O

Tm

NGa

Bb

0c

I

a

ARAHA

KAHRFS

I

csDshi

h1

Mammalian Biology 80 (2015) 431–439

Contents lists available at ScienceDirect

Mammalian Biology

jou rn al hom epage: www.elsev ier .com/ locate /mambio

riginal Investigation

he home range and multi-scale habitat selection of the threatenedaned three-toed sloth (Bradypus torquatus)

ereyda Falconia, Emerson M. Vieirab,∗, Julio Baumgartenc, Deborah Fariac,astón Andrés Fernandez Ginéc

Universidade Estadual de Santa Cruz (UESC), Departamento de Ciências Biológicas, Programa de Pós-graduac ão em Ecologia e Conservac ão daiodiversidade, Rodovia Ilhéus Itabuna, Km 16, CEP 45650-000, Ilhéus, Bahia, BrazilLaboratório de Ecologia de Vertebrados, Departamento de Ecologia, Instituto de Ciências Biológicas, Universidade de Brasília, Campus Darcy Ribeiro, C.P.4457, Brasília, DF 70910-900, BrazilUniversidade Estadual de Santa Cruz (UESC), Departamento de Ciências Biológicas, Laboratório de Ecologia Aplicada à Conservac ão, Rodovia Ilhéus

tabuna, Km 16, CEP 45650-000, Ilhéus, Bahia, Brazil

r t i c l e i n f o

rticle history:eceived 30 July 2014ccepted 30 January 2015andled by Francesco Ferrettivailable online 7 February 2015

eywords:tlantic forestabitat useadiotelemetryorest structurehaded cocoa plantation

a b s t r a c t

Habitat selection is a scale-dependent process of paramount importance to the understanding of howspecies deal with environmental variation. This process has practical implications for wildlife conserva-tion, aiding in the identification of key resources for animals and in the definition of scales relevant tothe proposal of practical conservation actions. In this study, we investigated in different spatial scales thehabitat selection of the maned three-toed sloth (Bradypus torquatus), a threatened and endemic arborealfolivore of the Atlantic rainforest (vulnerable, sensu IUCN). We radio-tracked and monitored seven slothsfor 18 months in landscapes of Southern Bahia, Brazil, the current core region of the species’ distribution.The average values of the home-range estimates were low, but varied considerably among individualsregardless of the estimator (from 0.95 to 27.8 ha, MCP method, and from 0.39 to 21.52 ha, fixed kernelmethod). At the landscape scale, the maned sloths preferred early secondary forest and shade cacaoplantations, avoided open areas, and occupied late secondary forest as expected compared to its avail-ability. At the home range scale, however, the sloths did not show preference for any forest category,though, again, avoided open areas. At smaller spatial scales, the sloths were highly selective towardsforest patches characterized by complex vegetation structures (i.e., areas with a high density of trees,closed and dense canopies), and selected large trees with lianas and bromeliads and also with connected

crowns. The high selectivity observed at finer scales appeared to be the result of limited spatial percep-tion and experience due to the species’ characteristic slow mobility. Our results support the notion thatmaned sloths can effectively occupy (and even select for) disturbed forested habitats. However, we donot know whether sloth populations are viable in landscapes containing only disturbed habitats or lowproportions of undisturbed habitats.

© 2015 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier GmbH. All rights reserved.

ntroduction

Habitat loss, disturbance, and fragmentation are synergistic pro-esses that account for the current and unprecedented rates ofpecies loss on a global scale (Fischer and Lindenmayer, 2007).ue to the fact that extinction proneness is closely linked with a

pecies’ capacity to utilize and thrive in disturbed, anthropogenicabitats, species-specific information on habitat use and selection

s of paramount importance for conservation purposes. Patterns

∗ Corresponding author. Tel.: +55 61 31072998; fax: +55 61 31072904.E-mail address: emvieira@unb.br (E.M. Vieira).

ttp://dx.doi.org/10.1016/j.mambio.2015.01.009616-5047/© 2015 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier Gmb

of habitat selection affect species distribution and abundance(Saavedra-Rodríguez et al., 2012) and are often scale dependent(Kelt et al., 1999; Morin et al., 2005; Finlayson et al., 2007). Dueto the fact that ecosystem processes encompass different spatialscales, the understanding of the effects of scale-dependent vari-ation in habitat and resource quality is highly relevant for theeffective conservation of any species (Milne et al., 1989).

The evaluation of habitat-use patterns is particularly needed forthe establishment of adequate strategies for the conservation of

rare, threatened, and highly-specialized species (Croak et al., 2013),such as the maned three-toed sloth (or simply maned sloth) Brady-pus torquatus (Mammalia, Bradypodidae). It is an arboreal folivoreendemic to the Brazilian Atlantic forest whose populations are

H. All rights reserved.

4 lian Bi

c(ae(tsas

sveSiastt

sgsfs

twswSlhc

pii(m1nctm

M

S

iRi(ih(wf

Ftiw

32 N. Falconi et al. / Mamma

urrently listed as vulnerable due to habitat loss and fragmentationMachado et al., 2008; IUCN, 2014). This species inhabits isolatedreas of central-northeastern to central portion of the Atlantic for-st, from Rio de Janeiro to Southern Sergipe, along the coast of BrazilHirsch and Chiarello, 2011). The region of Southern Bahia main-ains the largest and most genetically diverse population of thispecies (Lara-Ruiz and Chiarello, 2005; Hirsch and Chiarello, 2011),nd thus represents a key location where conservation actionshould be focused.

Our current limited knowledge on the space use of the manedloths shows wide variation in individual home range estimates,arying from 0.44 to 29.33 ha in the Southern Bahia state in North-astern Brazil (Cassano et al., 2011), 0.8 to 10.8 ha in the Espíritoanto state (Chiarello, 1998; Chiarello et al., 2004), and 4.7 to 16.2 han the Rio de Janeiro state (Pinder, 1997). Such variation may bettributed not only to differences in the methodologies of previoustudies, but also to distinct individual needs and to specific featureshat characterize distinct forest types (Chiarello, 2008), as well ashe landscape contexts of each study area.

