Int J Primatol DOI 10.1007/s10764-015-9884-7

The Distribution, Habitat Use, and Conservation Status of Three Atlantic Forest Monkeys (Sapajus xanthosternos, Callicebus melanochir, Callithrix sp.) in an Agroforestry/Forest Mosaic in Southern Bahia, Brazil Kevin Michael Flesher 1

Received: 9 June 2015 / Accepted: 27 October 2015 # Springer Science+Business Media New York 2015

Abstract If we are to develop effective conservation strategies for primates in heavily disturbed biomes, we must understand why they persist or perish in human-dominated landscapes. I investigated the potential of an agroforestry/forest mosaic in the Atlantic forest of Bahia, Brazil to sustain populations of three monkey species —capuchins (Sapajus xanthosternos), titi monkeys (Callicebus melanochir), and marmosets (Callithrix sp.)— sampling the primates in three forest and four agroforestry landscapes across a 1000 km2 landscape during a 17-yr study. I used 754 km of transects, reconnaissance walks, semistructured interviews, and key informant reports to collect information on the monkeys. Capuchins and titi monkeys are mostly restricted to the largest forests, with titi monkeys widely distributed and capuchins nearly extinct. Both species are absent from the Attalea funifera forests, but a few capuchins persist in the Brodiguesia santosii forests. Marmosets live throughout the landscape and are the only species inhabiting the agroforestry lands. The abundance data indicate that the disturbed forests characteristic of the landscape provide adequate habitat for all three species, but chronic hunting suppresses titi and capuchin populations and limits their distributions. Data from the Reserva Ecológica Michelin show that protected monkey populations can reach abundances equal to those at other protected sites, and that capuchins and titi monkeys enter agroforestry lands when protected. Long-term data indicate that deforestation rates are low and that illegal logging and hunting pressure are declining. The creation of reserves and the relatively large amount of forest left make the regional landscape a vital part of the strategy to conserve these species. This study demonstrates that agroforestry/forest landscapes can play an important role in primate * Kevin Michael Flesher [email protected] 1

Centro de Estudos da Biodiversidade, Reserva Ecológica Michelin, Igrapiúna, BA 45443-000, Brasil

K.M. Flesher

conservation when large blocks of contiguous forest with adequate escape space for hunted primates remain, but that the potential conservation value of agroforestry habitats and small forest fragments for hunted species will be realized only if hunting is controlled. Keywords Agroforestry . Atlantic forest . Hunting . Monkeys . Secondary forest

Introduction With ongoing human population growth and widespread poverty in most primate range countries, people continue to clear, degrade, and fragment natural habitats, and 48% of all primate species are in danger of extinction (Estrada 2013; Isabirye-Basuta and Lwanga 2008; Rylands et al. 2008). Although the number of protected areas has increased over the past decades, reserves will never encompass all, or even the majority, of primate populations (Estrada et al. 2012). Because of this, many primates will survive only if they are able to adapt to human-dominated landscapes (IsabiryeBasuta and Lwanga 2008). One of our handicaps in trying to initiate effective conservation actions is that we know little about the natural history of most species, which impedes us from predicting how alterations of the original forest ecosystem will affect them. Understanding how forest disturbance, the conversion of primary to secondary forest, fragmentation, matrix habitats, and hunting affect the capacity of primates to adapt to transformed forest landscapes is essential for devising conservation strategies (Isabirye-Basuta and Lwanga 2008). Although scientists are beginning to make progress in understanding how primates adapt to diverse human-dominated landscapes (Anderson et al. 2007; Arroyo-Rodríguez and Mandujano 2009; Estrada and CoatesEstrada 1996; Pozo- Montuy et al. 2011), we still know little about how most primates fare in most of them. The plight of the South American Atlantic forest primates is characteristic of that of monkeys in other devastated tropical biomes (Chapman et al. 2006; Corlett 2007). Centuries of widespread deforestation, degradation, fragmentation, and hunting have radically modified monkey habitats, and many species have disappeared from large parts of their former ranges (Canale et al. 2012; Santos et al. 1987). Much of the forest was converted to particularly hostile habitats for arboreal species, such as sugarcane, soy, and pasture (Dean 1995; Galindo-Leal and de Gusmão Camara 2003), isolating primates in small degraded fragments where they have little chance of long-term persistence (Canale et al. 2012). However, in some parts of the biome agroforestry systems replaced the native habitat, and although overwhelmingly planted with exotic trees, they are nonetheless trees, and therefore provide potential habitat for monkeys. Most of these agroforestry systems are pulpwood plantations with monocultures of Eucalyptus or Pinus covering hundreds to thousands of hectares (Galindo-Leal and de Gusmão Camara 2003), but there are also other agroforestry systems such as the cabruca cacao groves (native forests trees left to shade the cacao) of southern Bahia that cover an area >6000 km2 (Saatchi et al. 2001). Whereas primate use of cabruca cacao is fairly well studied (Cassano et al. 2012; Estrada et al. 2012; Mikich and Liebsch 2014), primate use of other agroforestry systems such as mixed tree crop groves, rubber monocultures, cacao/rubber plantations, and native palm fiber extraction

Distribution and Conservation of Three Atlantic Forest Monkeys

forests remains largely unexplored. As these habitats cover extensive areas in the range of several monkey species, it is important to understand their potential role for primate conservation. The forests of coastal Bahia are one of the richest centers of primate diversity in the Atlantic forest, with 11 species historically found along the 700-km coast and as many as 6 species occurring sympatrically (Mittermeier et al. 1982). However, with the exception of the Callithrix species, the Bahian monkeys have suffered drastic population declines over the past 70 yr following widespread deforestation, fragmentation, forest degradation, and hunting (Santos et al. 1987). In 2011 only 12.7% (24,6086 km2) of the forest remained (SOS Mata Atlântica 2012), and six species (Alouatta guariba, Brachyteles hypoxanthus, Leontopithecus chrysomelas, Sapajus xanthosternos, Sapajus robustus, and Callicebus coimbrai) are Endangered or Critically Endangered in the state and one (Callicebus melanochir) is Vulnerable (IUCN 2014). This situation makes Bahia a high-priority conservation area within the Atlantic forest, yet the lack of detailed studies outside of the Ilhéus/Una region makes conservation initiatives difficult. Despite decades of primate research in the state, we know little about how monkeys are faring in most of Bahia’s landscapes and why they are adapting or failing to adapt to particular landscape mosaics. One region of particular interest is the coastal strip between the Rio de Contas and the Rio Jequiriça, where forest still accounts for up to 40% of the landscape and agroforestry is one of the dominant agricultural practices (Flesher 2006). This agroforestry/forest mosaic encompasses an area of several thousand square kilometers in the heart of the ranges of three species (Sapajus xanthosternos, Callicebus melanochir, and Callithrix sp.), making it an important region for primate conservation. People have intensively logged, gathered forest products, and hunted these forests for centuries, but the high percentage of forest cover remaining and the dominance of agroforestry systems as the matrix habitats create conditions favorable for these monkeys. Given this, the region between the Rio de Contas and Rio Jequiriça should be a promising site for primate conservation in the largely devastated Atlantic forest of Bahia. If monkeys thrive there, understanding how they use the landscape will provide insights for designing landscapes that are both agriculturally productive and beneficial for the monkeys. If, however, they have difficulty surviving there, understanding why this is will help address the causes of monkey declines. I designed this 17-yr study to investigate the potential of agroforestry/forest landscapes between the Rio de Contas and Rio Jequiriça in southern Bahia to serve as habitat for the three monkey species, focusing my efforts on a 1000-km2 landscape in the municipalities of Ituberá, Igrapiúna, Nilo Peçanha, and Camamu. I sampled monkeys in the three forest types (lowland evergreen rainforest, piaçava resting, and jataipeba restinga) and four agroforestry systems (mixed tree crops/ manioc/forest, mixed tree crops, rubber monocultures, and cacao/rubber plantations) representative of the regional landscape, mapping their distribution, identifying the habitats they use, and measuring their relative abundance across four of the landscape types. I analyzed the effects of forest degradation, forest fragmentation, matrix habitats, and hunting on their distributions and abundances. I investigated trends in deforestation, logging, agricultural practices, the creation of protected areas, and hunting pressure to discuss the long-term outlook for primate conservation in the region.

