Trolle 2003 - Mammal Survey In Western Pantanal

Biodiversity and Conservation 12: 823–836, 2003.  2003 Kluwer Academic Publishers. Printed in the Netherlands.

Mammal survey in the southeastern Pantanal, Brazil MOGENS TROLLE Mammal Department, Zoological Museum, University of Copenhagen, Universitetsparken 15, DK-2100 Copenhagen, Denmark (e-mail: mtrolle@ zmuc.ku.dk; fax: 145 -3532 -1010) Received 2 January 2002; accepted in revised form 8 May 2002

Key words: Brazil, Camera trapping, Conservation, Habitat use, Mammal inventory, Pantanal Abstract. The Pantanal floodplain of southwestern Brazil, with its mosaic of marshes, seasonally flooded savannas, woodlands and forests, is known as one of the wildlife hotspots of South America. The region harbors a rich mammal fauna and is a stronghold for a number of mammal species; however, still very few thorough mammal surveys have been conducted in the Pantanal. This is the first mammal survey from the southeastern part of the region. An intensive, three-month study of the medium to large, non-volant species was conducted at a location in the upper Rio Negro Basin, Aquidauana region, Mato Grosso do Sul. Thirty species of mammals were recorded, including 5 xenarthrans, 1 primate, 10 carnivores, 7 ungulates and 5 caviomorph rodents. It was possible to analyze the field biology of many of the species to a level not previously published for the Pantanal. Recommendations are given for camera trapping in the Pantanal and the Amazon. Finally, the effects of intensive cattle ranching on conservation in the Pantanal are discussed.

Introduction The Pantanal floodplain of southwestern Brazil, covering approximately 140 000 km 2 , is known for its unique abundance of conspicuous wildlife. The region has a rich mammal fauna and is a stronghold for a number of attractive larger species (Alho et al. 1988b; Mittermeier et al. 1990; Alho and Lacher 1991). Still, only a couple of thorough general mammal surveys have been published from this major region (Schaller 1983; Alho et al. 1988a). In order to assess the conservation role of the Pantanal for mammals, and to plan an adequate conservation strategy for the region, additional knowledge of the demographics, abundance and field biology of the mammal fauna is of great importance. An intensive, three-month mammal survey was conducted (May–July 2001). The main purpose of this paper is to present information on the fauna of medium to large, non-volant species found in the study area. This is the first mammal survey from the southeastern Pantanal. Camera trapping is a highly efficient technique for mammal surveys; however, only recently have modern camera traps been applied on a larger scale in the Neotropics. Recommendations are given for camera trapping in the Pantanal and the Amazon. Most of the Pantanal consists of cattle ranches. In the study area the intensive cattle ranching was found to have a serious negative impact on the environment. Cattle ranching and its effects on conservation in Pantanal are discussed.

824 Study area ´ The survey encompassed Fazenda Sta Emılia and adjacent areas. Fazenda Sta ´ (headquarters at GPS position 19830.4039 S; 55836.7919 W) is a research / Emılia conservation reserve of Universidade para o Desenvolvimento do Estado e da ˜ do Pantanal – UNIDERP of Campo Grande. The area is situated in the Regiao upper Rio Negro Basin of the southeastern part of the floodplain, Aquidauana region, Mato Grosso do Sul. The core area of the study covered approximately 4 3 5 km. The climate of the Pantanal is highly seasonal, with warm, wet summers, cold, dry winters, and an annual flooding of the open areas following the rainy season (described in more detail by Schaller 1983). During the first two months of the study period much of the study area was flooded. Only in the last month of the study the water level started to fall substantially and the area began to dry out. Habitat types As is typical for the Pantanal, the study area consists of a mosaic of open and closed, and mesic and xeric habitat types, correlated with the light undulations of the landscape, the seasonal inundation of the open areas, and the soil types. The definitions of major habitat types used in this study were as follows (plant names according to Pott and Pott 1994; Lorenzi 2000; for descriptions of the Pantanal habitats see also Prance and Schaller 1982; Ratter et al. 1988). Acurı´ forest: Semideciduous forest islands with Acurı´ Palms (Schealea phalerata) dominating the understorey (up to a height of around 8–10 m and creating most of the shade), and interspersed with deciduous emergents of 20–30 m (such as Sterculia apetala, Ficus spp., Tabebuia impetiginosa, Plathyhenia reticulata, Hymenaea stigonocarpa, Albizia niopioides, and Enterolobium contortisiliquum). An often impenetrable belt of the thorny bromeliad Gravateiro (Bromelia balansae) generally surrounds these forest islands. Part of the acurı´ forest had an extremely dense undergrowth of the Taboca Bamboo (Guadua cf. angustifolia). Acurı´ forest is found on higher elevations meandering through the open landscape (cordilheiras) and is not subject to flooding. It is the major forest type of the study area. Gallery vegetation: Gallery / riparian / swamp vegetation (forest, woodland, and scrub) along watercourses and at some sites bordering marshes. Typically the palm Bactris glaucescens, Vochysia divergens trees and the bush-like tree Licania parvifolia are structurally dominant components. Subject in varying degrees to seasonal flooding. Gallery forest, with its year-round water supply, has a strong floristic connection to the Amazon (Pott and Pott 1994). Cerrado vegetation: Woodland in a broad sense (i.e. open to closed; when dense ˜ known as cerradao), characterized by typical trees of the Cerrado region (low, gnarled trees with thick, furrowed bark and xeromorphic leaves), and interspersed with scrub. More open cerrado has a high proportion of grass coverage. Cactus (Cereus peruvianus) occurs in this habitat. Not subject to seasonal flooding.