The available information on the habitat selection of manedloths comes from one adult sloth and two juveniles from a sin-le plantation in Southern Bahia (Cassano et al., 2011). This studyuggested that sloths use mature forests and that they may pre-er shaded cacao plantations and avoids pastures, swamps, or earlyecondary forests (Cassano et al., 2011).

Given the scarcity of data on the home range and habitat selec-ion of maned sloths and their necessity for conservation purposes,e evaluated patterns of maned sloths’ habitat use at different

cales within the landscape of Southern Bahia. More specifically,e aimed to (i) estimate the home range size of maned sloths in

outhern Bahia, (ii) investigate the species’ habitat selection and itsevel of congruence across four different spatial scales (landscape,ome range, forest patch, and tree levels), and (iii) identify featuresonnected to habitat preference or avoidance in each spatial scale.

Sloths are slow-moving and low-metabolism animals whosehysiological and locomotion constraints do not allow for leap-

ng between trees to cross canopy discontinuities, which resultsn their weak dispersal ability compared to other mammal speciesEmmons, 1995; Peery and Pauli, 2014). In addition, due to slow

obility, sloths depend on crypticity to avoid predators (Eisenberg,978). Thus, considering the species’ locomotory limitations andeed to maintain cryptic habits, we anticipated that the structuralharacteristics of vegetation in their environment would be impor-ant factors in the habitat selection of the species across the study’s

ultiple scales.

aterial and methods

tudy area

Our study was conducted in landscapes of Southern Bahia,ncluding areas located within the limits of the Una Biologicaleserve (Rebio-Una) and neighboring farms located in the munic-

pality of Una, the Southern Bahia state, and Northeastern Brazil39◦18′–39◦00′W, 15◦23′–15◦03′S) (Fig. 1). The forest of this regions classified as tropical lowland rainforest characterized by a canopyeight of 25 to 30 meters and abundant epiphytes and woody lianasAmorim et al., 2008). The average rainfall is around 2000 mm/year,ith no regular dry season, and an annual temperature that ranges

rom 24 to 25 ◦C (Mori et al., 1983).The region of Una contains the largest remains of the Atlantic

orest in the Northeast of Brazil, a biome currently representing lesshan 16% of the original area (Ribeiro et al., 2009). Its landscapes composed primarily of forests in different successional stages,

ith nearly 50% of the land cover comprised of late secondary

ology 80 (2015) 431–439

forest fragments, and 16% early secondary forests (Faria et al., 2007).Shaded cacao plantations cover 6% of the landscape, the majority ofwhich are quite small (median patch size: 2.73 ha) and immersed ina forest-dominated landscape (Faria et al., 2007). Open areas makeup the remaining ∼27% of the landscape. Therefore, forested areas,including shaded cacao plantations, make up the bulk of the avail-able habitat in the region. These plantations are the result of specificmanagement interventions that vary according to region. Previousstudies that compared native forests with shaded plantations inthe Una region of Southern Bahia indicated that these latter forma distinct forest category, characterized by decreasing densities ofwoody lianas and foliage within the superior forest strata (Faria,2002). In particular, the shaded cacao plantations of Rebio-Una andits neighboring farms commonly show higher structural complex-ity and greater density of native trees compared to the shaded cacaoplantations of other regions of Southern Bahia, this mainly due tothe Rebio-Una’s laws and conservation goals.

Sloth capture and radiotracking

We captured seven maned sloths between May 2011 and March2012 (Table 1). The sloths were physically immobilized followinga technique used in previous studies (Chiarello, 1998; Chiarelloet al., 2004; Cassano et al., 2011), and each sloth was fitted with aball-chain collar attached to a VHF radiotransmitter (Biotrack Ltd.;Model TW3SM) weighing 30 g (<1% of the weight of the animals).Four maned sloths were captured inside the forest and releasedback into the same tree where they had been captured. The otherthree animals were captured after being sighted by local people orpark rangers on the roads of the Rebio-Una near high hunting pres-sure areas. We translocated these sloths into more protected areaswithin the Rebio-Una. Except for one individual whose radio signalwas lost at the end of the study, we recaptured the remaining sixsloths and removed the radio collars without injury. We followedthe recommended procedures of the American Society of Mam-malogists (Sikes and Gannon, 2011) in the capture, handling, andtagging of the species.

Each sloth was radio-tracked with the aid of a receptor (TelonicsInc.; Model TR-4) coupled to a Yagi three-element antenna(Telonics Inc.). We began monitoring each sloth seven days afterits release to ensure a period of habituation. For each monitoredsloth, we collected one to three locations per week during diurnalperiod (0600 to 1700 hrs) from 8 to 17 months, totaling 23 to 50localizations per animal (Table 1). We sampled during all hours ofdiurnal periods homogeneously across individuals in order to avoidtemporal bias.

Habitat evaluation

For the coarse scale, we mapped and categorized the vegetationtypes present in the surrounding landscape of each studied slothaccording to their structural characteristics and definitions foundin Brazilian legislation (CONAMA, 1994). These categories includedlate secondary forests (LSF, a forest with canopy height greaterthan 12 m and an average diameter at breast height (DBH) greaterthan 18 cm), early secondary forests (ESF, a forest with a canopyheight lower than 12 m and an average DBH smaller than 18 cm),shaded cacao plantations (SCP, including shaded cacao plantationsand forested crops with similar structures), and open areas (OA,including pastures and monocultures). Mapping was done to definethe limits of each type of vegetation using aerial photographs (scaleof 1:5000; taken in 2001–2002; IESB, unpublished data) and Google

Earth® imagery (Spot image, captured May 26, 2010) and also con-firming such limits in the field with a GPS (GPSMAP 60CSX, GarminInternational, Inc., Olathe, KA, USA). We then calculated the area ofeach vegetation type within each individual home range and the

N. Falconi et al. / Mammalian Biology 80 (2015) 431–439 433

F highlia radio-

lahoWtt

oratt1eaTawui

THa

ig. 1. Map of the Una Biological Reserve in the Bahia state of Northeastern Brazil,

nd the geographic position of the center of the home range of seven maned sloths

andscape. Landscape limits were defined by a circular area with 1000-m radius (area = 314 ha) around the center of each sloth’some range. This radius was based on the distance from the centerf a sloth’s home range to the furthest location where it was found.e assumed that such a defined area would allow us to detect at

he landscape level the potential importance of habitat mosaics inhe sloths’ selection processes for a home range.

