Dynamics of hunting territories and prey distribution in Amazonian Indigenous Lands

Applied Geography 56 (2015) 222e231

Contents lists available at ScienceDirect

Applied Geography journal homepage: www.elsevier.com/locate/apgeog

Dynamics of hunting territories and prey distribution in Amazonian Indigenous Lands P.A.L. Constantino* SHIS QI 29, conjunto 12, casa 1, Brasília, DF, Brazil

a r t i c l e i n f o

a b s t r a c t

Article history: Available online

Indigenous Lands in the Brazilian Amazon intend to guarantee indigenous rights and conserve forests, although many do not correspond to peoples' territorial needs and may not effectively preserve wildlife. Most indigenous people rely on game for subsistence, and the spatial distribution of hunts and prey determine hunting sustainability and wildlife conservation. I examined the Kaxinawa hunting territory dynamics through the participatory monitoring and mapping of 10 ILs. The Kaxinawa are central-place foragers whose ideal hunting territories have a circular shape with a radius of 5 km. The geopolitics of the Kaxinawa combined with spatial occupation distort hunting territories to maintain indigenous control while respecting the territories of nearby villages. The fission of large villages leads to reduced hunting territories but increases the overall hunted area, consequently affecting game populations. Kaxinawa hunting did not lead species to extinction. The Kaxinawa hunted 65% of prey within 2.5 km of the villages and the other 30% within 5 km. Although all of the species were hunted close to villages, the prey were smaller, and several sensitive species were rarely hunted. The replacement of such sensitivity for more resilient low-ranked species on hunting bags suggests that these species might be depleted near villages. These findings provide objective standards for titling Indigenous Lands and for improving wildlife management within these lands. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Kaxinawa Acre Central-place forager Subsistence hunting territory Participatory mapping Participatory monitoring

Introduction Wildlife remains fundamental to most Amazonian indigenous peoples as the main meat source and for structuring their society (Nasi et al., 2008). However, in a feedback loop, the social and cultural organization of an indigenous group may be unsustainable, impacting game populations, while the reduction of game availability affects the indigenous social organization as well as the ecosystem on which these people rely (Milner-Gulland, Bennett, & SCB 2002 Annual Meeting Wild Meat Group, 2003). Therefore, when Indigenous Lands1 are titled in Brazil to ensure the indigenous rights to maintaining the wellbeing, physical and cultural reproduction of the indigenous people, wildlife conservation is one of the mechanisms that accomplish indigenous rights (Brasil, 1988; Stocks, 2005). Particularly in the Brazilian Amazon, ILs are expected to contribute to wildlife conservation as much as other protected

* Tel.: þ55 61 8177 3797. E-mail address: [email protected]. 1 Abbreviation: IL(s) e Indigenous Land(s). http://dx.doi.org/10.1016/j.apgeog.2014.11.015 0143-6228/© 2014 Elsevier Ltd. All rights reserved.

areas given their strategic location and extension (Azevedo-Ramos, Amaral, Nepstad, Soares-Filho, & Nasi, 2006; Nepstad et al., 2006). Many features of huntereprey dynamics, such as hunting effort, strategy, technology, extension and distribution, may have differential effects on game populations and communities that can threaten both wildlife and the indigenous people relying on them. Estimates on wildlife population size and hunters' offtake are used to determine the causes of variation in Amazonian wildlife communities (Alvard, Robinson, Redford, & Kaplan, 1997; Peres, 2001; Robinson & Redford, 1994) and to evaluate the hunting sustainability to set wildlife-management standards (Bodmer & Robinson, 2004). However, the spatial dynamics of hunters and their prey have been overlooked despite their high relevance in Amazonian hunting systems (Milner-Gulland et al., 2003; Novaro, Redford, & Bodmer, 2001; Peres, 2001). The arbitrary choice of catchment area shape in these models, for example, resulted in inadequate conclusions regarding the sustainability of subsistence hunting in the Amazon, where species would theoretically be locally extinct but were still present (Levi, Shepard, Ohl-Schacherer, Peres, & Yu, 2009). The shape of the catchment area of a group of hunters would not differ from that of a hunting territory (Novaro et al., 2001; Peres &