825 Grassland: Commonly found between marshes and more closed habitats. Occurs in vast areas or as narrow grass beaches between marshes and acurı´ forest. Much of the grassland is seasonally flooded. Marsh: Areas dominated by aquatic vegetation that are isolated or connected with water courses / drainage courses, and may dry out seasonally. Various types of marshes are found, from open and low with water hyacinths (Eichornia spp.) to closed with tall reeds, bunchgrass or scrub. ´ was purchased by UNIDERP three years before this study, Fazenda Sta Emılia and in connection with this most of the cattle was removed from the part of the reserve covered by the study. Due to the lack of cattle, in most open areas the grass was tall. The neighboring fazendas (cattle ranches) had a high cattle ranching intensity, which had an evident impact on the habitats: (1) the open areas were often grazed and trampled down to the level of exposing a high proportion of bare sand, and (2) the acurı´ forest was open and with hardly any undergrowth. The southeastern corner of the Pantanal as a whole is characterized by a relatively high cattle ranching intensity and, judging by a satellite photo of the region (Willink 2000), a relatively high degree of habitat alteration. In addition, outside the UNIDERP reserve, burning of the land by the end of the dry season (as evidenced by scorching marks on the trees) and some local hunting of certain native mammal species (although the introduced wild pig Sus scrofa is usually the preferred game of the locals) has apparently been normal practice, at least until recently.

Materials and methods A range of methods were used: camera trapping, census by foot as well as by horse, boat, motor vehicles and airplane, nightspotting, observations of tracks (Becker and Dalponte 1991), feces, burrows and skulls, and interviews. Thirty sites were camera trapped successfully. The total camera trapping effort was about 450 camera trapping nights. The walked census added up to approximately 325 h and 500 km. Six Trailmaster infrared trailmonitors (five passive TM550s and one active TM1550) with TM35-1 camera kits (standard, automatic, weather proof, 35 mm Yashica cameras with auto flash) were used (Goodson and Associates, USA). The active system works with an infrared beam between a transmitter and a receiver, and is triggered when the beam is broken. The passive system consists only of a transmitter that emits a wedge-shaped infrared field, and is triggered by warmblooded animals moving through the wedge. On the passive TM550s I covered the sides of the infrared sensor with tape, leaving only a 1 cm gap in the center (assuring that an animal is in the center of the photo when it triggers the trap). This works well with the following sensitivity settings: P 5 2–3, Pt 5 1–2 (the infrared wedge is divided into a number of ‘windows’. P relates to the number of windows that must be broken by a warm-blooded animal for the trap to be triggered, and Pt is the number of seconds the animal has to break these windows). The advantages of the passive system are that it is cheaper than the active one and easier to set up. The disadvantages are that in an open, tropical area like the Pantanal the trap may be

826 ´ and adjacent areas, and assessment of Table 1. List of mammal species recorded at the Faz Sta Emılia relative abundance.

Xenarthra Dasypodidae

Myrmecophagidae Primates Cebidae Carnivora Canidae Procyonidae Mustelidae

Felidae

Perissodactyla Tapiridae Artiodactyla Tayassuidae Cervidae

Caviomorpha Erethizontidae Hydrochaeridae Cuniculidae Dasyproctidae Echimyidae Small mammals Marmosidae Sigmodontinae

Scientific name

Common name

Total no. of records

Survey methods

Relative abundance

Dasypus novemcinctus Euphractus sexcinctus a Priodontes maximus Myrmecophaga tridactyla Tamandua tetradactyla

Nine-banded Long-nosed Armadillo Six-banded Armadillo Giant Armadillo Giant Anteater Southern Tamandua

16 b 20 b 1 21 11

VCBT VB B VCT VC

C C R C C

Alouatta caraya

Black Howler Monkey

12

VAF

C

Cerdocyon thous Nasua nasua Procyon cancrivorus Eira barbara Pteronura brasiliensis Lontra longicaudis Leopardus pardalis Leopardus tigrinus Panthera onca Puma concolor

Crab-eating Fox South American Coati Crab-eating Raccoon Tayra Giant Otter Neotropical Otter Ocelot Oncilla Jaguar Puma

35 30 29 1 11 1 59 1 1 19

VCT V VCT V VA V VCT C CI CT

C C C U? C ? C ? R? C

Tapirus terrestris

Brazilian Tapir

30

VCTF

C

Pecari tajacu Tayassu pecari Blastocerus dichotomus Mazama americana Mazama gouazoubira Ozotoceros bezoarticus a

Collared Peccary White-lipped Peccary Marsh Deer Red Brocket Deer Brown Brocket Deer Pampas Deer

13 6 52 2 10 3

VC VC VCT C VCT VI

C U C U? C R(s?)