At the finer scale, we measured all trees used by sloths in termsf diameter at breast height (DBH), total tree height, and the occur-ence of lianas, bromeliads, and connected crowns. For each forestrea used, we established a 100 m2 plot around each used tree andook the same measurements for all trees with a DBH > 3.2 cm insidehe plot, as well as measured tree density (number of trees per00 m2), canopy openness, and the vertical profile of the foliage. Westimated canopy openness using hemispheric photographs takent 1-m height from three points located 1 m from the used tree.hese points were on the vertices of an imaginary equilateral tri-

ngle located 1 m from the center (used tree), one of these verticesas located to the north of it. Posterior estimates were obtainedsing the Gap Light Analyzer (GLA) 2.0 (Frazer et al., 1999). Modify-

ng the method described in Hubbell and Foster (1986), the vertical

able 1ome ranges (in ha) of seven maned sloths radio-tracked in Southern Bahia, Brazil. Estimnd 50%) methods. Each animal was identified by a code that included an acronym for Br

Code Sexa/age Period

BT171b F/adult May 2011–Dec 2011

BT163 M/sub-adult Jun 2011–Jan 2012

BT323b F/sub-adult Jul 2011–Aug 2012

BT393 F/adult Dec 2011–Oct 2012

BT182 F/adult Jan 2012–Oct 2012

BT033b F/sub-adult Mar 2012–Oct 2012

BT122 F/adult Apr 2012–Oct 2012

a Sex determination based on Lara-Ruiz (2004).b Translocated animals.

ghting the Una region (black rectangle). The delimitation of Una Biological Reservetracked are shown on the right.

profile of the foliage was evaluated at the same three points using a3 m pole to allow for vertical sighting and estimation of the foliagedensity of the superior (height >10 m) and inferior (height <10 m)stratum as percentages of the foliage in these strata. We used themean of the three measurements of canopy openness and foliagedensity obtained in the “used” plot for the analysis of habitat selec-tion. Lastly, for each animal location (Table 1), we randomly choseone tree within each home range, established a 100 m2 plot aroundit, and then took the same measurements as described previously.

Home range estimates

In the present study, there was a high frequency of no or smalldisplacement between successive fixes obtained for the sloths.Thus, we estimated home ranges using only one fix per weeklyinterval in an effort to increase independence between successivelocations (Cassano et al., 2011). Home ranges were estimated using

the 100% minimum convex polygon method (MCP100%; Mohr,1947) and the 95% fixed kernel estimator (FK95%; Worton, 1989)with bandwidth (h) estimated using the least square cross vali-dation process (hLSCV). In addition, we estimated core area sizes

ates were made using the minimum convex polygon (MCP) and fixed kernel (95%adypus torquatus (BT) followed by a specific number.

Fixes (weekly) MCP Fixed kernel

95% 50%

23 (23) 27.8 21.52 4.6124 (23) 10.64 6.89 1.5450 (34) 13.71 6.59 1.3536 (24) 2.92 7.14 1.5343 (21) 11.5 2.35 0.4341 (21) 1.09 0.39 0.8634 (18) 0.95 3.32 0.78

434 N. Falconi et al. / Mammalian Biology 80 (2015) 431–439

Fig. 2. Habitat selection by B. torquatus at the landscape (second spatial-scale order; a) and home-range (third spatial-scale order; b) scales in the Atlantic Forest of SouthernB interva availac

utR9brs((

daaC

H

Saadasiheb

totihto(fmssow

ahia, Brazil. Dark dots indicate mean proportional utilization using the confidence

bility. Selection occurs when utilization falls above (positive) or below (negative)

acao plantations; and OA = open areas.

sing a fixed kernel method based on 50% of the utilization dis-ribution (FK50%). We performed these analyses using the Homeange Tools (HRT) extension (Rodgers et al., 2007) to the ArcGIS®

.3 software (Environmental Systems Research Institute, 2008). Weuilt a cumulative curve in order to check the stability of the homeanges in terms of home range size (MCP) in relation to the sampleize of each animal using the Animal Movement Extension, v.2.04Hooge and Eichenlaub, 1997) of the ArcView GIS 3.2 programEnvironmental Systems Research, 1999).

We used the non-parametric Mann–Whitney test to evaluateifferences in the home range sizes of adult and sub-adult sloths,s well as between translocated and non-translocated sloths. Thesenalyses were performed using version 3.0.2 of the R software (Rore Team, 2013).

abitat selection analysis

We performed the habitat selection analysis using Resourceelection for Windows version 1.0 (Leban, 1999) with each animals a sampling unit. We employed a chi-square goodness-of-fit testnd Bonferroni confidence intervals a posteriori (Neu et al., 1974) toetect any resource used disproportionately compared to its avail-bility. When the observed use of a habitat category did not differignificantly from the expected one compared to its availability,t was considered as not selected (a random use). Otherwise, theabitat was considered to be preferred (more frequently used thanxpected by chance) or avoided (less frequently used than expectedy chance) (Byers et al., 1984).

According to Johnson (1980), we evaluated habitat selection inerms of four spatial scales. At the landscape scale (Johnson’s sec-nd order), the proportion of the area occupied by each vegetationype within each sloth’s home range (FK95%) was compared withts availability in the landscape (1000 m-radius circular area). At theome range scale (Johnson’s third order), the proportion of localiza-ions for each vegetation type was compared with the proportionf area occupied by each vegetation type within the home rangesFK95%). At the patch scale (Johnson’s fourth order), the relativerequency of the measured structural variables reported in the 100-

2 plots was compared with randomly selected plots of the same

ize. Finally, we considered an additional “fifth order” at the treecale in terms of the relative frequency of the measured attributesf the used trees compared to those measured for non-used treesithin the same 100-m2 plot. This order was used to calculate sloth

als of Bonferroni (Neu et al., 1974). Dark triangles indicate mean proportional avail-bility. Codes: LSF = late secondary forest; ESF = early secondary forest; SCP = shaded

preferences in terms of tree scale. Spearman correlations wereperformed prior to the habitat selection analysis between all treeattribute pairs, and due to the high correlation (r > 0.5) betweenDBH and tree height, the latter variable was removed from the anal-ysis. The other paired comparisons showed low correlation values(r < 0.5).