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

Nascimento, 2006); however, the latter concept is associated with hunter's social organization. A hunting territory is the parcel of the territory as recognized by the indigenous social unit, either a single village or the entire ethnic group, as the area of hunting over which they exert some degree of control and ownership (Descola, 1994). These territories are constructed as a consequence of the indigenous relationship to space (Read et al., 2010), which frequently involves characteristics of the landscape and environment, prey availability, indigenous belief systems, their occupation pattern, and space use by other people (Lopez, Beard, & Sierra, 2013). Similar to territories, the boundaries and “ownership” of hunting territories are dynamic and defined by local norms that are not necessarily recognized, regulated or granted by the state (Little, 2001). Therefore, the social dynamics of the indigenous groups influence the shape and boundaries of hunting territories, consequently reflecting the distribution of game species and hunting sustainability. Recent studies have described the spatial distribution and dynamics of hunters, hunts, prey, and hunting territories (OhlSchacherer et al., 2007; Parry, Barlow, & Peres, 2009; Smith, 2008), but few have analyzed the causes of territory variation relating to hunters and their prey (Lopez et al., 2013; Read et al., 2010). The studies on indigenous hunting, however, were conducted within large, well-protected areas (Ohl-Schacherer et al., 2007) or in villages without titled land (Read et al., 2010), which are rare in the Amazon. Therefore, describing hunter behavior, prey distribution, and the processes influencing the dynamics of hunting territories of contemporary traditional societies in delimited protected areas represent necessary knowledge not only for improving wildlife conservation but also for complying with indigenous rights (Levi, Lu, Yu, & Mangel, 2011; Read et al., 2010). This study intends to determine 1) the processes driving the variation in the shape of indigenous hunting territories and 2) how

223

prey distribution might be affected by hunting territory dynamics. I hypothesize that hunting territories vary according to geopolitical processes as composed by village social organizations and constrains that are imposed by other settlements and societies. In addition, I hypothesize that variation in hunting territories indicates in changes in prey catch distribution. I described the patterns of indigenous hunting strategies, hunting territory dynamics, and hunted prey distribution inside the Kaxinawa ILs in the state of Acre, Brazilian Amazon. Material and methods Study area The Kaxinawa, which are auto-denominated Huni Kuin, are a Panoan-language-speaking indigenous people occupying the Brazilian-Peruvian border. After the arrival of non-indigenous people in the region in the late nineteenth century, the Kaxinawa corresponded to 55% of the total indigenous population in Acre State, Brazil, with more than 7200 people (IBGE, 2010) living in 12 ILs that were demarcated from the late 1970s to the early 1990s (Iglesias, 2003). Game meat is a highly valued good that is mostly hunted for subsistence and is relevant to determine social relations within and among villages even when Kaxinawa families have access to alternative meats (Constantino et al., 2008; Kensinger, 1985; Navarro, 2004). Prey can be classified as preferred species, which are mainly represented by large game; species of secondary preference; and low-preferred species, which are only hunted when hunters cannot find any other species and must return home. This research was conducted in 45 Kaxinawa villages of 10 Indigenous Lands that are located in Acre State, western Brazilian
and Amazon, in a region of app. 72,000 km2 between the Alto Jurua Alto Purus Rivers (Fig. 1). The study encompasses 83% of the ILs that

Fig. 1. Kaxinawa Indigenous Lands and villages that are studied in the state of Acre. Indigenous Land numbers correspond to Table 1.