K.M. Flesher

Methods Study Area The research landscape encompasses an area of 1000 km2, radiating 20–30 km out from the town of Ituberá (13°43.9′S; 39°09′′W) in a region of the southern Bahian coast (Baixo Sul) known as the Costa do Dendê (Oil Palm Coast). The mean annual precipitation is 2000 mm, with rainfall throughout the year and daily temperatures between 18° and 30°C [Reserva Ecológica Michelin (REM) unpub. data]. The topography is hilly with elevations of 40–586 m. The regional ecosystems include lowland Atlantic broadleaf evergreen rain forest, piaçava (Attalea funifera) restinga (forests growing on sandy soils) forests, jataipeba (Brodriguesia santosii) restinga forests, mangrove-lined estuaries, and diverse agroforestry systems with 200 cultivars planted, including 60 tree crops (Flesher 2006). Before the 1950s the region was largely forested with subsistence agriculturalists scattered throughout the hills, but by the middle of the 1970s farmers cleared most of the forest to establish large cacao and rubber plantations and smallholder farms (Devore 2014; Flesher 2006). The landscape today supports 42,640 ha of forest (ca. 40% of the study area), and although no pristine forests remain (Flesher 2006), the flora is rich (Rocha Santos and Talora 2012). The human population density is 64/km2, with people living throughout the landscape and no place in the forest is >2 km from a road. The rural economy is depressed and most people live on less than twice the minimum wage per month (ca. US$600) (Devore 2014; Flesher 2006). Landscape Characterization To characterize the regional landscape, I first conducted reconnaissance surveys using a 1:100,000 scale topographical map (SUDENE 1976). I hiked throughout the study area and made five over-flights, taking >1000 photographs to serve as a visual memory to help classify the landscape, and noting the agricultural habitats, forest cover and state of conservation, and house distribution. I used a 1999 Landsat image with a 30 × 30 m resolution to classify the forest cover using an ERDAS program (version 1), transferring the classified map to an ArcMap program (version 9.2) to calculate fragment sizes, interfragment distances, and total forest cover. I ground-truthed the forest map by hiking throughout the landscape (Flesher 2006). Based on this analysis, I mapped seven landscape types in the region: large forests, jataipeba (Brodiguesia santosii) restinga forests, piaçava (Attalea funifera) restinga forests, mixed tree crops/manioc/ forest, mixed tree crops, rubber monocultures, and cacao/rubber plantations (Appendix 1). Forest Characterization I used six categories to classify the forest according to the history of human use: pioneer forest, early secondary forest, late secondary forest, heavily logged forest, medium-impact logged forest, and lightly logged forest (Appendix 2). I gathered this information by interviewing farmers and men who previously worked on timber crews (Flesher 2006). In the case of the REM, I used aerial photographs (1964 and 1974) and

Distribution and Conservation of Three Atlantic Forest Monkeys

a SPOT image (1989) to determine the age of manioc fallows and forest clearing. I used the species lists and forest structure parameters of Rocha Santos and Talora (2012), who sampled the REM forests using the same classification system. Study Species Three monkey species inhabit the study landscape: the yellow-breasted capuchin (Sapajus xanthosternos), the Bahian masked titi monkey (Callicebus melanochir), and the black tuft-eared marmoset (Callithrix penicillata and/or hybrids with Callithrix penicillata × C. kuhlii) (Appendix 3). Hunters killed the last howlers (Alouatta guariba) in the 1950/60s, and there are no records of golden-headed lion tamarin (Leontopithecus chrysomelas) or northern muriqui (Brachyteles hypoxanthus), supporting biogeographic surveys that mark the Rio de Contas as the northern limit of their distributions (Santos et al. 1987). Census Methods I conducted the study between May 1997 and July 2013, sampling in every year except 2005, with a total of 1194 days and >5000 h in the field. I used line transects, reconnaissance hikes, semistructured interviews, key informant reports, and sighting reports to collect the data. I worked throughout the landscape between 1997 and 2004, but concentrated my efforts in the REM (forest fragments in the rubber monoculture landscape) from 2006 onward. I continued to collect data in the greater landscape after 2006 through periodic visits and key informant reports. I include several primate observations between August 2013 and April 2015 to discuss changes in ranging behavior and habitat use in the REM after the termination of the study. Transect Site Selection I established transects at 13 census sites to collect quantitative data on relative abundance and habitat use, sampling forest fragments of various sizes and the agricultural landscapes typical of the region (Fig. 1 and Table I). Owing to logistical constraints I did not have fixed transects in the piaçava or jataipeba restinga forests and relied exclusively on reconnaissance censuses, interviews, and key informant reports to collect data there. Because people regularly use the forests for hunting and wood gathering, I was unable to design a trail network free of human use. Instead I chose the least used trails available at each site (used once a week or less), opening new trails at only six of the forest sites. All trails were loops that passed through a representative sample of the site habitats. Census walks To census monkeys I walked quietly along the trails at ca. 1.2 km/h, scanning the surrounding vegetation with 7 × 35 binoculars and stopping every 50 m to listen for animal sounds (Emmons 1984). I began morning walks within the first hour of sunrise, ending no later than 11:00 h and afternoon walks between 14:00 h and dusk. I

K.M. Flesher

Fig. 1 Distribution of marmosets (Callithrix) in the municipality of Ituberá, Bahia, Brazil, 1997–2013. Closed circles = encounters; open circles = interviews. Census sites as listed in Table I.

conducted most census walks in the agricultural lands in the afternoon to minimize encounters with farm workers. I decided on sampling effort based on the census area and logistic constraints, with a minimum sample effort of 24 km and a maximum of 138.3 km (Table I). For each encounter I recorded the species, group size, time, location, strata used, and habitat. I was able to collect these data for all marmoset and capuchin encounters, but for 40.5% of the titi monkey encounters I was able to collect only location data because the groups were far from the trail. I plotted all encounters on individual species maps (Figs. 1, 2 and 3). Apart from the census walks, I spent thousands of hours in the landscape conducting other fieldwork. Each time I encountered a primate, either by sight or sound, I recorded the same information as I did for the census walks. Interview Census Design I conducted 177 interviews to plot species distributions and habitat use in areas without fixed transects. I attempted to interview at least one person in each 4-km2 grid of agricultural

Distribution and Conservation of Three Atlantic Forest Monkeys Table I Abundances of three species of primates (encounters/10 km–1), from transect walks, Ituberá, Bahia, Brazil, 1997–2013 Census site

Census area

Census Sapajus Callicebus km

100 ha in a 400-ha forest

24

Forest F1 Piauí

Callithrix

N = 13

N = 130

N = 51

0

1.25

1.67

F2 Karin/Guadalupe

500 ha in a 13,000-ha forest

43

0.23

0.23

0.47

F3 Ondulada

100 ha in a 13,000-ha forest

25

0

*

0.40

F4 Pacangê

500 ha in a 13,000-ha forest

138.26 0.83

0.76 1.81 (1.97; 1.50)a (0.49; 0.37)a

F5 Vila 5

190 ha in a 625-ha forest

116.1

*

5.00 1.89 (3.17; 5.74)a (2.67; 0.61)a

F6 Pancada Grande

170 ha in a 625-ha forest

50.25

**

5.17 0.80 (1.54; 6.49)a (1.54; 0.43)a

F7 Luis Inácio

60 ha in a140-ha forest

52.35

0.19

3.44 0.76 (1.94; 3.53)a (1.29; 0.59)a

F8 Novo Milenio

280 ha in a 580-ha forest

27

**

Total 10 km–1

475.96 0.27

Rubber R1 Michelin Club Loop

320 ha of a 7500-ha landscape 62

*

1.48

2.73

1.07

N=0

N=0

N=5

0

0

0.65

R2 Michelin Rubber Ridge 240 ha of a 7500-ha landscape 34.2

0

0

*

R3 Michelin West Rubber

0

0

0.25

360 ha of a 7500-ha landscape 39.5

Total 10 km–1

135.7

Cacao/rubber

—

—

0.37

N=0

N=0

N=1

C1 Fazenda Juliana

200 ha of a >3000-ha property 20

0

0

0

C2 Fazenda Ondulada

180 ha of a >3000-ha property 31

0

0

0.32

—

—

0.20

N=0

N=0

N = 16

Total 10 km–1

51

Mixed tree crops S1 Cavaloc Farm

30 ha in a smallholder mosaic

30.5

0

0

1.64

S2 Brito Lima Farm

41 ha in a smallholder mosaic

29.5

0

0

2.03

S3 Oliveira Farm

49 ha in a smallholder mosaic

31

0

0

1.61

Total 10 km–1

91

—

—

1.76

Total: Regional landscape

753.66 0.17

Total: Agricultural lands

277.7

*

—

1.72

0.97

—

0.79

Observed at site by author, but not during census walks

**

Observed at site by informants, but not by the author

a

Relative abundances from monitoring census: first estimate = 1997–2004 before protection; second estimate = 2011–2013, 6–8 yr after protection.

land on the SUDENE map. I chose a 4-km2 area as the minimum unit of interview density because most people were unfamiliar with areas beyond 2 km from their farms. I chose interviewees by visiting each section of the landscape and speaking with as many people as possible, revisiting areas until I had interviewed at least one person in most or all of the 4-km2 pixels. I interviewed any man who was willing to converse, determining his willingness based on his reaction to my presence. When I approached

K.M. Flesher

Fig. 2 Distribution of titi monkeys (Callicebus melanochir) in the municipality of Ituberá, Bahia, Brazil, 1997–2013. Closed circles = encounters; open circles = interviews; crosses = groups extirpated during the study period. Census sites as listed in Table I.

an individual and he showed an interest in my presence, I began the conversation by explaining my interest and asking if he would like to participate in an interview. If he was willing, I would proceed to ask questions. If he actively ignored me or said hello but showed no interest in conversation, I moved on. I did not interview women because they did not have answers to my questions. All interviews started as open-ended conversations during which I asked about the monkeys present, the habitats used by monkeys, and when the person last saw them. If I doubted a claim that a particular species was present, I asked specific details about the life history and physical characteristics of the species and accepted the information only if, through his explanation, the person convinced me of his knowledge. Owing to the illegality of hunting, I restricted questions on this subject to the 50 interviewees (28.2%) who knew me well enough to know that the interview was confidential. Starting in 2006 I collected detailed information on hunting in the REM, using hunting incident data (encounters with hunters, gunshots, dogs, traps, and hides) gathered daily by the forest guards while making their rounds. Seventeen key informants participated in the study, providing regular updates of wildlife sightings and hunting incidents. These men were acquaintances who were born