Results

Home range

The estimated mean [SD] home range sizes were 9.80 [9.53] ha(MCP100%) and 6.89 [6.95] ha (FK95%), with mean core areas(FK50%) of 1.59 [1.40] ha (Table 1). The estimated median val-ues of home ranges were 10.64 ha (MCP100%), 6.88 ha (FK95%),and 1.35 ha (FK50%). Home range size was not significantlydifferent between translocated and non-translocated animalswhen estimated using the MCP100% method (translocated:median = 13.71 ha, n = 3; non-translocated: median = 6.78 ha, n = 4;Mann–Whitney U = 9, P = 0.29) or the FK95% method (translocated:median = 6.59; non-translocated: median = 5.10; Mann–WhitneyU = 6, P = 1.0). Furthermore, we did not detect any significant dif-ferences between the adult and subadult sloth home ranges usingthe MCP100% method (adult: median = 7.21 ha, n = 4; subadult:median = 10.64 ha, n = 3; Mann–Whitney U = 6, P = 1.0) or the FK95%method (adult: median = 5.23 ha; subadult: median = 6.59 ha, n = 3;Mann–Whitney U = 4, P = 0.48). The home range size of the onlymale sloth monitored (BT163) held an intermediary positionbetween those of females, being similar or very close to the medianvalues considering the home ranges of all sloths. Home range sizestabilized around seven months for most sloths, except for BT171(a translocated individual) and BT182 (non-translocated), whosecumulative curves suggested that their home ranges could increase(Supplemental Material, Fig. A).

Habitat use and selectionHabitats such as LSF (57.4%) and ESF (30.7%) represented the

major part of sloths’ home ranges (FK95%), followed by OA (6.4%)and SCP (5.6%). Within their home ranges, the tagged sloths were

located predominantly in LSF (56.4%) and ESF (34.9%), and less fre-quently in the SCP (6.6%) and OA (2.1%). At both the landscapeand the home range scales, sloths did not use habitats ran-domly (�2 = 306.2, df = 3, P < 0.0001; and �2 = 25.8, df = 3, P < 0.0001,

N. Falconi et al. / Mammalian Biology 80 (2015) 431–439 435

F e ordec ropor(

rart

Ttacips(s1dotla

tdPnp

ig. 3. Habitat selection by B. torquatus at the patch scale, the fourth spatial-scalonfidence intervals of Bonferroni (Neu et al., 1974). Dark triangles indicate mean pnegative) availability.

espectively). At the landscape scale, the sloths preferred both ESFnd SCP, avoided OA, and did not select LSF (Fig. 2). At the homeange scale, the sloths used all forested vegetation types accordingo their availability, though avoided OAs (Fig. 2).

The maned sloths were also selective at the patch scale (Fig. 3).he observed use of forest patches differed significantly fromhe expected values considering the availability of habitat vari-bles in terms of the density of trees (�2 = 449.3, df = 2, P < 0.0001),anopy opening (�2 = 90.3, df = 2, P < 0.0001), and foliage densityn the stratum above 10 m (�2 = 85.2, df = 2, P < 0.0001). The slothsreferred structurally complex patches characterized by high den-ities of trees (>20 trees per 100 m2), intermittently open canopies10–20% of openness), and moderate densities of foliage in theuperior stratum (20–40% of foliage density at the stratum above0-m high). In contrast, the sloths avoided forest patches with lowensities of trees (<20 individuals per 100 m2) or in extreme casesf canopy openness (<10% or >20%). They were also indifferento variation in foliage densities below 10 m, using them in simi-ar proportion to what was expected to be used compared to theirvailability (�2 = 2.6, df = 2, P = 0.26).

At the finer spatial scale (trees scale), the use of trees byhe tagged sloths was not random in terms of DBH (�2 = 3677.7,

f = 2, P < 0.0001) and the occurrence of lianas (�2 = 1254.2, df = 1,

< 0.0001), bromeliads (�2 = 1293.7, df = 1, P < 0.0001), and con-ected crowns (�2 = 188.5, df = 1, P < 0.0001). The maned slothsreferred thick (>20 cm of DBH) trees, as well as trees with sections

r of habitat selection. Dark dots indicate mean proportional utilization using thetional availability. Selection occurs when utilization falls above (positive) or below

of lianas, bromeliads, and connected crowns (Fig. 4). Conversely,the trees that did not have such structures were avoided bysloths.

Discussion

To our knowledge, our study is the most complete publishedresearch addressing the use of space and habitat selection of thevulnerable maned sloth. Our results indicated that the home-rangesizes of these sloths vary markedly in Southern Bahia, and that thisspecies is selective in its habitat requirements across different spa-tial scales. The sloths selected early secondary forests and shadedcacao plantations at the landscape level, while avoiding open areasat several other scales (Fig. 5). At the lower spatial scales, manedsloths preferred forest patches with a high tree density, interme-diate canopy openness, and a well-developed canopy stratum, aswell as larger trees with lianas, bromeliads, and connected crowns.

Home range

The wide variation in terms of home range size reported in thisstudy (0.95 to 27.8; minimum convex polygon 100% method) has

also been observed in previous studies (0.56 to 29.33 ha; Pinder,1997; Chiarello, 1998; Chiarello et al., 2004; Cassano et al., 2011).This variation may reflect some aspects, such as environmental con-ditions, individual sloth histories (translocated or not translocated),

436 N. Falconi et al. / Mammalian Biology 80 (2015) 431–439

F of habi l availa

od2uh

ol(esatwi1(

H

mrvd

ig. 4. Habitat selection by B. torquatus at the tree scale, the fifth spatial-scale order

ntervals of Bonferroni (Neu et al., 1974). Dark triangles indicate mean proportionavailability.

r else distinct individual needs (e.g., related to sex, age or repro-uctive condition; Pinder, 1997; Chiarello, 1998; Chiarello et al.,004; Chiarello, 2008; Cassano et al., 2011). Unfortunately, we werenable to identify which factors cause significant variation in theome range size of our study animals.