224

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

are occupied by the Kaxinawa, 67% of Kaxinawa villages, and app. 63% of the Kaxinawa population living in ILs in Acre State, Brazilian Amazonia (Table 1). All of the Kaxinawa ILs are titled except for the Kaxinawa do Seringal Curralinho, which had no formally identified boundaries in 2014. A Kaxinawa IL in Brazil can host several villages averaging 80 people. All of the villages are located alongside the main rivers (Fig. 1), whereas only three have immediate access to roads. The predominant vegetation is the terra firme forest (noninundated forest), which is dominated by open canopy with palm trees and natural patches of bamboo in clay soils (Acre, 2006). Precipitation ranges from 1600 to 2000 mm/year (Sombroek, 2001). Acre State has a relatively dense human population living in forests that have been under intensive use for rubber and timber extraction and increasingly deforested for cattle (Hayashi, Souza, Sales, & Verissimo, 2009). Wildlife has been under elevated hunting pressure by indigenous and non-indigenous people for decades compared to elsewhere in the Amazon (Peres & Palacios, 2007); however, most large game or threatened species are hunted in protected areas (Calouro, 2005; Constantino, unpuplished data; Ramos, 2005). Participatory mapping and monitoring of hunting The Kaxinawa collected hunting data as part of a participatory mapping and monitoring program that was led by the local NGO ~o Pro
-
Indio do Acre (CPI-AC) and the Environmental SecComissa retary of Acre State to improve the environmental and territorial management of ILs. The methodology of participatory research was designed and implemented to compose a long-term education program for indigenous environmental agents (Constantino, Carlos, et al., 2012), which are hired as rural extension agents to motivate environmental debates and sustainable initiatives in their villages. Between 2006 and 2009, a 10-to-15-day natural resource management workshop was held twice in seven ILs (Table 1) with an average participation of 30 Kaxinawa hunters from 51 villages. The hunters drew the boundaries of the villages' hunting territories and drew and estimated the distance of associated landscape features (e.g., hunting trails, saltlicks and campsites) on a 1:50,000 A0 size map containing the cartographic features “IL boundaries”, “village location”, and “hydrographic system”. The hunters of a village did not access the mapping of other villages while mapping their hunting territory. The output maps were digitalized, and spatial analyses were performed using ArcMap 9. The size and shape of the ILs were downloaded from the Federal Agency for Indigenous Affairs (FUNAI) website (http://www.funai.gov.br/).

During the workshops, the Kaxinawa filled out to questionnaires that assessed the village age, population size, number of hunters, frequency of hunts, hunting techniques and most-employed tactics. In addition, the Kaxinawa were organized into focal groups per village and per group of villages to debate the results of hunting monitoring and mapping, exploring the reasons and consequences for hunting territories and the status of game populations. Community representatives of 42 villages from ten Kaxinawa ILs, some of which mapped their hunting territories, engaged in a hunting monitoring system between 2004 and 2011, recording the features of all of the hunted animals (species, weight, sex and age), hunting trips (date, type, number of hunters, and duration), and kill events (location, distance from the village or campsite, and gun used) (Constantino, unpublished data). I excluded the first year of data collection, which is considered a learning period, and the last two years, when several villages quit the monitoring program, to ensure data quality and selected a set of 33 Kaxinawa villages from seven Kaxinawa ILs (Table 1). The range of months with monitored hunts varied between seven and 45, averaging 18 monitored months per village. The hunting records during the learning period of 2004 and from 9 villages that had less than seven months of monitoring were not included. To calibrate the indigenous estimate of distance and to check for data mapping and hunting quality, I followed along on hunting trips, recording the indigenous estimation of distance from 9 villages and the geographic position of 27 landscape features that were mapped and 65 animals that were killed using a GPS Garmin CSX60. The straight distance from each of those points to the village center in meters was calculated in ArcMap 9. Given that the Kaxinawa estimate distances used “minutes” as units (e.g., Kensinger, 1975), I used these 92 points to establish the association between the two distance estimations to convert the indigenous unit “minute” of distance into meter. The indigenous distance of 60 min corresponds to a distance of 1.72 km, similar to that found among the Matsigenka in Manu National Park, 60 min for 2 km (OhlSchacherer et al., 2007). Data analysis For the sample of mapped villages (n ¼ 51), I proceeded with the following analyses: 1) pairwise Person's correlation to verify the association between hunting territory area and maximum distance of hunting territory, and 2) linear multiple regression to identify the factors that could explain the variation in size and the maximum distance of one-day hunting territories across villages. I selected the best model that explained the studied system after excluding

Table 1 Scope of research and wildlife management in Kaxinawa Indigenous Lands in Acre State, Brazil. NM e not mapped. Map ref.