Distribution and Conservation of Three Atlantic Forest Monkeys

Fig. 3 Distribution of capuchins (Sapajus xanthosternos) in the municipality of Ituberá, Bahia, Brazil, 1997– 2013. Closed circles = encounters; open circles = interviews. Census sites as listed in Table I.

in the region and who I knew had an active interest in wildlife. I also collected sighting reports from people I knew or spoke with who told me about monkeys they had seen recently. Analysis I plotted all interview and sighting records directly on the regional map (Figs. 1–3) to determine which of the landscape types each species inhabited and their distributions within each landscape type. I used relative abundances, calculated as the number of encounters per 10 km walked, as the measure for analyzing the relative value of each landscape type for each species. I chose to use relative abundances instead of calculating densities because low encounter rates and probable violation of assumption that monkeys act independently of the observer precluded the calculation of species-specific densities (Peres 1999). Owing to the low number of replicates for some of the landscape types I was unable to use inferential statistical tests and used descriptive statistics instead. As primates in heavily hunted forests frequently adopt cryptic behavior, this may have affected abundance estimates (Remis and Jost Robinson 2012), and

K.M. Flesher

all relative abundance estimates presented in this study should be considered as minimums. This is especially true for the capuchins, which appear to learn to avoid trail areas and cease calling, but is also the case of titi monkeys, which are in the best of circumstances cryptic monkeys difficult to observe (Price et al. 2002), and marmosets that move quietly. I analyzed the REM data to test for habitat selection, using a chi-square test (I set the α at 0.05 for all statistical tests). To do so I classified each 50 m segment of 27,050 m of trails according to the history of human land use as described in the preceding text, assigning a habitat category for each monkey record according to the location on the trail. I kept a record of each trail walked and summed the habitat sampled along all the trails to calculate the total of each habitat type sampled during the study. I excluded pioneer forest encounters because these were <5% of the total sampled habitats. I sampled 2326.75 km of trails for the marmoset and titi monkey analysis, but owing to the low number of encounters of capuchins outside of the Pacangê forest, I only used the Pacangê data (999.35 km sampled) for the analysis of this species. Because lightly logged forest consisted of <5% of the sampled habitats in Pacangê, I combined this and medium impact logged habitats for the capuchin analysis. In addition to the REM habitat selection analysis, I listed the monkey encounters for the entire landscape to show the diversity of habitats used and the number of times I encountered each species in each habitat. To test for habitat stratum selection, I analyzed all visual encounters using a chisquare test. For each register I categorized the vertical habitat profile as upper third, middle third, lower third, and ground. If individuals of the same group were in different strata at the time of the encounter, I used the stratum used by the first individual sighted for the analysis. To test for the effect of hunting on the relative abundance of marmosets and titi monkeys I analyzed the 1997–2004 census data (when none of the forests were protected) from the eight forest sites using the degree of hunting pressure as the independent variable. I include only the forest sites >100 ha in the analysis to control for habitat variables that might confound interpretations of the variation in abundances between the forests and the agricultural lands. The criterion used for determining hunting pressure was the density of houses within 2 km of the forest edge, measured by walking through these areas and counting houses. House counts were: Pancada Grande 280; Karin/Guadalupe 143; Novo Milenio 117; Vila 5 85; Ondulada 68; Piauí 30; Pacangê 12; Luis Inácio 5. The correlation of decreased game abundance with increased settlement density is a well-established relationship (Jerozolimski and Peres 2003), suggesting that settlement density may be used as a proxy for hunting pressure when directly quantifying hunting behavior is not practicable. I used a Spearman rank correlation two-tailed test for the analysis. Encounter rates of capuchins were too low to include in this analysis. To measure the effects of changes in hunting pressure on monkey abundances, I used long-term walking transect data of the REM (130 km walked during each monitoring period), using a data set collected before the forests were protected (1997–2004) and the most recent census (2011–2013) 6–8 yr after protection. Owing to the low number of forest replicates (4), I was unable to use inferential statistics and use descriptive statistics instead. I also use off-transect monkey registers and interview data to discuss changes in abundances and distributions within the REM.

Distribution and Conservation of Three Atlantic Forest Monkeys

Ethical Note All research followed the legal requirements of the country. I conducted all interviews according the human subjects protocol (Rutgers University authorization #EO1-437).

Results Distribution and Abundance Each of the species responded differently to the regional landscape (Tables I and II, Figs. 1–3). Marmosets are widespread, inhabit the seven landscape types, and occupy at least 70% of the landscape. The close proximity of records suggests that they have a continuous distribution across the landscape. It is likely that their population numbers in the thousands and is panmictic over a much larger regional landscape. Although interview records suggest that they are common in all of the landscape types, transect and off-transect results indicate that they prefer some landscapes over others. Marmoset abundances were the highest in the mixed tree crop landscape, being 1.6, 4.8, and 8.8 x more abundant than in the forests, rubber, and cacao/rubber landscapes, respectively. While interviews suggest that marmosets are common in the rubber monocultures and cacao/rubber groves, it is not common to see them away from the forest edge and I only recorded them four times in the rubber monocultures and once in the cacao/rubber groves during transect walks and saw them infrequently in these habitats during the hundreds of hours I spent conducting other work there. Titi monkeys occupy ca. 20% of the landscape, inhabiting three of the landscape types. However, they are widespread only in the main forest block, have a patchy distribution in the mixed tree crop/manioc/forest landscape, and are almost extinct in the mixed tree crop landscape. The regional population appears to be divided into ≥16 isolated subpopulations, with the largest subpopulation along the eastern ridge of the main forest block. Their total population may be >1500 (using Müller's (1995) Table II Results of interview censuses for three monkey species by landscape type, Ituberá, Bahia, Brazil, 1997–2013 Landscape type

No. of interviews No. of key informants Sapajus Callicebus Callithrix (%) (%) (%)

Large forest

30

11

30

70

73.3

Piaçava restinga forest

10

1

0

0

60

Jataipeba restinga forest

6

0

83.3

0

83.3

Mixed tree crops/manioc/forest 17

2

17.6

52.9

82.4

Mixed tree crops

62

6

0

6.5

82.3

Rubber monocultures

32

8

6.3

0

68.75

Cacao/rubber

20

6

0

5

80

Total

177

17a

10.7

20.3

76.8

Data are expressed as the percentage of interviewees who had seen the species in each landscape type. a

Some key informants provided information on more than one landscape type.

K.M. Flesher

estimation of 24-ha home ranges and an average group size of 3), with perhaps as many as 1000 monkeys in the largest subpopulation. Though I never encountered titi monkeys in the mixed tree crop landscape, interviews indicate that they persisted in four fragments, inhabiting three 40–100-ha and one 16-ha fragment, but it appears that at least two of these groups were extinguished during the study period. None of the interviewees remember titi monkeys inhabiting either the piaçava or jataipeba restinga forests. Capuchins occupy <10% of the landscape, and I recorded them only during transect walks in three of the eight forest sites. Although I recorded them off-transect in the Vila 5 forest and informants confirmed their presence in two other forests, they are absent from the remaining two forests. This monkey inhabits the large forest block, the mixed tree crop/manioc/forest landscape, and the jataipeba restinga forests, but interviews and searches indicate that they have a patchy distribution in each of these landscape types. Counting the known groups I estimate that there are likely 14 groups in 9 subpopulations, and the total population may not be >70–100 monkeys. The largest subpopulation is that of the Pacangê, Luis Inácio, and Itapema I forests, where at least three groups, and perhaps as many as five, have contiguous home ranges. The remaining groups are mostly isolated from one another by distances of several to >10 km and the subpopulations are unlikely to have contact. None of the informants remember capuchins inhabiting the piaçava restinga forests and few animals remain in the jataipeba restinga landscape. Use of Forest Habitats Each of the species used all of the forest types found in the main forest block (Table III), but each used them differently (Table IV). Marmosets selected strongly for agricultural fallows in all stages of development and against the other forest habitats . Titi monkeys showed a significant preference for mature fallows and heavily logged forests, using the other forest types less than expected. Use of Agricultural Habitats Informants reported marmosets as present in all of the agroforestry habitats. I encountered marmosets 284 times in agricultural lands (51.5% of the records), seeing them in rubber, cacao, oil palm, and piaçava groves and in orchard trees, but 56.3% of these records were in narrow pioneer forests along the waterways (Table III). Marmosets used the mixed tree crop groves and REM rubber groves with pioneer forest in the inter-rows extensively, but the low number of records in the rubber with cleared interrows and cacao/rubber groves indicates that these are not productive habitats for marmosets. Though informants occasionally see marmosets foraging in the rubber groves, they appear to use this habitat mostly to move between riparian forests. They enter the cacao/rubber groves along the forest edge but spend little time there. There are several groups that live on the outskirts of town that move between hillside forests and people’s gardens to feed in the orchards. One group had a home range completely within town, living in the orchard and oil palm groves of a rubber factory. I did not encounter titi monkeys in the agricultural lands during the study, and know of only two cases in which they entered agricultural land. The first was of a group of

Distribution and Conservation of Three Atlantic Forest Monkeys

Table III Total number of records of three species of monkey by habitat type, Ituberá, Bahia, Brazil, 1997–2013