Except for the sloth that seemed to be dispersing (BT171), thebserved home range values (range: 0.95 to 13.71 ha) were simi-ar to those of the non-dispersing maned sloths previously studied0.56 to 10.8 ha; Chiarello et al., 2004; Chiarello, 2008; Cassanot al., 2011), as well as similar to the values observed for otherloth species, such as B. variegatus (0.5 to 3.7 ha; Montgomerynd Sunquist, 1975; Queiroz, 1995) and Choloepus hoffmanni (0.4o 3.9 ha; Montgomery and Sunquist, 1975). Moreover, the valuesere lower than those estimated for herbivorous mammals of sim-

lar size to the maned sloth (based on data presented in Damuth,981), as expected for a low-metabolism arboreal folivore mammalEisenberg, 1978).

abitat selection

At the landscape scale, the studied sloths predominantly used

ore than one habitat type. Similarly, Vaughan et al. (2007)

eported that, in a diverse landscape within Central America, B.ariegatus and Choloepus hoffmanni commonly had two or threeistinct habitats, such as cacao agroforests, living-fence rows, open

itat selection. Dark dots indicate mean proportional utilization using the confidenceability. Selection occurs when utilization falls above (positive) or below (negative)

areas, and riparian forest strips. This suggests that sloths require amosaic of habitats to fulfill their basic living requirements.

Our results indicated that maned sloths preferred shaded cacaoplantations and early secondary forests among their available habi-tats to establish a home range. The floristic composition of theseareas likely makes their vegetation types attractive to sloths despitethe fact that these structurally simplified habitats may increaseexposure to potential predators. Shaded cacao plantations and earlysecondary forests harbor high abundances of pioneer tree species,such as Ficus, Tapirira, and Ocotea, which are commonly consumedby sloths (Mariano-Neto, 2004; Sambuichi, 2006; Cassano et al.,2011). The preference of the sloths for shaded cacao plantationsand early secondary forests was observed even for those that weretranslocated to the interior of late secondary forests. Thus, at thisscale (Johnson’s second order), the composition and abundanceof tree species seemed to be more important in determining amaned sloth’s home range than the high structural complexity ofthe forested habitat.

At the home range scale (Johnson’s third order), the varioustypes of forested habitats did not appear to influence the move-ment of the maned sloths. The sloths appeared to be capable ofadapting their diets according to differences in the composition and

abundance of consumed plant species available in a given habi-tat, feeding mostly on shade-tolerant species in forest remnantsand predominantly on shade-intolerant plants present at shadedcacao plantations (Cassano et al., 2011). Such adaptation indicates

N. Falconi et al. / Mammalian Bi

Fig. 5. Conceptual diagram of factors determining the habitat use of the maned slothwas

aqpa

awpsspawa

tslilsssbtm

td1sa(twTaa

ithin four spatial scales of the Atlantic Forest in Northeastern Brazil (modifiedfter Bunnell & Huggard, 1999). Habitat variables that were positively or negativelyelected in each scale are indicated by the correspondent signals (+) or (−).

certain flexibility of the sloth when engaging habitats of differinguality, a flexibility which has been previously reported (cf. marsu-ials, Finlayson et al., 2007; Smith et al., 2013; ungulates, Schaefernd Messier, 1995).

The observed habitat selection of the maned sloths was gener-lly congruent across spatial scales, although some inconsistenciesere observed. The main discrepancy between the sloths was theirreference for more complex components of habitat at the finercale (trees and patches), yet this was not the case at the coarsercale (vegetation types). However, sloths may prefer more com-lex forest patches even within more simplified habitat types, suchs shaded cacao plantations. They also selected large-sized treesithin areas where such trees are not the most abundant type, such

s in early secondary forests.In particular, there was a notable congruence in terms of habi-

at preferences between the finer and coarser scales. Part of thetructural differences between the shaded cacao plantations andate secondary forests from this region is that the former habitats characterized by lower densities of vegetation in the middle andower forest strata (Faria, 2002). Shaded cacao plantations and lateecondary forests had no influence on habitat selection within theloths’ home ranges and our fine-scale results indicated that theloths’ habitat selection was not influenced by vegetation densitieselow 10 m high. In particular, open areas were the only habi-ats avoided at both coarse scales, which is to be expected from

ammals that are strongly adapted to arboreal life.The studied sloths seemed to be highly selective in terms of habi-

at features at the lower scales. Sloths are known to stay for severalays in the same tree or forest patch (Montgomery and Sunquist,975; Queiroz, 1995; GAFG, pers. obs.). Thus, we anticipated highelectivity at these scales. As expected for a non-leaping and crypticrboreal mammal, structural features linked to canopy connectivityconnected crows and with lianas) and vegetation complexity (highree densities and well-developed foliage in the higher stratum)

ere selected by the sloths’ at both the patch and tree scales.

hese features, linked to tree connectivity and animal crypticity,re commonly preferred also by other sloth species (Montgomerynd Sunquist, 1978; Acevedo-Quintero et al., 2011).

ology 80 (2015) 431–439 437

The studied maned sloths preferred thick (diameter at breastheight >20 cm) trees, similarly to other sloth species (Choloepushoffmanni and Bradypus variegatus; Montgomery and Sunquist,1978; Queiroz, 1995) and arboreal folivorous mammals (Ganzhorn,1995; Milner and Harris, 1999; Cunningham et al., 2004; GAFG,unpublished data). Predation risk from terrestrial predators, suchas pumas and ocelots (still present in the study area), can be poten-tially reduced when an arboreal mammal locates itself in tall trees(Sweitzer and Berger, 1992). Considering that the species’ mainaerial predator (the harpy eagle Harpia harpyja) is locally very rareor absent in the study region, exposure in tall crowns is likely anon-risky action for these sloths.