Indigenous Land

# Monitored villages

Total # villages

Size (km2)

Hunting territory (km2) One-day

Overnight

1 2 3 4 5 6 7 8 9 10

Kaxinawa/Ashaninka do Rio Breua ~o Kaxinawa do Rio Jorda Kaxinawa do Baixo Rio Jord~ ao ~ Kaxinawa da Praia do Carapana ^ncia Kaxinawa do Seringal Independe Kaxinawa Nova Olinda Kaxinawa do Seringal Curralinho
b Kaxinawa do Rio Humaita
do Caucho Igarape Alto Rio Purusc Total

4 11 6 5 0 0 0 3 1 3 33

5/6* 20 8 9 2 4 3 6 4 21/44* 82/106*

312.78 872.94 87.26 606.99 115.84 275.33 ** 1273.84 123.18 2631.30 6299.46

290.02 534.42 80.27 229.20 83.31 219.95 60.98 NM NM NM 1437.18

80.96 213.79 24.27 93.87 9.23 97.59 No NM NM NM 438.75

*Kaxinawa villages/total villages; **Indigenous Land not titled. a Also occupied by 1 Ashaninka village. b Also occupied by unknown number of isolated people. c Also occupied by 23 Kulina villages.

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

225

unrelated explanatory variables from the full model (p > 0.1) through a backward stepwise elimination and after analyzing the largest adjusted R2. Apart from looking at the patterns of hunting territory shapes, the size of the one-day hunting territory and the maximum distance between the center of the village to the boundary of the oneday hunting territory are metrics of shape design and were used as proxies for the hunting territory shape in the multiple regression analysis. These variables were the dependent variables for hunting territory shape, whereas the explanatory variables that are likely to influence the shape of hunting territories were selected based on scientific literature in anthropology (e.g., Levi et al., 2011), conservation (e.g., Jerozolimski & Peres, 2003), and geography (e.g., Lopez et al., 2013): village population size, village age, number of neighboring villages in a 5-km radius, and shorter village distance to the border of the IL. The size, shape, and spatial distribution of the hunting territories and wildlife-management zones were estimated using ArcMap 9. The villages in the Kaxinawa do Seringal Curralinho IL were not included in the regression analysis because this IL has no demarcated borders, but these villages represent well the situation of non-indigenous restriction to Kaxinawa hunting territories. Based on the size of the ideal Kaxinawa one-day hunting territory (78.5 km2, see Discussion) and on monitoring data, I used the average distance within which prey of preferred species were hunted from the village to create three classes of hunting distances. Distance Class 1 corresponds to the area within an app. 2.5-km radius from the village, Distance Class 2 corresponds to the area between app. 2.5- and 5-km radii from the village, and Distance Class 3 corresponds to the area greater than a 5-km radius from the village. I compared the assemblage of prey that were hunted in the three distance classes 1) using an ANOVA and Tukey's HSD test to compare the mean prey weight, 2) ranking the species according to the frequency of hunted animals in each distance class, and 3) using the BrayeCurtis dissimilarity test to compare the composition of the assemblage of the 10 most-frequently hunted species. Whilst these analyses regard the overall pattern in all of the studied Kaxinawa Indigenous Lands and villages, the local cases were selected for illustrative purposes. Results Kaxinawa hunting strategies All of the hunters residing in a village can forage in the same hunting territory but may request permission to forage in the hunting territories of other villages. The Kaxinawa completely substituted the bow and arrow for shotguns of a variety of calibers, which were the most common weapon that was used to hunt, accounting for 94% of the prey killed. The other animals were hunted using a machete or fire, in the case of armadillos, or were caught by hand. The most common Kaxinawa hunting strategy was the one-day hunting departing from the village of residence (82.7% of hunting events), when hunters spent app. 7 h on average (n ¼ 2653). Usually, a single man (74.3%) left the village at the dawn following a hunting trail searching for fresh signs of large-bodied preferred game species. During the beginning of the hunt, hunters ignore smaller prey of less-preferred species. If the hunter kills a large prey and judges that there is enough meat, he returns home. However, if the time passes and the distance from the village increases without encountering prey of large preferred species, the hunter does not ignore less preferred species any longer. Because hunters have good knowledge of the availability of species in their hunting territory, the moment that they decide to kill less-preferred animals depends on their expectation of killing a preferred one. Usually, the