Coendou prehensilis Hydrochoeris hydrochaeris Cuniculus paca Dasyprocta azarae Thrichomys apereoides

Brazilian Porcupine Capybara Paca Azara’s Agouti Punare´

1 94 2 72 Many

V VTF TI VCAT C

? C ? C C

cf. Monodelphis sp. Holochilus sp. or Nectomys sp.

Short-tailed opossum Marsh/Water Rat

Several 1

C T

C ?

When available, scientific names follow Voss et al. (2001) and common names Emmons and Feer (1997). Additional scientific names follow Anderson (1997). Abbreviations: Survey methods: A 5 acoustic record; B 5 burrows; C 5 camera trapping; F 5 feces; I 5 interview; T 5 tracks; V 5 visual record. Assessment of relative abundance: C 5 common; U 5 uncommon; R 5 rare; s 5 seasonal variation in a b ´ local abundance. Observed outside Faz Sta Emılia; observations of burrows excluded.

triggered by shades moving in front of the sensor (e.g. caused by a tree moving in the wind in front of the sun), and that it is difficult for the infrared sensor to ‘see’ warm-blooded animals when the temperature of the environment is high. I got around this by programming the units to work only between the afternoon and morning, which should not be a problem when working with mainly crepuscular and nocturnal species. The advantages of the active system are that even in an open area like the Pantanal it can work both day and night, and that one can easily choose the

827 Table 2. Summary of camera trapping results.

Dasypus novemcinctus Myrmecophaga tridactyla Tamandua tetradactyla Cerdocyon thous Procyon cancrivorus Leopardus pardalis Leopardus tigrinus Panthera onca Puma concolor Tapirus terrestris Pecari tajacu Tayassu pecari Blastocerus dichotomus Mazama americana Mazama gouazoubira Mazama sp.? a Dasyprocta azarae No. of captures No. of species Camera trapping nights (ctn) No. of captures/ctn

Fo (n 5 10)

Ga (n 5 6)

Ce (n 5 10)

Gr (n 5 7)

Total (n 5 33)

Cap.

Sites

Cap.

Sites

Cap.

Sites

Cap.

Sites

Cap.

Sites

2 1 1 3 3 13 – – 1 2 3 1 2 2 – 7 13 54 13 141 0.38

2 1 1 3 1 5 – – 1 2 1 1 2 2 – 3 6

1 – 1 – 2 7 – 1 1 5 – 1 – – – – 11 30 9 59 0.51

1 – 1 – 2 2 – 1 1 2 – 1 – – – – 2

3 – 1 14 10 5 1 – 2 3 – – 1 – – 1 6 47 10 119 0.39

2 – 1 4 4 3 1 – 2 3 – – 1 – – 1 1

6 2 5 14 20 5 – – 2 2 1 – – – 1 – – 58 10 93 0.62

3 2 3 3 6 3 – – 2 2 1 – – – 1 – –

12 3 8 31 35 30 1 1 6 12 4 2 3 2 1 8 30 189 16 412 0.46

8 3 6 10 13 13 1 1 6 9 2 2 3 2 1 4 9

a Photos of Mazama deer where I was unable to identify the species; this category includes a reddish species recorded in forest, but it is not possible to distinguish whether it is M. americana or M. nana. Number of captures (cap.) and sites: Fo 5 acurı´ forest; Ga 5 gallery forest; Ce 5 cerrado; Gr 5 open grassy areas (including grass beaches).

minimum size of the animals that one wants to monitor by setting the altitude of the infrared beam (the reason it is called ‘active’). Trailmonitors were set so that even smaller mammals were recorded. All non-flooded habitats were covered. Most trapping sites were along trails (animal, cattle and human). Bait was often used; mostly sardines in oil that proved to be quite attractive to carnivores (as evidenced by many photographs of animals sniffing or eating the bait).

Results A total of approximately 550 records of medium to large mammals and their habitat use was obtained. The final list presents 30 species, including three small mammals (Table 1). Camera trapping gave around 250 photos and recorded 16 species of medium to large mammals (Table 2), along with a number of small rodents (Sigmodontinae, Echimyidae), mouse opossum (Marmosidae), and eight species of birds. The overall camera trapping success for medium to large mammals was about 50% (i.e., on average each camera trap gave one capture per two nights). Table 3 gives a summary of the habitat observations, and Table 4 the crepuscular and nocturnal activity patterns recorded by camera trapping.