In addition, thick trees tend to have wide crows, a factor thatcan potentially be related to high availability of foliar biomass(Ganzhorn, 1992). The maned sloth’s selection of large trees mayoptimize its rate of food intake, representing an energy-savingstrategy important for an arboreal folivore energetically con-strained by its diet (McNab, 1978). Indeed, the tagged maned slothsselected sites with a canopy openness of 10 to 20%, which is rela-tively high compared to the values reported for native forests andshaded cacao plantations in Southern Bahia (3.1 to 7.7% and 2.5 to6.1%, respectively; Pinheiro et al., 2013). These animals also pre-ferred (at the landscape scale) habitats with typically more opencanopies, such as in the early secondary forests and shaded cacaoplantations. The sloths’ preference for forest patches with rela-tively high canopy openness could be the result of their preferencefor forested sites with canopies of high sunlight exposition, foliarproduction, and dominated by pioneer tree species.

Conservation remarks

The shaded cacao plantations surrounding the Una BiologicalReserve included in both our and Cassano et al.’s (2011) studies onthe habitat selection of maned sloths are undermanaged and havegreater structural complexity than other shaded cacao plantationsof the Southern Bahia region (Cassano et al., 2011). We believe thatthe presence and abundance of specific tree species, coupled withthe maintenance of a relatively open, yet complex canopy, are themain factors that account for the use and selection of the shadedcacao plantations by maned sloths.

Among the local management practices currently recom-mended to increase farm productivity in the region is the use of lessshade, which often reduces tree density in these newly-managedplantations to only a third of what is found on traditional shadedcacao plantations (Schroth et al., 2014). This procedure markedlyreduces the quality of these highly managed plantations as habitatsfor maned sloths and other mammals (Rolim and Chiarello, 2004;Cassano et al., 2014) by decreasing the availability of trees whiletriggering structural simplification on a local scale, also throughthe consequent reduction of canopy connectivity and overall foliagedensity.

Although our results support the observations made by Cassanoet al. (2011) that the maned sloth can effectively use and evenselect shaded cacao plantations and other modified habitats, thispattern was observed within a landscape with a large proportionof mature forests. It has already been shown that, in the Bahiastate, the capacity of shaded cacao plantations to support diversemammal assemblages is strongly context-dependent and largelyinfluenced by the presence of large forest tracts in the landscape(Faria et al., 2007). It is not clear whether maned sloths can main-tain long-term viable populations in shaded cacao plantations inthe absence of undisturbed forests, or even simply with a low

proportion of such habitats in a landscape. For instance, althoughPeery and Pauli (2014) estimated the population growth rates of B.variegatus (brown-throated sloths) and Choloepus hoffmanni (Hoff-mann’s two-toed sloths) in a shade-grown cacao farm in Central

4 lian Bi

Aalf

iwsoawaI

saatrsitttcaiits

A

1afiC1pftvC(

A

t0

R

A

A

B

C

Paulo.

38 N. Falconi et al. / Mamma

merica, only the two-toed sloth populations were stable in thebsence of immigration, while the brown-throated sloth popu-ations could only be maintained via immigration from nearbyorests.

Therefore, it is possible that the population of the maned slothsn Southern Bahia can only be maintained in protected landscapes

ith a high proportion of forests, such as Rebio-Una. In fact, thispecies is rarely found in landscapes with a limited occurrencesf forests and high proportions of shaded cacao plantations, suchs those from the municipality of Linhares, Espirito Santo state,hich is located roughly 580 km south of our study region (Rolim

nd Chiarello, 2004), and those of the neighboring municipality oflhéus, 60 km to the north (GAFG, pers. obs.).

The selection of shaded cacao plantations as a habitat by manedloths can be influenced by local management of such plantationsnd, most likely, landscape context. Thus, we predict that the over-ll abundance of B. torquatus has already been greatly reduced inhe cocoa producing region of southern Bahia, mainly where tree-eduction management has been intensified and native forests arecarce. On the other hand, the small home range and high plastic-ty of this species in the disturbed and human-modified forests ofhe Atlantic Forest biome (Ribeiro et al., 2009) speak positively tohe possible conservation of this species. This plasticity suggestshat the species’ conservation situation is not as alarming as oneould suppose. However, further studies are required to addressnd understand the influence of environmental variables and vary-ng landscape contexts on the species. Given the currently availablenformation on B. torquatus habitat requirements, we suggest thathe conservation status of this species (vulnerable, sensu IUCN)hould be maintained.

cknowledgements

We thank Universidade Estadual de Santa Cruz (Project 00220.100.1134) CNPQ through Rede SISBIOTA (Process 563216/2010-7)nd CASADINHO/PROCAD (Process 552198/2011-0) and Idea Wildor their financial support. We are also in debt to CAPES for provid-ng fellowships for N. Falconi and G. Giné (Process 02566/09-5) andNPq for the research fellowships to D. Faria (Process 307221/2012-) and E.M. Vieira (Process 305981/2010-2). We thank severaleople who made this research possible especially Bruno Marchenaor his help throughout this entire study. We also acknowledgehe Instituto de Estudos Socioambientais do Sul da Bahia for pro-iding the aerial photographs and the Instituto Chico Mendes deonservac ão de Biodiversidade (ICMBio) for the research permitslicense number: 27021-1).

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at http://dx.doi.org/10.1016/j.mambio.2015.01.09

eferences

cevedo-Quintero, J.F., Granada, D.S., Plese, T., 2011. Abundancia y preferencia dehábitat de Bradypus variegatus y Choloepus hoffmanni durante la época seca endos fragmentos de bosque seco en Arboletes, Antioquia, Colombia. Edentata 12,36–44.

morim, A.M., Thomas, W.W., Carvalho, A.M., Jardim, J.G., 2008. Floristic of the UnaBiological Reserve, Bahia, Brazil. In: Thomas, W.W. (Ed.), The Atlantic CoastalForest of Northeastern Brazil. New York Botanical Garden Press, New York, pp.67–146.

yers, C.R., Steinhorst, R.K., Krausman, P.R., 1984. Clarification of a technique for

analysis of utilization-availability data. J. Wildlife Manage. 48, 1050–1053.

assano, R.C., Kierulff, M.C.M., Chiarello, A.G., 2011. The cacao agroforests of theBrazilian Atlantic forest as habitat for the endangered maned sloth Bradypustorquatus. Mamm. Biol. 73, 243–250.

ology 80 (2015) 431–439

Cassano, C.R., Barlow, J., Pardini, R., 2014. Forest loss or management intensification?Identifying causes of mammal decline in cacao agroforests. Biol. Conserv. 169,14–22.