Fig. 2. Territories of Kaxinawa wildlife management: a) Hunting territories of isolated villages and wildlife reserve area of Kaxinawa/Ashaninka do Rio Breu and Kaxinawa do ~o, b) one-day hunting territories of villages at the Kaxinawa do Baixo Rio Rio Jorda ~o, and c) Hunting territory of villages in the untitled Kaxinawa do Seringal CurJorda ralinho. The hunting territories of other nearby villages were not included to improve information clarity. Maps are in the same 1:127,000 scale.

226

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

Table 2 Variables explaining the variation in the one-day Kaxinawa hunting territory remaining in the minimal regression model after backward stepwise selection. Villages in the non-titled IL Kaxinawa do Seringal Curralinho were not included in the analysis because there is no demarcated border. Response variable

Parameter

Estimate

t Ratio

Prob > t

5.91 1.32

5.35 6.01

<0.0001 <0.0001

0.27 <0.01

3.78 4.69

0.001 <0.0001

281.94 0.06

22.88 3.07

<0.0001 0.004

SE

Size Intercept Number of neighboring villages in a 5-km radius Village age Minimum distance to the IL border Maximum distance Intercept Minimum distance to the IL border

31.61 7.91 1.03 0.004

6450.07 0.18

~o IL exemplify this variation (Fig. 2b). The Kaxinawa do Baixo Jorda average maximum distance between the village and the limits of the one-day hunting territory was 5.78 km (SD ¼ 1.38 km, n ¼ 51), and the average size was 35.93 km2. There was no significant correlation between the size and the maximum distance of the oneday hunting territory (r ¼ 0.07, p > 0.05). One-day hunting territories of close villages frequently overlap in the portions that are most distant from the villages (Fig. 2b) without leading to severe conflicts, although there was occasional animosity, particularly when hunters from one IL entered an adjacent IL. The overnight hunting territories usually begin in a campsite extending over an additional area to the one-day hunting territory (Fig. 2a), beginning at app. 5 km away and ending up to 13 km from the hunters' residence villages. It was not possible, however, to identify clear shape patterns for these territories because there was a wide variety of shapes. Drivers of variation in hunting territory shape