828 Table 3. Summary of habitat observations.

Dasypus novemcinctus Euphractus sexcinctus Priodontes maximus Myrmecophaga tridactyla Tamandua tetradactyla Alouatta caraya Cerdocyon thous Nasua nasua Procyon cancrivorus Eira barbara Pteronura brasiliensis Lontra longicaudis Leopardus pardalis Leopardus tigrinus Panthera onca Puma concolor Tapirus terrestris Pecari tajacu Tayassu pecari Blastocerus dichotomus Mazama americana Mazama gouazoubira Ozotoceros bezoarticus Coendou prehensilis Hydrochoeris hydrochaeris Cuniculus paca Dasyprocta azarae Totals No. of species

Fo

Fo (e)

Ga

Ce

Sc

Gr

1 4 1 5 5 11 3 21 3

1 3

1

6 3

2

6 8

1

4 6

12

8 5 18

Ri

La

4

1

1

1

1

13

7

3 3 7 1 1 2 1

6

Ma

1 1 5

6 1

9 1 4 1 8 10 1

7

1

2 2

4 1

5 7 4 1

2 8

3

2

1 47

1 3

1

28 115 20

6 –

2 12 35 12

15 83 14

3 –

83 14

52

39

3

112 7

49 3

5 2

Number of records in various habitats: Fo 5 acurı´ forest; Fo(e) 5 forest edge; Ga 5 gallery vegetation; Ce 5 cerrado; Sc 5 scrubland; Gr 5 grassland / grass beach; Ma 5 by or in marsh; Ri 5 by or in river / channel; La 5 by or in lake.

The camera trapping gave 55 photos of Ocelot (Leopardus pardalis). An analysis of the photos revealed that individual recognition is possible, which allowed an ´ estimate of density of Ocelot using capture–recapture analysis (Trolle and Kery, unpublished data).

Discussion Interviews with locals indicated that three species not recorded in this survey occurred in the study area: Three-banded Armadillo (Tolypeutes matacus), Bush Dog (Speothos venaticus), and a dwarf brocket (Mazama nana?); however, these are all species that are often confused by locals, and whose presence should be verified. Maned Wolf (Chrysocyon brachyurus) was known from other parts of the southeastern region by the locals, but it is presumably rare. A couple of additional secretive

829 Table 4. Crepuscular and nocturnal activity patterns between 16 P.M. and 8 A.M. recorded by camera trapping. 16–18 Dasypus novemcinctus Myrmecophaga tridactyla Tamandua tetradactyla Cerdocyon thous Procyon cancrivorus Panthera onca Puma concolor Leopardus pardalis Tapirus terrestris Pecari tajacu Tayassu pecari Blastocerus dichotomus Mazama americana Mazama gouazuobira Dasyprocta azarae

18–20

20–22

22–00

00–02

02–04

II

IIII

II

I

I IIII IIII

IIII III

II IIII IIIII

III IIIII IIIIII

II IIIIII

I IIII

III

04–06

06–08

I I

II

IIIIIII II

I I I

III I I

I II I I IIIIII I I

I I

II

I I

IIIIIIIII

mammals could occur in the region. The following species have been listed for the southern half of the Pantanal: White-eared Opossum (Didelphis albiventris), Nakedtailed Armadillo (Cabassous tatouay), Hoary Fox (Pseudalopex vetulus), Grison (Galictis sp.) (probably an edge species that does not reach the study area), Jaguarundi (Herpailurus yaguarondi) (relatively common in other parts of the Pantanal), Geoffroy’s Cat (Oncifelis geoffroyi), and Margay (Leopardus wiedii) (Allen 1916; Alho et al. 1988a; Alho and Lacher 1991; Lourival 1993; PCBAP 1997; personal observation). Pampas Cat (Lynchailurus pajeros) was reported from the northern Pantanal (Allen 1916).

Field biology information Detailed information about something as basic as habitat use and activity patterns of many of the Brazilian mammals is still available from only relatively few sites (Eisenberg and Redford 1999). Such descriptive information is important for the basic understanding of the biology and ecology of the mammal fauna and for the planning of future studies in the area. The use of camera trapping in this study makes it possible to analyze the field biology of many of the species to a level not before published for the Pantanal. In this section a summary of the field observations of each species is given. In the assessment of relative abundance of a number of the species, I have compared the results of this survey with camera trapping and sighting frequencies from 4.5 months of similar field work in the Rio Negro region (Brazil) and the Rio Madre de Dios region (Peru) of the Amazon (Trolle, unpublished data) and sightings during two years in other parts of the Pantanal (Trolle, unpublished data). A summary of the assessments of abundance is given in Table 1.