Habitat type

Forest

Callithrix

Callicebus

Sapajus

N = 552

N = 413

N = 57

28

1

1

Pioneer forest Early secondary

79

61

6

Late secondary

39

59

6

Heavily logged

54

125

11

Medium-impact logging

46

131

24

Lightly logged

6

36

5

Jataipeba restingas

7

—

3

Riparian pioneer forest

160

—

—

Rubber monocultures

41

—

—

Mixed tree crops

37

—

—

Gardens

19

—

—

Piaçava restingas

18

—

—

Beach/coconut

4

—

—

Cacao/rubber

2

—

—

Agricultural

protected titi monkeys that regularly entered a mixed tree crop smallholder farm adjacent to a large forest, occasionally sleeping in remnant forest trees and feeding on bananas. The other case was of a group that an informant saw in a banana tree along the forest edge on a cacao/rubber plantation. Table IV Observed and expected forest habitat selection for three species of monkey, Ituberá, Bahia, Brazil, 1997–2013, with results of chi-square tests demonstrating whether or not each species selected forest habitats according to their availability Callithrix N = 161

Callicebus N = 390

Sapajus N = 42

Forest habitat type

Expected Observed Expected Observed Expected Observed

Early secondary

26

49

62

58

12

5

Late secondary

20

34

49

59

6

6

Heavily logged

40

36

96

119

7

9

Medium-impact logged

60

36

146

122

—

—

Lightly logged

15

6

37

35

—

—

—

—

—

17

22

Medium-impact logged + lightly loggeda — df

4

4

3

χ2

46.32

11.28

6.00

P

<0.001

0.024

0.112

Only Reserva Ecológica Michelin data used. For Callithrix and Callicebus the total km sampled = 2327 km; for Sapajus only Pacangê data used = 999 km. a

Medium-impact logged and lightly logged habitats combined for Sapajus because the expected value for LL was <5.

K.M. Flesher

Though capuchins do not inhabit any of the agricultural landscapes, they do enter them on occasion, especially in the jataipeba restingas and since 2008 in the REM. The groups of the jataipeba forest landscape raided watermelon and squash gardens near Barra de Serinhaém until the 1970s and locals reported that during the study period monkeys entered oil palm groves to feed on the fruits and to sleep in the trees. In the REM they entered abandoned pupunha palm (Bactris gasipaes) and rubber groves adjacent to the forest, crossing the ground and foraging on the fruits of the former and foraging among the bromeliads in the rubber trees. Capuchins recolonized the Luis Inácio forest during the study period by crossing 300 m of rubber monocultures from the Pacangê forest, and by the close of the study period had incorporated these rubber groves as part of their home range. An informant saw capuchins feeding on cacao fruits on a cacao/rubber plantation adjacent to the large forest block. Vertical Use of Habitat Marmosets, titi monkeys, and capuchins all selected strongly for the upper stratum of the habitats they used (Table V). While marmosets selected the middle stratum almost as expected according to availability, neither the titi nor capuchins did so. Marmosets use the ground when necessary to cross open areas or roads between tree patches, but I never saw them foraging on the ground and they crossed open ground quickly. I never recorded capuchins on the ground but informants saw them descend to eat bananas at research capture sites, in pupunha palm groves when moving out of the forest to reach the fruiting trees, and crossing up to 200 m of open sand flats between forest patches in the jataibeba landscape. Informants reported that capuchins forage for crabs in the understory of these forests and in the mangroves. Until the 1970s they would go to the ground to raid gardens in this landscape as well. The only case I recorded of the titi monkeys going to the ground was that of a habituated group that did so to eat bananas on a smallholder farm.

Table V Vertical habitat use by three species of monkey, Ituberá, Bahia, Brazil, 1997–2013, with the results of the chi-square tests demonstrating that the three monkey species do not select forest strata according to their availability Callithrix N = 208

Callicebus N = 187

Sapajus N = 53

Stratum

Expected

Observed

Expected

Observed

Expected

Observed

Ground

52

5

46.75

0

13.25

0

Lower 1/3

52

12

46.75

2

13.25

2

Middle 1/3

52

45

46.75

18

13.25

3

Upper 1/3

52

146

46.75

167

13.25

48

df

3

χ2 P

3

3

244.11

416.57

121.87

<0.001

<0.001

<0.001

Distribution and Conservation of Three Atlantic Forest Monkeys

Effect of Hunting Marmosets are not a game species and thus are not directly affected by hunting. No one ever listed titi monkeys as a preferred game species, although when asked directly, people considered them edible and claimed that people kill them opportunistically. The unprotected forest data from 1997–2004 indicate that the number of houses within 2 km from the forest edge was not significantly related to the abundances of marmosets (rs = –0.0714, df = 6, P = 0.88) or titi monkeys (rs = –0.3832, df = 6, P = 0.35), corroborating informant claims. However, the long-term monitoring data from the REM show that in the 8 yr after the initiation of guard patrols, marmoset abundances decreased by 71.5% (1.65/10 km–1 vs. 0.47/10 km–1) while titi monkey increased by 69.7% (2.41/10 km–1 vs. 4.09/10 km–1). In the Pancada Grande forest (the most heavily hunted REM forest from 1997 to 2004), titi monkey abundances increased 321.4% (1.54/10 km–1 vs. 6.49/10 km–1) and marmoset abundances decreased 258.1% (1.54/10 km–1 vs. 0.43/10 km–1). A comparison between titi monkey abundances in the heavily hunted Karin/Guadalupe forest and the protected Pancada Grande forests (data from 2011–2013 monitoring census), which were similar in terms of habitat availability, shows a difference of 2,722% (0.23/10 km–1 vs. 6.49/10 km–1). These data suggest that hunting pressure does affect these species, but differently. All hunters interviewed considered hunting as a principal cause of the decline of capuchins, although not all hunters killed these monkeys. The lack of encounters during the census walks at five of the eight forest sites, the small group sizes, and the frightened behavior of the capuchins I encountered all indicate a heavily persecuted population. The most lightly hunted groups (Pacangê and adjacent forests) had 15–20 individuals while all of the heavily hunted groups (all others encountered) had 4–6 individuals. All capuchins reacted negatively to my presence by alarm calling and fleeing and only after several years of protection did the Pacangê groups start to tolerate researchers’ presence. The Pancada Grande/Vila 5 group remained extremely wary even after 8 yr of protection.

Discussion The three species persist in the study landscape despite decades of deforestation, forest degradation and fragmentation, and intensive hunting, but each has adapted to the mosaic distinctly. Marmosets thrive in this landscape, exploiting the agroforestry lands, small and large fragments of both logged and fallow forests, the restinga forests, mangroves, and towns, but abundances vary across landscape types. Titi monkeys adapt well to the logged and fallow clay soil forests >100 ha, but their abundances varied greatly between forest sites. They rarely use the agroforestry lands, surviving in only one or two of the small forest fragments therein, and are absent from both restinga forests. Capuchin monkeys adapt well to the logged and fallow clay soil forests and the jataipeba restingas, and will enter rubber and pupunha groves when protected, but are absent from the majority of the landscape. Hunting appears to be the principal factor affecting the distribution and abundances of titi monkeys and capuchins as it is for many monkeys throughout the tropics (Chapman et al. 2006; Corlett 2007; Peres 1990). Informants >60 yr old remember

K.M. Flesher

capuchin monkeys as being common and widespread, but few people today are familiar with this animal. All forests are easily accessible to hunters and, outside of the REM, capuchins suffer chronic hunting pressure. This lack of escape space and their low reproductive potential (Auricchio 1995) make it difficult to compensate for hunter offtake. One possible explanation for the persistence of capuchins in hunted forests is that some hunters do not kill monkeys. Two of the principal hunters in the forests with capuchins claimed that they do not kill these monkeys because they resemble people, and this is likely the reason they survived there. A second explanation is that capuchins adopt cryptic behavior where they are heavily persecuted, reducing vocalizations and avoiding trail areas. I saw capuchins only once during 1153 h spent in the REM Vila 5 Forest with sighting intervals of up to several years despite the frequent use of this forest by scientists and forest guards. However, behavioral adaptations to avoid hunters were effective only in forests >500 ha and not even in all of these (Remis and Jost Robinson 2012), and hunters continued to kill capuchins during the study period. Hunters eliminated titi monkeys from almost all of the small fragments and suppressed their abundances in the large forest blocks. Although hunters do not rank the titi monkey as a preferred game animal (Santos et al. 1987), the best explanation for the variation in abundances between forests is hunting pressure, as the flora and disturbance histories of the sampled forests were similar (Flesher 2006). Nonetheless, they persist in most forests >100 ha despite decades of chronic hunting pressure. Because of their cryptic coloration and behavior and their strategy of fleeing upward and hiding in dense vegetation when disturbed, titi monkeys are able to escape hunters on most occasions, as most hunters are unwilling to waste time pursuing a monkey that offers little return for effort invested (Alvard 1995). In parts of the landscape with only small (<30 ha) forest fragments, the constant proximity of people makes encounters with monkeys a certainty and their cryptic behavior and coloration were not sufficient to protect them. Hunting pressure keeps titi monkeys from using agroforestry lands to cross between forests and only in early 2015, after a decade of protection and a year after the REM rubber groves were abandoned, did titi monkeys venture into the agroforestry landscape. Unless hunting pressure is eliminated, the smaller subpopulations will remain isolated in their respective fragments and may not be viable in the long term. Hunting may benefit marmosets by suppressing populations of their predators (Hilário and Ferrari 2015; Michalski and Peres 2005), as indicated by the decline of marmosets in the REM after the initiation of guard patrols and the subsequent recovery of the fauna (Flesher and Laufer 2013). Although part of the recorded decrease could be the result of marmosets adopting more cryptic behavior because of an increase in predator abundances, predation pressure likely increased following the recovery of the fauna. Differences in the availability and quality of escape space from nonhuman predators is a possible explanation for the difference in marmoset abundances between agroforestry landscapes (Chiarello and de Melo 2001; Oliveira and Dietz 2011; Sussman and Kinzey 1984). The regional carnivore community includes raptors, snakes, and mammals known to hunt marmosets (Ferrari and Beltrão-Mendes 2011; Hilário and Ferrari 2015; Sussman and Kinzey 1984). Marmoset abundances were lowest in the rubber and cacao/rubber groves where the canopies are open and animals travel exposed, whereas they reached their highest abundances in the mixed tree crop groves with denser and more structurally complex canopies (Rylands 1996). A second