Kaxinawa consider widening the range of target species when returning home. The second-most-common hunting strategy is overnight hunting (9.9%), when a group of kinsmen travel to a campsite far from the village and actively forage on an area around the campsite. These hunts may last from two to seven days. The strategies of night hunting with flashlight, sit-and-wait close to a fruiting tree or saltlick, and dog hunting are rarely used by the Kaxinawa. The use of motored canoes for hunting is relatively rare, as oil is expensive and limited and usually used to travel to cities. Shape of the hunting territories The total Kaxinawa hunting territory size in this study was 1869.82 km2, which was divided between one-day hunting territories and overnight hunting territories. All of the villages had a one-day hunting territory that covered a total of 1437.18 km2; 76.5% of the villages had an overnight territory covering 438.75 km2 (Table 1). There was 96.05 km2 of overlapping area between the hunting territory categories. Excluding the overlapping areas, the size of the total overnight hunting territory equals 29.5% of the total one-day hunting territory. Hunting territories are mostly located within the ILs, but the hunting territory of some villages near the IL limits extended beyond the borders into protected or private lands (Fig. 2a, b). Given that villages' mapping hunting territories had 3786 people, the Kaxinawa hunting had to supply the needs of 2.6 people/km2 within a one-day hunting territory. The one-day hunting territory of some villages approximated a circular shape. Novo Segredo (Kaxinawa do Rio Jord~ ao IL) and Jacobina 1 (Kaxinawa/Ashaninka do Rio Breu IL) exemplify the circular one-day hunting territory (Fig. 2a). These villages had two of the largest one-day hunting territories, with 69.23 km2 and 50.2 km2, respectively. Variations of this pattern included shapes that covered a narrow area near the village, widening with distance at one or both margins of the river or covering wider areas near the village and narrower with distance. Neighboring villages in the

The single significant minimal model that was composed of the number of Kaxinawa neighboring villages, the village age and the distance to the border of the IL (Table 2) explained nearly 60% of the variation across one-day hunting territory sizes (Radj ¼ 0.59, F ¼ 21.2, df ¼ 3, p < 0.01, n ¼ 49). The fewer villages neighbored a Kaxinawa village, the more distant it was to the border of the IL, and the older was the foundation of a village, the larger was the one-day hunting territory. The single minimum significant model composed by the distance to the border of IL explained only 16% of the variation of the one-day hunting territory maximum distance (Radj ¼ 0.16, F ¼ 9.45, df ¼ 1, p < 0.01, n ¼ 49) (Table 2), where the maximum distance of the one-day hunting territory decreased as the villages were distanced from the border of IL. Nevertheless, the villages within 5 km of the IL borders had more than 0.75 km of maximum distance of one-day hunting territory than did the villages that were distant from the IL borders, a marginally statistically significant difference (meanvillages close to border ¼ 6.18 km, SD ¼ 1.46 km, n ¼ 24; meanvillages distant to border ¼ 5.43 km, SD ¼ 1.18 km, n ¼ 24; t ¼ 1.88, df ¼ 44, p ¼ 0.06). The population size in the village was not linearly associated with the variation in one-day hunting territory shape. Distribution of hunted prey The Kaxinawa recorded the distance from the village in which 8150 animals from 54 species of higher taxa were hunted (Table 3). The Kaxinawa hunted 95.9% of prey within app. 5 km, from which more than 65% of all species were within distance class 1, almost 30% of 48 species were in distance class 2, and only approximately 4% of 25 species were in distance class 3. The average prey distance from the village was app. 2.2 km, ranging from 1.15 km to 3.9 km across villages. The minimum distance of the hunted animals was virtually zero, given that 5.6% of the animals were killed close to houses, whereas the maximum distance was app. 13 km from the villages during overnight hunts. The Kaxinawa hunted prey of

Table 3 Overall distribution of hunted animals within the different distance classes from Kaxinawa villages. Distance class

Number of species

Number of prey/%

Weight (kg)/%

Mean prey weight (kg ± SE)a

1 e Up to 2.5 km 2 e Between 2.5 and 5 km 3 e Beyond 5 km Total

54 48 25 54

5536/67.93 2276/27.93 338/4.15 8150

49,692/64.28 24,249/31.37 3363/4.35 77,304

8.98 10.66 9.95 9.51

a b

Different superscripted letters indicate significant differences with p < 0.05. Species preference as in Table 2.