830 Xenarthra Dasypus novemcinctus: Frequently camera trapped and burrows were common. Recorded in all habitats, but most often in grassland and cerrado. Observed around dusk (4.45 P.M.→), and frequently camera trapped between 8 P.M. and 4 A.M., but never recorded around dawn and during the day. The nocturnal habits may be due to poaching. Unlike the diurnal E. sexcinctus, the meat of the Nine-banded Armadillo is valued for human consumption. Euphractus sexcinctus: Frequently observed (once three individuals were encountered within 15 min of walking), and burrows were common. Recorded both in open areas, forest edges (often seen in thorny bromeliad scrub), and inside acurı´ forest. Seems to be more common inside acurı´ forest than D. novemcinctus. Seems strictly diurnal (most records between 10 A.M. and midday), reflected also by the fact that it was never camera trapped at night. An individual was observed to carry acurı´ palm fruits into its burrow. Priodontes maximus: One old burrow was found in acurı´ forest and the species was never recorded by camera trapping. It had not been encountered recently in the area by the locals and seems to be rare. Myrmecophaga tridactyla: Encountered relatively frequently, traveling through all habitats, even flooded areas and rivers. Only observed in the afternoon (1–6 P.M.) and was rarely recorded by the night-active camera traps (this, however, probably also reflects that it mostly does not follow trails). Tamandua tetradactyla: Observed and camera trapped relatively frequently. Recorded both in all forest types, and traveling on the ground in open areas. Often feeds and rests in Acurı´ palms. Forages both diurnally and nocturnally, but the open area traveling was recorded only at night (6 P.M.–2 A.M.) and at dawn. Primates Alouatta caraya: Relatively common in acurı´ forest; however, less so than in other regions of the Pantanal (personal observation). Troops of up to eight individuals were observed. Carnivora Cerdocyon thous: Camera trapping revealed that this fox was common. Favors open areas, though it was also recorded occasionally in forest. Camera trapped throughout the night (peak between 6 P.M.–2 A.M.), and is probably mainly nocturnal and crepuscular, though it also has some diurnal activity. Frequently travels in pairs; however, often the second individual of a pair was photographed 1–5 min after the first one passed by, which indicates that they look for food separately. Nasua nasua: Common. Forest dwellers mostly encountered in acurı´ forest, but frequently forage in open areas close to cover such as the edge of wetlands. In forest they forage both terrestrially and arboreally, and rest in tall trees and Acurı´ palms.

831 Diurnal. Groups of between 4 and 14 Coatis as well as large solitary individuals were seen. Procyon cancrivorus: The species most often camera trapped and tracks were very common. Nocturnal. Frequently travels in pairs, walking close together. Eira barbara: Was only recorded once; active in early morning in a tall tree. The lack of camera trap photos of the Tayra may reflect that it is diurnal and rarely uses trails; however, it seems uncommon compared to the Amazon. Lontra longicaudis: A single individual was observed in a minor river. Pteronura brasiliensis: At least two groups frequented the part of a minor river covered by the study: a family of 3–4 and a group of seven individuals. In addition, a single individual was observed far up a small, at parts densely vegetated channel. The smaller family was observed travelling through acurı´ forest at some distance from the river. Leopardus pardalis: Nine individuals were recorded by camera trapping: four males, four females and one subadult. Capture–recapture analysis of the camera trapping data gave a density estimate of 2.82 independent individuals / 5 km 2 (SE ´ unpublished data). Ocelots were recorded in all non-flooded 1.00) (Trolle and Kery, habitats. They were recorded throughout the night by camera trapping, and twice observed active during the day. Evidence of home range overlap with individuals of the opposite sex was seen for seven of the eight adult individuals. In no cases were two males caught at the same site, and only in one case were two adult females caught at the same trap. This site was visited various times by the one female, but only once by the second. The scant information indicates overlap of home ranges between sexes, but intrasexual territoriality. A lactating female that traveled widely was recorded on both sides of the minor river. Leopardus tigrinus: Camera trapped once in cerrado. Panthera onca: Only one camera trap photo was obtained of the Jaguar and tracks were never found. A neighboring rancher had not seen the species in the area during more than 50 years, and it seems to be relatively uncommon in the study area. Puma concolor: Camera trapped at six different sites and tracks were common. Unlike the Jaguar seems quite common. Recorded in all non-flooded habitats. Perissodactyla Tapirus terrestris: Frequently camera trapped and tracks were common. Recorded in all habitats, and tracks revealed that it travels widely. Artiodactyla Pecari tayacu: I observed this species more frequently in the study area than in any other parts of the Pantanal I have worked in, and several skulls were found; however, the peccary was camera trapped relatively infrequently. It was recorded in all non-flooded habitats, and forages both in open and closed vegetation. Nocturnal, crepuscular and occasionally diurnal. Seems to typically hide in dense cover during