Distribution and Conservation of Three Atlantic Forest Monkeys

landscape feature that may explain this difference is quick recourse to forest for escaping predators. Marmosets rarely rest in the agroforestry habitats, preferring to retreat to the dense canopies and vine tangles of secondary forests for shelter (Silva Ferreira 2014), and the scarcity of natural vegetation in the cacao/rubber groves may partially explain why marmoset abundances are lower in this landscape type than in the others (Oliveira and Dietz 2011). Food resources are also variable between landscape types, and the mixed tree crop groves provide more fruit resources and possibly a more diverse arthropod community than either the rubber or cacao/rubber groves, which also might explain the difference in abundances (Rylands 1996). The reasons for the absence of titi monkeys and capuchins from the agricultural lands are difficult to assess because of the confounding effect of hunting (Blanco and Waltert 2013). However, evidence from other landscapes indicates that when not hunted titi monkeys and capuchins can persist in small fragments for decades after major deforestation events, often surviving radical transformations of their native habitat (Benchimol and Peres 2014; Cardoso de Sousa et al. 2008; Michalski and Peres 2005). As such, it is probable that in the absence of hunting, both titi monkeys and capuchins would be more widespread and likely to exploit at least parts of the agroforestry landscapes (Cassano et al. 2012; Kierulff et al. 2005). The small distances (≤400 m) between fragments and an agroforestry matrix with trees that provide adequate movement substrates suggests that these highly arboreal monkeys should be able to move between fragments throughout the landscape (Blanco and Waltert 2013; Pozo-Montuy et al. 2011) while avoiding terrestrial predators such as dogs (Frigeri et al. 2014; Oliveira et al. 2008). The mixed tree crop landscape with diverse edible agricultural fruits and closely spaced fragments should be habitable for both species in the absence of hunting. It is easy to imagine monkeys of both species incorporating several fragments in their home ranges (Chagas and Ferrari 2011) and exploiting agroforestry and forest habitats. It is unclear, however, whether the pioneer forests that dominate the natural vegetation of the rubber and cacao/rubber landscapes could sustain these monkeys without their having recourse to the more floristically diverse large forest blocks (Heiduck 2002; Trevelin et al. 2007). Capuchins began using the REM rubber and pupunha groves once hunting was eliminated, and Sapajus species show a remarkable adaptability to limited food supplies in other parts of the biome (Freitas et al. 2008; Galetti and Pedroni 1994; Ludwig et al. 2006; Mikich and Liebsch 2014; Siemers 2000), so it is possible that they would be able to survive relying solely on resources from the pioneer forests. The scarcity of large trees adequate for shelter (Heiduck 2002; Kinzey and Becker 1983) and a yearly period of fruit scarcity lasting 5 mo (Flesher 2006) suggest that titi monkeys would be vulnerable to starvation and predation in the rubber and cacao/rubber landscapes even in the absence of hunting. However, the fact that they began colonizing the REM abandoned rubber groves in 2015 shows that they can find food and shelter in this habitat of rubber trees and pioneer vegetation. Whether they can live exclusively in this habitat without recourse to mature forest stands remains to be seen. It is surprising that the piaçava restingas do not support large populations of titi monkeys and capuchins as these species or those of the same genera use disturbed sandy soil forests in other parts of the biome (Beltrão-Mendes et al. 2011; Kierulff et al. 2005), and other highly frugivorous arboreal animals, e.g., Potos flavus, inhabit this

K.M. Flesher

forest. Furthermore, capuchins inhabit the jataipeba restingas to the east and the forests to the west, so they had to have inhabited this part of the landscape at some time in the past. Interviews with men in their 60s whose fathers hunted these forests have no memory of either species living there, so they must have been extirpated >80 yr ago. The jataipeba forest landscape, with its forest patches, mangroves, and oil palm groves, offers quality resources for capuchins and they were common in this landscape until the 1970s. Capuchins readily cross the treeless sand flats between forests, which enables them to incorporate several remnants in their home ranges. The mangroves provide rich foraging grounds and the oil palm trees nutritious fruits, but chronic hunting pressure prevents them from exploiting the full potential of this landscape. There is no indication that titi monkeys ever inhabited the jataipeba restingas, even though the coastal forest was continuous until the 1950s. Today, with only small fragments remaining separated by treeless sand flats it is unlikely that this landscape is suitable for these highly arboreal monkeys (Müller et al. 1997). Primate abundances in the study landscape are comparable to those of other Atlantic forest sites for the same species (Bastos Moreira 2009; Kierulff et al. 2005) or species of the same genera (Chagas and Ferrari 2011; Chiarello 1999; Faria Franco Negrão and Valladares-Pádua 2006; Oliveira et al. 2003; Price et al. 2002; Trevelin et al. 2007), indicating that the study landscape provides quality habitat for these species. Titi monkey abundances in the REM after 8 yr of protection were comparable to the highest relative abundances recorded for the genera in the Atlantic Forest, even when just counting visual encounters (Chagas and Ferrari 2011). Titi monkey abundances in the REM after 8 yr of protection (4.09/10 km–1) were 31.5 and 37.2 × greater than those for the same species in the Serra das Lontras/Javi (0.13/10 km–1) and Una Biological Reserve (0.11/10 km–1), respectively (Bastos Moreira 2009). Capuchin abundances in the REM (0.83/10 km–1), were higher than those recorded in the Serra do Conduru State Park (0.35/ 10 km–1) and the Una Biological Reserve (0.58/10 km–1) (Kierulff et al. 2005), and lower than those in only one fragment in Sergipe (1.4/10 km–1) (Chagas and Ferrari 2011). Marmoset abundances fell within the range of those registered at other sites in the region (Bastos Moreira 2009). These results show that the study landscape provides as good habitat as exists within the range of these species, and that even with the history of intensive logging and manioc cultivation, the conservation potential of the landscape for these species is still high. The study raises several questions that could not be answered with the sampling methods used. Hunting pressure made calculating monkey densities impossible, which limits our ability to understand the carrying capacity, and thus the full conservation potential, of the landscape for these primates. I spent limited time searching for monkeys in some parts of the largest forest blocks, and given the capacity of capuchins to avoid people, a more thorough search throughout these forests might have revealed a larger population than that reported here. Studies designed to identify the particular habitat features that influence the monkeys' choices to use particular landscape types are needed for a more fine-grained understanding of the potential of agroforestry and forest habitats to serve as habitat. We also need to understand how the widespread use of agrochemicals in certain landscape types influences the monkeys’ capacity to inhabit these areas.

Distribution and Conservation of Three Atlantic Forest Monkeys

Prospects for the Future Marmosets are widespread and abundant and have undoubtedly benefitted from the forest disturbances that have favored the growth of the pioneer vegetation they prefer (Sussman and Kinzey 1984). As they are not hunted, they are likely to continue thriving in the landscape for the foreseeable future. Hunting, deforestation, and forest degradation and fragmentation, however, continued to threaten titi monkeys and capuchins at the end of the study. Although there was a surge of environmental law enforcement between 1998 and 2002 when the authorities successfully closed the regional timber mills and created the APA (Área de Proteção Ambiental: Environmental Protection Area) de Pratigi, little has happened since and there are no longer any federal or state environmental law enforcement officials stationed in the study landscape. Nonetheless, some of the threats appeared to decline throughout the study period. Hunting is the most imminent threat to both species, and although the hunting tradition remains strong among the elder and there are still young men who enjoy hunting, interviews indicate that people hunt much less than they did in the past (Flesher and Laufer 2013). Few people today depend on game meat for subsistence and hunting is mostly a leisure activity, but most of the rural youth prefer town life for leisure. Even the relatively high prices paid for game meat in town are not enough inducement for most to take up hunting. Although this decline in hunting pressure may allow for a population recovery for less preferred game species such as the titi monkey, whether hunting will decline sufficiently to enable the capuchin population to recover remains to be seen. Farmers cleared 650 ha of the largest forest block, including the Fazenda Karin forest (F2), which supported one of the few remaining capuchin troops and several titi monkey groups. Elsewhere in the landscape smallholders cleared small plots of secondary forest to plant manioc, but overall, the forest mosaic remained stable over the 17-yr study period, with only 1.5% cleared. The clearing of the Karin forest fragmented the main forest block along BA-250, but other deforestation events did not alter the fragmentation pattern that existed at the beginning of the study. As people were still clearing forest at the end of the study period, deforestation still represents a long-term, if not imminent, threat. Although illegal loggers continue to degrade forests by removing the larger trees, logging pressure declined throughout the study period. This was mostly due to the closing of the regional timber mills and the environmental police check posts along the federal highways, which make it increasingly difficult to transport illegally felled logs outside of the region. Another reason was that several of the largest landowners who had previous allowed logging in their forests prohibited it, denying loggers access to some of the principal forests. Nonetheless, there remains a chronic local demand for hardwoods, and without a viable alternative to local timber, logging is unlikely to stop completely and will continue to degrade the forests in the coming years. Efforts to legally protect the forests advanced throughout the study period. The most significant achievement in terms of area affected was the creation (1998) and expansion (2001) of the multiple-use APA de Pratigi, which now encompasses most of the regional landscape (857 km2). Owing to the lack of effective management, however, the reserve has largely failed to stop illegal resource use and offers dubious protection