± ± ± ±

0.15A 0.24B 0.61C 0.12

Importance of preferred species (%)b 49.28 67.88 70.41 55.35

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

average size that were significantly different among the distance classes (F ¼ 18.31, DF ¼ 2, p < 0.001, n ¼ 8147). The prey that were hunted in distance class 1 were smaller than those in distance class 2 but did not differ from those in distance class 3 (Table 3). Species of varying preference were killed at significantly different distances from the villages (F ¼ 155.49, p < 0.001, df ¼ 2, n ¼ 8150). On average, the preferred species were caught significantly farther away (app. 2.54 km, SD ¼ 2.04 km, n ¼ 4511) than were species of secondary preference (app. 1.84 km, SD ¼ 1.76 km, n ¼ 2818, p < 0.001) and low-ranked species (app. 1.64 km, SD ¼ 1.63 km, n ¼ 872, p < 0.001), whereas the low-ranked species were hunted significantly closer than were species of secondary preference (p ¼ 0.02). Consequently, the representativeness of

227

preferred species increased with distance classes from the villages. Preferred species represented almost 50% of the prey in distance class 1, 67.9% in distance class 2, and 70.4% in distance class 3 (Table 3). Additionally, the proportion of all of the preferred species except for the piping guan increased in the hunters' prey profile from distance class 1 to distance class 2 (Fig. 3a), particularly the spider and woolly monkeys, lowland tapir, and tortoise, which ranged from 90% up to 160%. However, the proportion of spider and wooly monkeys and the caiman that were hunted in distance class 3 increased in relation to the proportion within 5 km (Fig. 3b). In contrast, some smaller and frequently hunted species decreased their representativeness when distant to the villages, such as

Fig. 3. Difference in the percentage of occurrence of prey species with more than 30 hunted animals between distance classes from Kaxinawa villages: a) between distance class 1 and 2 and b) between distance class 3 and 1 þ 2. Preferred species: black bar; secondary species: gray bar; low-ranked species: white bar.

228

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

armadillos, agoutis, squirrel monkeys, tamarins, paca and parrots (Fig. 3a). The small-bodied (i.e., tegu and oropendola) or largebodied animals that were subject to food taboo (i.e., giant armadillo and porcupine) in particular were not hunted at further distances from villages. The composition of the ten most-hunted prey assemblages among the distance classes varied as indicated by the BrayeCurtis dissimilarity. There was a gradient in which the prey composition in distance class 1 was 42% different from that of distance class 2 and 88% different from that of distance class 3. The composition of distance classes 2 and 3 differed by 74%. While the white-lipped and collared peccaries, red brocket deer, howler monkey, and tortoise were frequently hunted independently of the distance class, other species were substituted. The paca and agouti were frequently hunted within 2.5 km but substituted by the brown capuchin and spider monkey between 2.5 km and 5 km from the villages. Beyond 5 km, the caiman and trumpeter substituted the squirrel and armadillos (Fig. 3). Other important species in terms of meat provision to the Kaxinawa also varied greatly between the distance classes. The curassow, wooly monkey, and lowland tapir ranked low in their relative contribution to the number of prey that were hunted close to the villages, being the 19, 24 and 27 mosthunted species, respectively. Nevertheless, these species ranked amongst the 20 most frequently hunted in distance class 2, at 19, 17, and 18, respectively, and in distance class 3, at 16, 15, and 18, respectively. Discussion Kaxinawa central-place foraging behavior and hunting territory dynamics Contemporary Kaxinawa hunters behave as central-place foragers, covering the core hunting territory around the village of residence, rarely foraging on overnight hunting territories. A hunter normally leaves his village in the morning for a one-day walking hunt searching for preferred, large-prey species that satisfy his selfimposed daily meat quota. If such prey are killed, the hunter returns home. If not, after achieving the imaginary boundaries of his oneday hunting territory, he returns home foraging on smaller prey that, when summed, satisfy his quota. This behavior is similar to the first days after contact with non-indigenous people (Kensinger, 1985), despite the Kaxinawa engagement in the rubber economy that influenced the perception of territory, then linked to the “seringais” (Iglesias, 1992). Several differences, however, are observed between the contemporary Kaxinawa and those that were contacted. Not all adult men currently hunt, and the frequency of hunts per men can be very low depending on the linkage with the surrounding society and market and the reliance on alternative meat. The exclusive use of shotguns has enhanced the preference of the Kaxinawa for large game compared to the past bow-and-arrow hunting (Kensinger, 1975). Although there is also no mention of overnight hunts in the studies of on Kaxinawa hunting that were performed right after contact (Kensinger, 1975, 1983, 1985, 1995), the contemporary Kaxinawa expanded their overall hunting territory by almost 30% by hunting in overnight territory. A similar pattern was observed among the Achuar, who hunt for 2e3 days in areas as far as 12 km from village centers (Descola, 1994), and the Siona-Secoya, who have a one-day hunting area of 6 km around the village and an overnight hunting territory around distant campsites (Vickers, 1985). The geopolitics of the ideal Kaxinawa village (mae kuin) are defined by a limited social space in which households occupy its center, from which an ample hunting territory radiates in which hunters can ideally pursue all types of prey (Kensinger, 1975).