Chiarello, A.G., 1998. Activity budgets and ranging patterns of the Atlantic forestmaned sloth Bradypus torquatus (Xenarthra: Bradypodidae). J. Zool. (Lond.) 246,1–10.

Chiarello, A.G., 2008. Sloth ecology: an overview of field studies. In: Vizcaíno, S.F.,Loughry, W.J. (Eds.), The Biology of the Xenarthra. University of Florida Press,Gainesville, pp. 135–152.

Chiarello, A.G., Chivers, D.J., Bassi, C., Maciel, M.A., Moreira, L.S., Bazzalo, M., 2004.A translocation experiment for the conservation of maned sloths. Biol. Conserv.118, 421–430.

CONAMA, 1994. Resoluc ão CONAMA no5 de 04 de maio de 1994. www.mma.gov.br/port/conama/legislacao/CONAMA RES CONS 1994 005.pdf (08.10.14).

Croak, B.M., Crowther, M.S., Webb, J.K., Shine, R., 2013. Movements and habitat useof an endangered snake, Hoplocephalus bungaroides (Elapidae): implications forconservation. PLoS ONE 8 (4), e61711, http://dx.doi.org/10.1371/journal.pone.0061711

Cunningham, R.B., Pope, M.L., Lindenmayer, D.B., 2004. Patch use by the greaterglider (Pteuroides volans) in a fragmented forest ecosystem. III. Night-time useof trees. Wildlife Res. 31, 579–585.

Damuth, J., 1981. Home ranges, home range overlap, and species energy use amongherbivorous mammals. Biol. J. Linn. Soc. 15, 185–193.

Eisenberg, J.F., 1978. The evolution of Arboreal Herbivores in the Class Mammalia.In: Montgomery, G.G. (Ed.), The Ecology of Arboreal Folivores. Smithsonian Insti-tution Press, Washington, pp. 135–152.

Emmons, L.H., 1995. Mammals of rain forest canopies. In: Lowman, M.D., Nadkarni,N.M. (Eds.), Forest Canopies. Academic Press, San Diego, pp. 199–223.

Environmental Systems Research Institute (ESRI), 1999. ArcView. Version 3.2. Envi-ronmental Systems Research Institute, Redlands, USA.

Environmental Systems Research Institute (ESRI), 2008. ArcGIS. Version 9.3. Envi-ronmental Systems Research Institute, Redlands, USA.

Faria, D., (Ph.D. Dissertation) 2002. Comunidade de morcegos em uma paisagemfragmentada da Mata Atlântica do sul da Bahia, Brasil. Universidade Estadual deCampinas.

Faria, D., Paciencia, M.L.B., Dixo, M., Laps, R.R., Baumgarten, J., 2007. Ferns, frogs,lizards, birds and bats in forest fragments and shade cacao plantations of twocontrasting landscapes in the Atlantic forest, Brazil. Biodivers. Conserv. 16,2335–2357.

Finlayson, G.R., Vieira, E.M., Priddel, D., Wheeler, R., Bentley, J., Dickman, C.R., 2007.Multi-scale patterns of habitat use by re-introduced mammals: a case studyusing medium-sized marsupials. Biol. Conserv. 141, 320–331.

Fischer, J., Lindenmayer, D.B., 2007. Landscape modification and habitat fragmenta-tion: a synthesis. Glob. Ecol. Biogeogr. 16, 265–280.

Frazer, G.W., Canham, C.D., Lertzman, K.P., 1999. Gap Light Analyzer (GLA), Version2.0. Imaging software to extract canopy structure and gap light transmissionindices from true-colour fisheye photographs. Users Manual and Program Docu-mentation. Simon Fraser University, Burnaby, British Columbia, and the Instituteof Ecosystem Studies, Millbrook, New York.

Ganzhorn, J.U., 1992. Leaf chemistry and the biomass of folivorous primates in trop-ical forest. Oecologia 91, 540–547.

Ganzhorn, J.U., 1995. Low-level forest disturbance effects on primary production,leaf chemistry, and lemur populations. Ecology 76, 2084–2096.

Hirsch, A., Chiarello, A.G., 2011. The endangered maned sloth Bradypus torquatus ofthe Brazilian Atlantic forest: a review and update of geographical distributionand habitat preferences. Mammal Rev. 42, 35–54.

Hooge, P.N., Eichenlaub, B., 1997. Animal Movement Extension to ArcView. Version1.1. Alaska Science Center – Biological Science Office, U.S. Geological Survey,Anchorage, USA.

Hubbell, S.P., Foster, R.B., 1986. Canopy gaps and the dynamics of a neotropical forest.In: Crawley, M.J. (Ed.), Plant Ecology. Blackwell, Oxford, pp. 77–95.

IUCN, 2014. IUCN Red list of threatened species. Version 2014.1. www.iucnredlist.org (11.07.14).

Johnson, D.H., 1980. The comparison of usage and availability measurements forevaluating resource preference. Ecology 61, 65–71.

Kelt, D.A., Meserve, P.L., Patterson, B.D., Lang, B.K., 1999. Scale dependence and scaleindependence in habitat associations of small mammals in southern temperaterainforest. Oikos 85, 320–334.

Lara-Ruiz, P., (Master’s Thesis) 2004. Tamanho corporal, dimorfismo sexual ediversidade genética da preguic a-de-coleira, Bradypus torquatus Illiger, 1811(Xenarthra: Bradypodidae). In: Pontifícia Universidade Católica de MinasGerais.

Lara-Ruiz, P., Chiarello, A.G., 2005. Life-history and sexual dimorphism of the Atlanticforest maned sloth Bradypus torquatus (Xenarthra: Bradypodidae). J. Zool. 267,63–73.

Leban, F., 1999. Resource Selection for Windows version 1.00. University of Idaho,Moscow.

Machado, A.M.B., Drummond, G.M., Paglia, A.P., 2008. Livro vermelho da faunabrasileira ameac ada de extinc ão. Fundac ão Biodiversitas, Belo Horizonte.