K.M. Flesher

for the monkeys. Several large-holders (people owning >50-ha properties and who hire laborers to work their farms) created reserves by registering their forests with the state and/or federal authorities while others simply declared their forests Bprotected^ by posting signs. Although this has been effective at stopping most illegal logging, it has been ineffective for stopping hunting because few property owners have hired forest guards. Unless there is a willingness on the part of the large-holders to pay for forest guards or outside assistance to help finance guard patrols, it is unlikely these Bprotected^ forests will provide refuge for these monkeys. The one exception was the REM, as Michelin invested in active protection by hiring forest guards that resulted in a substantial increase in wildlife abundances, including those of capuchin and titi monkeys (Flesher and Laufer 2013). Despite the many threats facing the titi and capuchin monkeys, the pressures on the forest today are less than they were when the study began and there is still hope for the long-term persistence of these species in the study landscape. The mosaic of logged and fallow forests provides suitable habitat, and as there is still a significant amount of contiguous forest cover, there is ample space for these monkeys to thrive. The relatively low deforestation rate and the fact that 40,000 ha of forest remains, means that deforestation and fragmentation are not imminent threats to either of the species and there is still time to initiate conservation actions. The long-term decline in hunting allows for a cautious optimism that monkey populations will recover with time and expand their ranges into unoccupied forests, perhaps one day using the agroforestry lands. Although the APA status of the landscape has offered dubious protection for the monkeys, the legislation necessary for sustainably managing the landscape exists, providing a legal foundation for future conservation actions. The creation of the REM and the fact that most large-holders have declared their forests protected is encouraging, and though much remains to be done, the experience in the REM shows that effective protection is possible (Flesher and Laufer 2013). The abundance of data shows that the populations of both species in the study landscape can be equal to or greater than those in the largest reserves left within their ranges, making this landscape a priority site for the conservation of these species despite that fact that the agroforestry lands currently offer little value for either capuchin or titi monkeys. Acknowledgments For help during my doctorate I thank Pete Vayda, Tom Rudel, Louise Emmons, George Clark, Colleen Hatfield, Peter Morin, Brad Walters, Kristi MacDonald, and Jay Kelly. At Michelin, I thank Bernard Francois, Lionel Barré, Gerard Bockiau, Eric Cavaloc, Ubirajara Swinerd, André Souza dos Santos, Antônio Camilo de Jesus, Edenildo Norberto Oliveira, Rosivan dos Santos, and Valdir Kleber de Jesus Sena. For logistic and intellectual support in the field, I thank Marcia Maria Flesher, Richard Hartley, Ronaldo Jupará Dread and Luis Carvalho, Jonathan Devore, Juliana Laufer and Andrea Dechner. Special thanks to the editor and two anonymous reviewers for comments that improved the manuscript.

Appendix 1 Landscape types in the municipality of Ituberá, Bahia, Brazil, 1997–2013 Large Forests Forests >100 ha, mostly concentrated in a >13,000 ha block that extends for 30 km along two ridges in the western part of the landscape. The forest is mostly <2 km wide, 5 km at its widest and several hundred meters at its narrowest, with forest gaps of 10–30 m and a maximum of 2 km. Large-holders own most of these

Distribution and Conservation of Three Atlantic Forest Monkeys

forests as part of their cacao/rubber plantations and logged them repeatedly over the past 70 yr. People hunt throughout these forests. Jataipeba (Brodiguesia santosii) Restingas Forest on coastal sand flats separated by open low herbaceous vegetation with ground bromeliads and small bushes and shrubs. Forest covers 8–10% of landscape with sixty remnants, mostly 10–40 ha, and the largest is 215 ha. Interfragment distances are mostly ≤200 m, with a maximum distance of 1 km. Open sand flats, wetland, mangroves, and large-holder oil palm and coconut monocultures make up the rest of the landscape. The few people who inhabit this landscape live in four coastal villages, but people hunt throughout. Piaçava (Attalea funifera) Restingas Piaçava forests occupy 13,000 ha of the coastal sand hills east of highway BA-001, split into six principal blocks separated by an estuary and oil palm plantations, with interfragment distances mostly ≤30 m. Mangroves, wetlands, and water cover 35% of the landscape, with oil palm plantations and manioc fields occupying ≤6%. Owing to the practice of clearing and burning forest to increase piaçava palm densities, the forest is secondary vegetation. No agrochemicals are used in the piaçava groves. Few people inhabit this landscape, mostly concentrated in village of Jatimane, but people hunt throughout. Mixed Tree Crops/Manioc/Forest Large and smallholder farms with mixed tree crops and manioc on the smallholder farms; commercial manioc plantations on medium size farms; and cacao/rubber, piaçava, and forest on the large-holder properties. The use of herbicides, fungicides, and pesticides is widespread on the cacao/rubber plantations but varies on the smallholder farms according to the capacity of the farmer to purchase them. This landscape type occupies two areas of ca. 6000 ha each. Forest covers 40–65% of the landscape with interfragment distances rarely >200 m. People live and hunt throughout the agricultural lands, with houses generally ≤1 km apart. Mixed Tree Crops Mostly 5–15 ha smallholder farms with a high diversity of agroforestry crops including cacao, rubber, cloves, allspice, cupuaçu, bananas, black pepper, guarana, piaçava, orchard fruits, oil palm, and manioc, with mixed plantings over small spatial scales. The use of herbicides, fungicides, and pesticides varies according to the capacity of the farmer to purchase them. Forest covers 20–30% of the landscape and consists of chronically disturbed logged forest and manioc fallows generally <10 ha (only 13 fragments are ≥40 ha and the largest two are 129 and 512 ha), with interfragment distances rarely >200 m. People live and hunt throughout the landscape, with houses generally ≤1 km apart. Rubber Monocultures 7500 ha of large-holder rubber monocultures planted at a density of 500 trees/ha, with lesser areas planted with cacao/rubber. Herbicides are applied once or twice each year to maintain the rubber tapping trails open. Natural vegetation occupies 21% of the landscape and consists of pioneer forests along streams, <10 ha pioneer forest fragments on rocky outcrops, extensive cattail wetlands, and pioneer bush in the rubber inter-rows. Interfragment distances are ≤400 m. People are concentrated in three villages, with most of the

K.M. Flesher

landscape visited by rubber tappers on 3-day rotations. Hunting was widespread before the creation of the REM in 2006, after which it was reduced substantially because of daily guard patrols. Cacao/Rubber Plantations Large-holder cacao/rubber plantations form a largely contiguous block covering thousands of hectares in the 1–2 km wide valleys between the main forest ridges in the western section of the study landscape. Cacao/rubber groves are planted at a density of 900 and 400 trees/ha, respectively, with the rubber trees providing shade for the cacao trees. The use of herbicides, fungicides, and pesticides is widespread. Cattle pastures and other crops occupy 17% of the landscape, with small forest fragments and wetlands occupying <6%. Interfragment distances are several hundred meters to >1 km. People live concentrated in plantation villages and most of the landscape is visited by workers on a daily basis. Owing to the scarcity of game in these plantations, hunters mostly hunt in the nearby large forests.

Appendix 2 Forest classification in the municipality of Ituberá, Bahia, Brazil, 1997–2013 Pioneer Forest Manioc fallows, abandoned pasture, and riparian forests in agricultural lands cut and burned <25 yr ago. The structural profile is of dense low thickets of herbaceous and woody vegetation growing to 5–10 m with an abundance of vines and emergents (frequently Schefflera morototoni) reaching 12+ m. Dominant genera include Cecropia, Tapirira, Schefflera, Senna, Byrsonima, Stryphnodendron, Inga, Vismia, Kielmeyera, Himatanthus, Henrietta, Miconia, Tibouchina, Solanum, Piper, and Cyperus grasses. Early Secondary Forest Manioc fallows 25–50 yr old. Dense continuous vegetation to 4–8 m with spaced taller trees to 10+ m. Vines are generally abundant, forming dense mats and sometimes covering parts of the canopy, and thin lianas are moderately abundant. Dominant genera include Tapirira, Senna, Schefflera, Inga, Rollinia, Pogonophora, Byrsonima, Casearia, Albizia, Simarouba, Tibouchina, Miconia, Henrietta, Aparisthmium, Euterpe, Guettarda, Lacistema, Licaria, Himatanthus, Ocotea, Rauvolfia, and Pourouma. Late Secondary Forest Manioc fallows >50 yr old. Fairly continuous growth to 8–15 m, often with a denser lower stratum to 3–6 m. Dominant genera include Pourouma, Euterpe, Helicostylis, Vochysia, Carpotroche, Cupania, Tapirira, Maprounea, Inga, Byrsonima, Cordia, Clethra, Nectandra, Ocotea, Xylopia, Symphonia, Swartzia, Tachigali, Lacistema, Eschweilera, Apeiba, Miconia, Tibouchina, Guapira, Tetrastylidium, Hieronymam, and Simarouba. Heavily Logged Forest All large and most medium-sized trees logged 50–70 yr ago. Sapling thickets with dense continuous growth to 8–15 m and spaced medium-sized trees to 20 m. Dominant genera include Helicostylis, Tovomita, Eschweilera, Pogonophora, Garcinia, Pourouma, Virola, Anartia, Cordia, Apeiba, Eugenia,

