot and other secretive mammals in the northern pantanal

Documents