832 the day and come out to feed in more open areas around dusk. Groups of 4–7 members were encountered, including a female with three young. Tayassu pecari: Uncommonly camera trapped and observed. Herds were recorded both in forest and open habitats close to cover. Nocturnal and diurnal. Blastocerus dichotomus: Common in marshes during the study period. Feeds in marshes and rests in dense waterside habitats. Often observed feeding in the morning, but it is active throughout the day, and was also recorded active at night. Camera trapped travelling through forest and cerrado relatively far from water. Up to four individuals were observed feeding in the same marsh, but never more than one adult male. Mazama americana: Never observed and only camera trapped with certainty twice. Seems relatively uncommon, but this may be partly due to secretiveness and the fact that Mazama deer rarely walk along trails (as evidenced by tracks). Recorded at night inside forest. Mazama gouazoubira: Commonly observed. A species of open areas such as grassland and cerrado, though it may seek shelter in dense vegetation such as forest edges. Diurnal and nocturnal. Ozotoceros bezoarticus: The Pampas Deer, easy to record due to its open ´ during the study period. The habitats, was never observed inside Fazenda Sta Emılia tall grass and / or the flooding of the area may cause this absence. According to locals it can be encountered later in the dry season. Observed frequently in bordering fazendas to the east, both in open grassland and in cerrado landscapes. A male with antlers in velvet was observed in July. Caviomorpha Coendou prehensilis: Observed once during nightspotting, active in a tall tree inside acurı´ forest. Hydrochaeris hydrochaeris: Common around marshes, but particularly abundant along certain parts of a minor river where it meandered through relatively open landscapes (not gallery forest). Around marshes groups of up to around 20 individuals were seen, but the most frequent group size was 4–5. Agouti paca: Tracks of the Paca were recorded twice in gallery vegetation along a marsh, and the rodent had been observed once at night by the river by a ranchhand. Seems relatively uncommon. Dasyprocta azarae: Camera trapping revealed that it was a common species. Recorded most often in forest, but also frequently in cerrado. Was almost only camera trapped after sunrise, and seems strictly diurnal. Although the inventory methods did not allow for absolute estimates of density of the mammals studied (except in the case of Ocelot), the survey showed that the study area is relatively rich in at least the following species: Giant Anteater, Southern Tamandua, Nine-banded Long-nosed Armadillo, Six-banded Armadillo, Crab-eating Fox, South American Coati, Crab-eating Raccoon, Giant Otter, Ocelot, Puma, Brazilian Tapir, Collared Peccary, Marsh Deer, Brown Brocket Deer, ´ Capybara, Azara’s Agouti and Punare.

833 Recommendations for camera trapping in the Pantanal and the Amazon Within recent years the development of light-weight, weather-proof, simple-to-use camera traps with infrared sensors such as the Trailmaster system has made it both much easier and economically more feasible to use camera traps for mammal surveys. Since the 1990s the method has been used extensively in tropical forests of Asia as well as in Africa (Griffiths and Van Schaik 1993; Karanth and Nichols 1998; Franklin et al. 1999). Lately, also field workers in the Neotropics have been exploring the method (Rittl 1998; Yabe and Higuchi 1998; Yabe et al. 1998; Trolle, unpublished data). Undoubtedly, camera trapping will become one of the main methods for monitoring mammals in the Neotropics (A. Langguth, personal communication; Louise Emmons, personal communication). The method has applicability in general species inventories, presence–absence studies, and population surveys of individually recognizable species such as spotted cats. In this section, I give recommendations based on a total of 7.5 months of camera trapping in the Pantanal and the Amazon of Brazil and Peru (Trolle, unpublished data). Where to place traps: For a general species inventory all major habitat types should be covered. Generally productive sites are trails (especially where two trails cross), natural corridors, along fences, along rivers, natural licks, mud-wallows, drinking sites and underneath fruiting trees. In the Pantanal the narrow grass beaches between forest and swamp usually give a high trapping success. Felids, canids, raccoons, and tapirs are generally trapped well along trails. Animal signs are useful: tracks, scratching trees of Jaguar and Giant Anteater, Puma scrapes, Ocelot faeces piles, rubbing trees of Collared Peccary, sleeping sites of ungulates, armadillo burrows in use, grounds with many armadillo feeding holes, subterranean termite brooding chambers frequented by Giant Armadillo and Giant Anteater, etc. Many species shelter in hollow trunks, and this may also be an interesting place for traps. In the Pantanal, gallery vegetation, with its high floristic influence from the Amazon, should be trapped intensively. Bait: Using bait can highly improve the camera trapping success. Sardines in oil, cod liver oil and fur trapper lure are recommended for carnivores; salt or fruit for herbivores. Setting up the trap: Set up the trap at a pole cut at the site. Place the trap about 3–4 m from where you expect the animals to pass (the distance the flash works). The traps can be mounted at a fixed distance from trails throughout the survey, and / or a stick of known length can be placed in the photo, which will facilitate size comparisons of animals photographed. Smaller species can be excluded by placing the sensor higher. Ant, termite and rodent protection: To protect against ants and termites wrap adhesive tape around the base of the pole with the sticky side facing outwards. Above this, the bark can be peeled upwards with a machete so that an ‘umbrella’ is constructed; this umbrella protects the tape against rain and probably works against rodents (like the system used on ropes on ships, to prevent rats from entering). To protect wires from rodents wrap them up in duct tape.