Distribution and Conservation of Three Atlantic Forest Monkeys

Sloanea, Brosimum, Inga, and Swartzia with Sclerolobium, Balizia, Parkia, Eriotheca, Albizia, and Nectandra emergents. Medium-Impact Logged Forest Most large and some medium trees removed 50–70 yr ago. The structural profile is of a moderately dense understory to 2–5 m with less dense continuous vegetation to 10–15 m and small groves of mature trees or isolated individuals to 20+ m. Dominant genera include Helicostylis, Pogonophora, Garcinia, Tovomita, Eschweilera, Eugenia, Pourouma, Cordia, Swartzia, Amaouia, Pouteria, Xylopia, Mabea, Trichilia, Guettarda, Siparuna, Casearia, Diploon, Ecclinusa, Micropholis, Pradosia, Brosimum, Euterpe with large Eriotheca, Sloanea, Copaifera, Virola, Licania, Caryocar, Lecythis, Symphonia, Arapatiella, and Manilkara remnant trees. Lightly Logged Forest Selectively cut 50–70 yr ago with few trees removed per hectare. Fairly open continuous growth to 15–25 m with emergent old growth trees to 30–40+ m. Lianas sometimes form dense tangles on the larger trees. Dominant genera as for medium-impact logged forest.

Appendix 3 Study species in the municipality of Ituberá, Bahia, Brazil, 1997–2013 Sapajus xanthosternos The capuchin’s range includes the Bahian Atlantic Forest north of the Rio Jequitinhonha to the south bank of the Rio São Francisco delta in Sergipe, extending to the semideciduous forests and xeric shrublands as far west as the Rio São Franciso (Coimbra-Filho et al. 1992; Kierulff et al. 2005). Recent studies show that these omnivorous monkeys live in groups of up to 20+ monkeys ranging over areas of up to 1000 ha (Kierulff et al. 2005). They use disturbed and intact forests and some agroforestry lands (Cassano et al. 2012; Kierulff et al. 2005). This species is listed as Critically Endangered with the decline due to habitat loss, fragmentation, and principally hunting (Kierulff et al. 2008). Callicebus melanochir The titi monkey inhabits the evergreen forests of coastal Bahia south of the Rio Paraguaçu to the Rio Itaúnas in Espírito Santo, extending west as far as the semi-deciduous forests in northeastern Minas Gerais (Santos et al. 1987). The species is largely frugivorous (Heiduck 1997), living in small groups occupying home ranges of up to 24 ha (Müller 1996). Although adapting well to disturbed forests (Heiduck 2002), they are a forest-dependent species that rarely enters agricultural lands (Cassano et al. 2012; Flesher 1999). It is listed as Vulnerable, with the main threats being habitat loss and fragmentation (Veiga et al. 2008). Owing to their small size and cryptic behavior they are not a preferred game species (Santos et al. 1987). Callithrix There is some doubt as to the species of Callithrix inhabiting the Baixo Sul region (IUCN 2014), but the species in most of the study landscape resembles Callithrix penicillata more than any other species, following the criteria established by Coimbra-Filho et al. (2006). The marmosets along highway BA-001 appear to be a mix of C. penicillata and C. kuhlii. According to the IUCN (2014) C. penicillata has a

K.M. Flesher

wide distribution in the cerrado savannas of seven states and C. kuhlii inhabits the evergreen and semideciduous forests between the Rio de Contas and Rio Jequitinhonha. Marmosets use forests, agricultural lands, and urban areas, feeding on exudates, insects, small vertebrates, and fruit (Duarte and Young 2011; Fonseca and Lacher 1984). Though sometimes captured for the wildlife trade (Mittermeier et al. 1982), it is common in many areas and C. penicillata is listed as Least Concern and C. kuhlii as Near Threatened (IUCN 2014).

References Alvard, M. (1995). Intraspecific prey choice by Amazonian hunters. Current Anthropology, 36, 789–818. Anderson, J., Rowcliffe, J. M., & Cowlishaw, G. (2007). Does the matrix matter? A forest primate in a complex agricultural landscape. Biological Conservation, 135, 212–222. Arroyo-Rodríguez, V., & Mandujano, S. (2009). Conceptualization and measurement of habitat fragmentation from the primates' perspective. International Journal of Primatology, 30, 497–514. Auricchio, P. (1995). Primatas do Brasil. São Paulo: Terras Brasilis Editora Ltda. Bastos Moreira, L. L. (2009). Primatas das Serras das Lontras e Javi: Estado das populações e seu papel na conservação regional da comunidade de primatas no sul da Bahia. Master's thesis, Universidade Estadual de Santa Cruz, Ilhéus, Brasil. Beltrão-Mendes, R., Cunha, A. A., & Ferrari, S. F. (2011). New localities and perspectives on the sympatry between two endangered primates (Callicebus coimbrai and Cebus xanthosternos) in northeastern Brazil. Mammalia, 75, 103–105. Benchimol, M., & Peres, C. A. (2014). Predicting primate local extinctions within Breal-world^ forest fragments: a pan-neotropical analysis. American Journal of Primatology, 76, 289–302. Blanco, V., & Waltert, M. (2013). Does the tropical agricultural matrix bear potential for primate conservation? A baseline study from Western Uganda. Journal for Nature Conservation, 21, 383–393. Canale, G. R., Peres, C. A., Guidorizzi, C. E., Gatto, C. A. F., & Kierulff, M. C. M. (2012). Pervasive defaunation of forest remnants in a tropical biodiversity hotspot. PLoS ONE, 7, e41671. Cardoso de Sousa, M., dos Santos, S. S., & Coelho Marques Valente, M. (2008). Distribuição e variação na pelagem de Callicebus coimbrai (Primates, Pitheciidae) nos estados de Sergipe e Bahia, Brasil. Neotropical Primates, 15, 54–59. Cassano, C. R., Barlow, J., & Pardini, R. (2012). Large mammals in an agroforestry mosaic in the Brazilian Atlantic forest. Biotropica, 44, 818–825. Chagas, R. R. D., & Ferrari, S. F. (2011). Population parameters of the endangered titi monkey, Callicebus coimbrai Kobayashi and Langguth, 1999, in the fragmented landscape of southern Sergipe, Brazil. Brazilian Journal of Biology, 71, 569–575. Chapman, C. A., Lawes, M. J., & Eeley, H. A. C. (2006). What hope for African primate diversity? African Journal of Ecology, 44, 116–133. Chiarello, A. G. (1999). Effects of fragmentation of the Atlantic forest on mammal communities in southeastern Brazil. Biological Conservation, 89, 71–82. Chiarello, A. G., & de Melo, F. R. (2001). Primate population densities and sizes in Atlantic Forest remnants of northern Espírito Santo, Brazil. International Journal of Primatology, 22, 379–396. Coimbra-Filho, A. F., Rylands, A. B., Pissinatti, A., & Santos, I. B. (1992). The distribution and status of the buff-headed capuchin monkey, Cebus xanthosternos, in the Atlantic Forest region of eastern Brazil. Primate Conservation, 12(13), 24–30. Coimbra-Filho, A. F., Mittermeier, R. A., Rylands, A. B., Mendes, S. L., Kierulff, M. C. M., & Pinto, L. P. S. (2006). The taxonomic status of Wied’s black-tufted-ear marmoset, Callithrix kuhlii (Callitrichidae, Primates). Primate Conservation, 21, 1–24. Corlett, R. T. (2007). The impact of hunting on the mammalian fauna of tropical Asian forests. Biotropica, 39, 292–303. de Freitas, C. H., Setz, E. Z. F., Araújo, A. R. B., & Gobbi, N. (2008). Agricultural crops in the diet of bearded capuchin monkeys, Cebus libidinosus Spix (Primates: Cebidae), in forest fragments in southeast Brazil. Revisita Brasileira de Zoologia, 25, 32–39. Dean, W. (1995). With broadax and firebrand: The destruction of the Brazilian Atlantic Forest. Berkeley: University of California Press.