Mariano-Neto, E., (Ph.D. Dissertation) 2004. Efeitos da fragmentac ão sobre comu-nidades arbustivo arbóreas em Mata Atlântica, Una BA. Universidade de São

McNab, B.K., 1978. Energetics of arboreal folivores: physiological problems and eco-logical consequences of feeding on a ubiquitous food supply. In: Montgomery,G.G. (Ed.), The Ecology of Arboreal Folivores. Smithsonian Institution Press,Washington, D.C., pp. 153–162.

lian Bi

M

M

M

M

M

M

M

N

P

P

P

Q

R

N. Falconi et al. / Mamma

ilne, B.T., Johnston, K.M., Forman, R.T.T., 1989. Scale-dependent proximity ofwildlife habitat in a spatially-neutral Bayesian model. Landsc. Ecol. 2, 101–110.

ilner, J.M., Harris, S., 1999. Activity pattern and feeding behaviour of the tree hyrax,Dendrohyrax arboreus, in the Park National des Volcans, Rwanda. Afr. J. Ecol. 37,267–280.

ohr, C.O., 1947. Table of equivalent populations of North American small mam-mals. Am. Midl. Nat. 37, 223–249.

ontgomery, G.G., Sunquist, M.E., 1975. Impact of sloths on neotropical forestenergy flow and nutrient cycling. In: Golley, F.B., Medina, E. (Eds.), Tropical Ecol-ogy Systems: Trends in Terrestrial and Aquatic Research. Springer, New York,pp. 69–98.

ontgomery, G.G., Sunquist, M.E., 1978. Habitat selection and use by two-toed andthree-toed sloths. In: Montgomery, G.G. (Ed.), The Ecology of Arboreal Folivores.Smithsonian Institution Press, Washington, pp. 329–359.

ori, S.A., Boom, B.M., De Carvalho, A.M., Dos Santos, T.S., 1983. Southern Bahianmoist forest. Bot. Rev. 49, 155–232.

orin, P., Berteaux, D., Klvana, I., 2005. Hierarchical habitat selection by NorthAmerican porcupines in southern boreal forest. Can. J. Zool. 83, 1333–1342.

eu, C.W., Byers, C.R., Peek, J.M., 1974. A technique for analysis of utilization-availability data. J. Wildlife Manag. 38, 541–545.

eery, M.Z., Pauli, J.N., 2014. Shade-grown cacao supports a self-sustaining popula-tion of two-toed but not three-toed sloths. J. Appl. Ecol. 51, 162–170.

inder, L., 1997. Habitat use by resident and translocated maned sloth (Bradypustorquatus) in southeastern Brazil. Final Report to World Wildlife Found.

inheiro, M.P., Oliveira-Filho, J.A., Franc a, S., Amorim, A.M., Mielke, M.S., 2013.Annual variation in canopy openness, air temperature and humidity in theunderstory of three forested sites in southern Bahia State, Brazil. Cienc. Florest.23, 107–116.

ueiroz, H.L., 1995. Preguic as e guaribas: os mamíferos folívoros arborícolas doMamirauá. Sociedade Civil Mamirauá, Brasilia.

Core Team, 2013. R: A Language and Environment for Statistical Computing. RFoundation for Statistical Computing, Vienna, www.R-project.org/ (accessed20.07.14).

ology 80 (2015) 431–439 439

Ribeiro, M.C., Metzger, J.P., Martensen, A.C., Ponzoni, F.J., Hirota, M.M., 2009. TheBrazilian Atlantic forest: how much is left, and how is the remaining forestdistributed? Implications for conservation. Biol. Conserv. 142, 1141–1153.

Rodgers, A.R., Carr, A.P., Beyer, H.L., Smith, L., Kie, J.K., 2007. HRT: Home RangeTools for ArcGIS. Version 2.04. Ontario Ministry of Natural Resources, Centrefor Northern Forest Ecosystem Research, Ontario.

Rolim, S.G., Chiarello, A.G., 2004. Slow death of the Atlantic forest trees in cocoaagroforestry in southeastern Brazil. Biodivers. Conserv. 13, 2679–2694.

Saavedra-Rodríguez, C.A., Kattan, G.H., Osbahr, K., Hoyos, J.G., 2012. Multiscale pat-terns of habitat and space use by the pacarana Dinomys branickii: factors limitingits distribution and abundance. Endanger. Species Res. 16, 273–281.

Sambuichi, R.H.R., 2006. Estrutura e dinâmica do componente arbóreo em área decabruca na região cacaueira do sul da Bahia, Brasil. Acta Bot. Bras. 20, 943–954.

Schaefer, J.A., Messier, F., 1995. Habitat selection as a hierarchy: the spatial scales ofwinter foraging by muskoxen. Ecography 18, 333–344.

Schroth, G., Bede, L.C., Paiva, A.O., Cassano, C.R., Amorim, A.M., Faria, D., Mariano-Neto, E., Martini, A.M.Z., Sambuichi, R.H.R., Lôbo, R.N., 2014. Contribution ofcocoa agroforests to landscape carbon storage. Mitig. Adapt. Strateg. Glob.Change, http://dx.doi.org/10.1007/s11027-014-9570-7

Sikes, R.S., Gannon, W.L., 2011. The Animal Care and Use Committee of the AmericanSociety of Mammalogists. Guidelines of the American Society of Mammalogistsfor the use of wild mammals in research. J. Mammal. 92, 235–253.

Smith, A.G., Mcalpine, C., Rhodes, J.R., Seabrook, L., Baxter, G., Lunney, D., Bradley, A.,2013. At what spatial scales does resource selection vary? A case study of koalasin a semi-arid region. Austral. Ecol. 38, 230–240.

Sweitzer, R.A., Berger, J., 1992. Size-related effects of predation on habitat use andbehaviour of porcupines (Erethizon dorsatum). Ecology 73, 867–875.

Vaughan, C., Ramirez, O., Herrera, G., Guries, R., 2007. Spatial ecology and conser-vation of two sloth species in a cacao landscape in Limón, Costa Rica. Biodivers.Conserv. 16, 2293–2310.

Worton, B.J., 1989. Kernel methods for estimating the utilization distribution inhome range studies. Ecology 70, 164–168.