834 Flash mode: The ‘auto flash’ mode is recommended. The ‘fill in flash’ mode can cause an animal to be blurred if moving. Film: 400 ASA standard film for prints (e.g. Fujicolor) is recommended. In humid conditions such as the Amazon, 36exp film may swell and make rewinding of the film difficult. Cattle ranching and conservation in the Pantanal Most of the Pantanal consists of cattle ranches, and there seems to be a widespread opinion in the region that cattle ranching is an ecologically sustainable activity and can go hand in hand with conservation. However, intensive cattle ranching combined with annual fires to promote grass growth has a number of serious impacts on the environment. Here I will briefly offer some considerations on cattle ranching and conservation in the Pantanal and recommendations for future conservation biology studies and conservation measures in the region. A preliminary study of acurı´ forest on land adjacent to the UNIDERP reserve, with a high degree of cattle ranching, showed that the dominant woody components were all species typical of disturbed areas, species that can survive burning (e.g. the Acurı´ palm), and / or species that are not eaten by the cattle (Pott and Pott 1994). The forest, dominated by Acurı´ palms and often with hardly any undergrowth (typical of the Pantanal), was relatively species poor when it came to the woody plant species. The cattle browse inside the acurı´ forest, and scorching marks on the trees show that ´ fire also enters the forest. It is likely that the species poor, Acurı-dominated forest type is a direct result of the disturbances related to the intensive cattle ranching. The natural climax forest in the Pantanal is likely to be structurally different and much more diverse. In order to plan an adequate conservation strategy for the Pantanal, it is essential to answer three basic questions related to the cattle ranching activities: (1) what impacts does the cattle ranching have on the biodiversity?; (2) what happens to the habitats and the fauna when cattle ranching ceases in a conservation area?; and (3) would it be desirable to leave a certain number of cattle in the conservation area? Conservation biology studies comparing sites with and without cattle ranching, and studies monitoring the succession of areas where cattle have been taken out, should have the highest priority in the Pantanal. In the Pantanal the general opinion is that when cattle are taken out of an area the area will become ‘dirty’ and there will be fewer animals. This argument, although not based on scientific studies and highly biased, since it is easier to spot wildlife in open areas, is of great importance in the Pantanal because of ecotourism. Ecotourism has become a major industry in the region, and is increasing rapidly. In 2001 more than one million tourists were expected (Villela and Fuentes 2001). Ecotourism ought to promote conservation, but in the Pantanal this is not necessarily so. The tour operators are interested in showing the visitors as much wildlife as possible, and may decide to ‘clean’ their land for this reason. When an area is set aside for conservation and cattle ranching ceases or is diminished significantly, the natural succession will begin. The immediate result

835 might rightly be that the area becomes ‘dirty’. However, inside the forest the succession will continue and the pioneer species will eventually be replaced, as the forest approaches the climax state. This type of forest will most likely be much richer than the typical species poor acurı´ forest, and in the long run, leaving the forest undisturbed would undoubtedly pay off both when it comes to biodiversity and ecotourism.

Acknowledgements This project was a collaboration between the Zoological Museum University of ˜ do Copenhagen, Universidade para o Desenvolvimento do Estado e da Regiao Pantanal – UNIDERP, and Museu Nacional Universidade Federal do Rio de Janeiro. The project was made economically possible by the support of, among others, WWF-Denmark / Novo Nordisk, Zoological Museum of Copenhagen, University of Copenhagen, Copenhagen Zoo, NetTravel, Duracell, Photographica, and Fuji. Many thanks are due to UNIDERP for allowing the study to take place at Fazenda Sta ´ and for important logistical support. For essential help in preparing the Emılia ˜ Alves de Oliveira project and logistical support in Brazil I am grateful to Dr Joao ˜ Dr Hans and Neiva Maria Robaldo Guedes. Thanks are also due to Erika Guimaraes, ˚ and Mogens Andersen. J. Baagøe, Prof Jon Fjeldsa,

References Alho C.J.R. and Lacher T.E. Jr 1991. Mammalian conservation in the Pantanal of Brazil. In: Mares M.A. and Schmidly D.J. (eds), Latin American Mammalogy – History, Biodiversity, and Conservation. University of Oklahoma Press, Norman, Oklahoma, pp. 280–294. ´ Alho C.J.R., Lacher T.E. Jr, Campos Z.M.S. and Gonc¸alvez H.C. 1988a. Mamıferos da Fazenda ˜ de Nhecolandia, ˆ Nhumirim, sub-regiao Pantanal do Mato Grosso do Sul: levantamento preliminar de ´ especies. Revista Brasileira de Biologia 48: 213–225. Alho C.J.R., Lacher T.E. and Gonc¸alvez H.C. 1988b. Environmental degradation in the Pantanal ecosystem. BioScience 38: 164–171. Allen J.A. 1916. Mammals collected on the Roosevelt Brazilian Expedition, with field notes by Leo E. Miller. Bulletin of the American Museum of Natural History 35: 559–610. Anderson S. 1997. Mammals of Bolivia, taxonomy and distribution. Bulletin of the American Museum of Natural History 231: 1–652. ´ Becker M. and Dalponte J.C. 1991. Rastros de Mamıferos Silvestres Brasileiros. Editora Universidade de ´ ´ Brasılia, Brasılia, Brazil. Eisenberg J.F. and Redford K.H. 1999. Mammals of the Neotropics Vol. 3: The Central Neotropics: Ecuador, Peru, Bolivia, Brazil. University of Chicago Press, Chicago, Illinois. Emmons L.H. and Feer F. 1997. Neotropical Rainforest Mammals: A Field Guide. 2nd edn. University of Chicago Press, Chicago, Illinois. Franklin N., Bastoni, Sriyanto, Siswomartono D., Manansang J. and Tilson R. 1999. Last of the Indonesian Tigers: a cause for optimism. In: Seidensticker J., Christie S. and Jackson P. (eds), Riding the Tiger: Tiger Conservation in Human-Dominated Landscapes. Cambridge University Press, Cambridge, UK, pp. 130–147. Griffiths M. and Van Schaik P. 1993. Camera trapping: a new tool for the study of elusive rain forest animals. Tropical Biodiversity 1: 131–135.