Distribution and Conservation of Three Atlantic Forest Monkeys Devore, J. D. (2014). Cultivating hope: Struggles for land, equality, and recognition in the cacao lands of southern Bahia, Brazil. Doctoral thesis, University of Michigan. Duarte, M. H. L., & Young, R. J. (2011). Sleeping site selection by urban marmosets (Callithrix penicillata) under conditions of exceptionally high predator density. International Journal of Primatology, 32, 329– 334. Emmons, L. H. (1984). Geographic variation in densities and diversities of non-flying mammals in Amazonia. Biotropica, 16, 210–222. Estrada, A. (2013). Socioeconomic contexts of primate conservation: population, poverty, global economic demands, and sustainable land use. American Journal of Primatology, 75, 30–45. Estrada, A., & Coates-Estrada, R. (1996). Tropical rain forest fragmentation and wild populations of primates at Los Tuxtlas, Mexico. International Journal of Primatology, 17, 759–783. Estrada, A., Raboy, B. E., & Oliveira, L. C. (2012). Agroecosystems and primate conservation in the tropics: a review. American Journal of Primatology, 74, 696–711. Faria Franco Negrão, M., & Valladares-Pádua, C. (2006). Registros de mamíferos de maior porte na Reserva Florestal do Morro Grande, São Paulo. Biota Neotropica, 6, 1–13. Ferrari, S. F., & Beltrão-Mendes, R. (2011). Do snakes represent the principal predatory threat to callitrichids? Fatal attack of a viper (Bothrops leucurus) on a common marmoset (Callithrix jacchus) in the Atlantic Forest of the Brazilian northeast. Primates, 52, 207–209. Flesher, K. M. (1999). Primates of the Ituberá forest complex. Neotropical Primates, 7, 127–131. Flesher, K. M. (2006). Explaining the biogeography of the medium and large mammals in a humandominated landscape in the Atlantic forest of southern Bahia, Brazil: Evidence for the role of agroforestry systems as wildlife habitat. Doctoral thesis, Rutgers University. Flesher, K. M., & Laufer, J. (2013). Protecting wildlife in a heavily hunted biodiversity hotspot: a case study from the Atlantic forest of Bahia, Brazil. Tropical Conservation Science, 6, 181–200. Fonseca, G. A., & Lacher, T. E. (1984). Exudate-feeding by Callithrix jacchus penicillata in semideciduous woodland (cerradão) in central Brazil. Primates, 25, 441–450. Frigeri, E., Cassano, C. R., & Pardini, R. (2014). Domestic dog invasion in an agroforestry mosaic in southern Bahia, Brazil. Tropical Conservation Science, 7, 508–528. Galetti, M., & Pedroni, F. (1994). Seasonal diet of capuchin monkeys in a semi-deciduous forest in south-east Brazil. Journal of Tropical Ecology, 10, 27–39. Galindo-Leal, C., & de Gusmão Camara, I. (2003). The Atlantic Forest of South America: Biodiversity status, threats, and outlook. Washington: Island Press. Heiduck, S. (1997). Food choice in masked titi monkeys (Callicebus personatus melanochir): selectivity or opportunism? International Journal of Primatology, 18, 487–502. Heiduck, S. (2002). The use of disturbed and undisturbed forest by masked titi monkeys Callicebus personatus melanochir is proportional to food availability. Oryx, 36, 133–139. Hilário, R. R., & Ferrari, S. F. (2015). Dense understory and absence of capuchin monkeys (Sapajus xanthosternos) predict higher density of common marmoset (Callithrix jacchus) in the Brazilian northeast. American Journal of Primatology, 77, 425–433. Isabirye-Basuta, G. M., & Lwanga, J. S. (2008). Primate populations and their interactions with changing habitats. International Journal of Primatology, 29, 35–48. IUCN (2014). IUCN Red List of Threatened Species, version 2014.3. www.iucnredlist.org. Accessed 5 Oct 2014. Jerozolimski, A., & Peres, C. A. (2003). Bringing home the biggest bacon: a cross-site analysis of the structure of hunter-kill profiles in Neotropical forests. Biological Conservation, 111, 415–425. Kierulff, M. C. M., Santos, G. R., Canale, G. R., Carvalho, C. E. G., Cassano, C. R., Gouveia, P. S., Cassiano A. F. R. G., Araújo, M., Vila Nova, P., Marques, A. C., Santana Santos, P., & Pádua, J. C. (2005). Avaliação das populações do macaco-prego-do-peito-amarelo (Cebus xanthosternos) e proposta de estratégia para manejo e conservação da espécie. Instituto de Estudos Sócio-Ambientais do Sul da Bahia and Conservation International do Brasil. Kierulff, M. C. M., Mendes, S. L., & Rylands, A. B. (2008). Cebus xanthosternos. The IUCN Red List of Threatened Species, version 2014.3. www.iucnredlist.org. Accessed 5 Oct 2014. Kinzey, W. G., & Becker, M. (1983). Activity patterns of the masked titi monkey, Callicebus personatus. Primates, 24, 337–343. Ludwig, G., Aguiar, L. M., & Rocha, V. J. (2006). Comportamento de obtenção de Manihot esculenta Crantz (Euphorbiaceae), mandioca, por Cebus nigritus (Goldfuss) (Primates, Cebidae) como uma adaptação alimentar em períodos de escassez. Revista Brasileira de Zoologia, 23, 888–890. Michalski, F., & Peres, C. A. (2005). Anthropogenic determinants of primate and carnivore local extinctions in a fragmented forest landscape of southern Amazonia. Biological Conservation, 124, 383–396.

K.M. Flesher Mikich, S. B., & Liebsch, D. (2014). Damage to forest plantations by tufted capuchins (Sapajus nigritus): too many monkeys or not enough fruits? Forest Ecology and Management, 314, 9–16. Mittermeier, R. A., Coimbra-Filho, A. F., Constable, I. D., Rylands, A. B., & Valle, C. (1982). Conservation of primates in the Atlantic forest region of eastern Brazil. International Zoo Yearbook, 22, 2–17. Müller, K. H. (1996). Ranging in masked titi monkeys (Callicebus personatus) in Brazil. Folia Primatologica, 65, 224–228. Müller, K. H., Ahl, C., & Hartmann, G. (1997). Geophagy in masked titi monkeys (Callicebus personatus melanochir) in Brazil. Primates, 38, 69–77. Oliveira, L. C., & Dietz, J. M. (2011). Predation risk and interspecific association of two Brazilian Atlantic forest primates in cabruca agroforest. American Journal of Primatology, 73, 852–860. Oliveira, R. C. R., Coelho, A. S., & Melo, F. R. (2003). Estimativa de densidade e tamanho populacional de sauá (Callicebus nigrifrons) em um fragmento de mata em regeneração, Viçosa, Minas Gerais, Brasil. Neotropical Primates, 11, 91–93. Oliveira, V. B., Linares, A. M., Corrêa, G. L. C., & Chiarello, A. G. (2008). Predation on the black capuchin monkey Cebus nigritus (Primates: Cebidae) by domestic dogs Canis lupus familiaris (Carnivora: Canidae), in the Parque Estadual Serra do Brigadeiro, Minas Gerais, Brazil. Revista Brasileira de Zoologia, 25, 376–378. Peres, C. A. (1990). Effects of hunting on western Amazonian primate communities. Biological Conservation, 54, 47–59. Peres, C. A. (1999). General guidelines for standardizing line transect surveys of tropical forest primates. Neotropical Primates, 7, 11–16. Pozo-Montuy, G., Serio-Silva, J. C., & Bonilla-Sánchez, Y. M. (2011). Influence of the landscape matrix on the abundance of arboreal primates in fragmented landscapes. Primates, 52, 139–147. Price, E. C., Piedade, H. M., & Wormell, D. (2002). Population densities of primates in a Brazilian Atlantic forest. Folia Primatologica, 73, 54–56. Remis, M. J., & Jost Robinson, C. A. (2012). Reductions in primate abundance and diversity in a multiuse protected area: synergistic impacts of hunting and logging in a Congo Basin forest. American Journal of Primatology, 74, 602–612. Rocha Santos, L., & Talora, D. C. (2012). Recovery of Atlantic rainforest areas altered by distinct land-use histories in northeastern Brazil. Tropical Conservation Science, 5, 475–494. Rylands, A. B. (1996). Habitat and the evolution of social and reproductive behavior in Callitrichidae. American Journal of Primatology, 38, 5–18. Rylands, A. B., Williamson, E. A., Hoffmann, M., & Mittermeier, R. A. (2008). Primate surveys and conservation assessments. Oryx, 42, 313–314. Saatchi, S., Agosti, D., Alger, K., Delabie, J., & Musinsky, J. (2001). Examining fragmentation and loss of primary forest in the southern Bahian Atlantic Forest of Brazil with radar imagery. Conservation Biology, 15, 867–875. Santos, I. B., Mittermeier, R. A., Rylands, A. B., & Valle, C. M. C. (1987). The distribution and conservation status of primates in southern Bahia, Brazil. Primate Conservation, 8, 126–142. Siemers, B. M. (2000). Seasonal variation in food resource and forest strata use by brown capuchin monkeys (Cebus apella) in a disturbed forest fragment. Folia Primatologica, 71, 181–184. Silva Ferreira, A. (2014). Hábitat e recursos alimentares utilizados por Callithrix sp. em mosaico agroflorestal de Mata Atlântica na Bahia. Master's thesis, Universidade Estadual de Santa Cruz, BA. SOS Mata Atlântica (2012). Atlas dos Remanescentes Florestais da Mata Atlântica. http://www.sosma.org.br. Accessed 5 Oct 2014. SUDENE (1976). Mapa de Ituberá, Bahia. Folha SD. 24–V-D-VI. Ministério do Interior, Superintendência do Desenvolvimento do Nordeste. Sussman, R. W., & Kinzey, W. G. (1984). The ecological role of the Callitrichidae: a review. American Journal of Physical Anthropology, 64, 419–449. Treveline, L. C., Port-Carvalho, M., Silveira, M., & Morell, E. (2007). Abundance, habitat use and diet of Callicebus nigrifrons Spix (Primates, Pitheciidae) in Cantareira State Park, São Paulo, Brazil. Revista Brasileira de Zoologia, 24, 1071–1077. Veiga, L. M., Printes, R. C., Ferrari, S. F., Kierulff, C. M., de Oliveira, M. M., & Mendes, S. L. (2008). Callicebus melanochir. The IUCN Red List of Threatened Species, version 2 014.3. www.iucnredlist.org. Accessed 5 Oct 2014.