836 Karanth K.U. and Nichols J.D. 1998. Estimation of tiger densities in India using photographic captures and recaptures. Ecology 79: 2852–2862. ´ ´ Lorenzi H. 2000. Plantas Daninhas do Brasil: Terrestres, Aquaticas, Parasitas e Toxicas. 3rd edn. ˜ Paulo, Brazil. Instituto Plantarum de Estudos da Flora Ltda, Sao ˆ ´ MS, Brasil), Masters Thesis, Universidade Lourival R.F.F 1993. A cac¸a da Nhecolandia (Corumba, Federal de Minas Gerais, Brazil. ˆ ´ Mittermeier R.A., Camara I.G., Padua M.T.J. and Blanck J. 1990. Conservation in the Pantanal of Brazil. Oryx 24: 101–112. ´ ´ ´ ´ ´ PCBAP 1997. Mamıferos. In: Diagnostico dos Meios Fısico e Biotico: Meio Biotico, Plano de ˜ da Bacia do Alto Paraguai Vol. 2(3). Ministerio ´ do Meio Ambienteos Recursos Hıdricos ´ Conservac¸ao ˆ e da Amazonia Legal, Brazil. ´ Brazil. Pott A. and Pott V.J. 1994. Plantas do Pantanal. EMBRAPA, Corumba, Prance G.T. and Schaller G.B. 1982. Preliminary study of some vegetation types in the Pantanal, Mato Grosso, Brazil. Brittonia 34: 228–251. Ratter J.A., Pott A., Pott V.J., da Cunha C.N. and Haridasan M. 1988. Observations on woody vegetation ´ Brazil. Notes from the Royal Botanic Garden of Edinburgh 45: types in the Pantanal and Corumba, 503–525. ˜ seletiva da madeira sobre a communidade de pequenos mamıferos ´ Rittl C.E. 1998. Efeitos da extrac¸ao de ˆ uma floresta de terra firme na Amazonia central, Masters Thesis, INPA / Universidade de Amazonas, Manaus, Brazil. Schaller G.B. 1983. Mammals and their biomass on a Brazilian ranch. Arquivos de Zoologia 36: 1–36. Villela R. and Fuentes L. 2001. A vez do Pantanal e´ agora. Veja, 31 / 1 / 2001, Brazil. Voss R.S., Lunde D.P. and Simmons N.B. 2001. The mammals of Paracou, French Guiana: a Neotropical lowland rainforest fauna. Part 2. Nonvolant species. Bulletin of the American Museum of Natural History 263: 1–236. Willink P.W. (ed.) (2000) A biological assessment of the aquatic ecosystems of the Pantanal, Mato Grosso do Sul, Brazil. In: RAP Bulletin of Biological Assessment. Conservation International, Washington, DC. ˜ e consumo de frutos caıdos ´ Yabe T. and Higuchi N. 1998. Dispersao da uxi-amarelo (Endopleura uchi) ˜ ZF-2 do INPA, Amazonia ˆ ´ por mamıferos da floresta natural da Estac¸ao Central. In: Higuchi N., ˜ da Campos M.A.A., Sampaio P.T.B. and dos Santos J. (eds), Pesquisas Florestais para a Conservac¸ao ˜ de Areas Degradadas da Amazonia. ˆ Floresta e Reabilitac¸ao MCT / INPA, Manaus, Brazil, pp. 84–91. ˆ ´ ´ ´ Yabe T., Rittl C.E. and Higuchi N. 1998. Especies de mamıferos registradas por cameras fotograficas ˜ Experimental de Silvicultura Tropical do Instituto Nacional de Pesquisas da ´ automaticas na Estac¸ao ˆ ˆ Amazonia – EEST-INPA, Amazonia Central. In: Higuchi N, Campos M.A.A., Sampaio P.T.B. and ˜ da Floresta e Reabilitac¸ao ˜ de Areas dos Santos J. (eds), Pesquisas Florestais para a Conservac¸ao ˆ Degradadas da Amazonia. MCT / INPA, Manaus, Brazil, pp. 94–107.