Although the shape of the ideal hunting territory is clear, I could not find references to the dimensions of this hunting territory in ethnographies describing Kaxinawa society. Village dwellers tend to avoid day-to-day contact with other societies, even Kaxinawa from other villages, consequently reducing the overlap between social spaces. In relation to nearby Kaxinawa villages, hunters from a village ideally have total control and collective ownership over a hunting territory (Kensinger, 1995). Hunting in a one-day hunting territory of other villages occurs only under “owners” permission, and conflicts may arise when norms are disrespected (Kensinger, 1975). Therefore, an ideal model of Kaxinawa hunting territory would require villages being distanced from the settlements of any ethnic groups, with an ample one-day hunting territory of a circular shape without overlaps with other settlements or hunting territories. In Acre State, only the most remote Kaxinawa villages achieved a pattern that was similar to this model (Fig. 2a). These villages are distanced more than 10 km from other settlements, so the circumference-like one-day hunting territory has an approximately 5-km radius and 78.5 km2, within which most prey are caught, including preferred species. This pattern is similar to the Matsigenka, in Peru (Ohl-Schacherer et al., 2007), and the Yuquí, in Bolivia (Stearman, 1990), who hunt most prey in a one-day core catchment zone with a 5-km radius from the village. Currently, the one-day hunting territory of these Kaxinawa villages approaching the ideal is supplemented by a smaller, less-frequented overnight hunting territory that certainly requires more effort to scout (Hames, 1980). The maximum distance of a one-day hunting territory is, thus, defined by Kaxinawa' capacity of hunting for one day and not by prey availability or the availability of unhunted land. The ~ IL, for example, do not hunt beyond Kaxinawa of Praia do Carapana the 5 km because they consider having enough prey close to villages (Bant & Pessoa, 2008), albeit being one of the most depleted ILs in the state where the Kaxinawa rely less on preferred game (Constantino, Tavares, et al., 2012). Hunters also argue that it is difficult to carry prey back, increasing the chance of getting lost in the forest (Bant & Pessoa, 2008) despite always taking precautions not to get lost when reaching unfrequented areas beyond the oneday distance to the village (Kensinger, 1975). The same pattern is observed among the Achuar in Ecuador, whose villages center their one-day hunting territories that ideally have a circular shape of a nearly 5-km radius and rarely touch other hunting territories (Descola, 1994). These hunting territory dimensions are not static, as their shapes varied around the ideal model mainly related to geopolitics. I propose four processes that are related to the Kaxinawa geopolitics to explain the distortion of hunting territories from the ideal model. First, hunters tend to respect the social norms related to hunting territories and thee land use of neighbor villages, especially avoiding overlapping areas close to these villages, given the reduction of one-day hunting territory size with the increased number of neighboring villages. Second, overlapping areas are located far from villages because the density of preferred game is low, and the “enforcement” of ownership and control over the territory might be easier closer to villages, thereby reducing hunters' willingness to “invade” such areas, whereas space gets more undefined and day-to-day encounters are less likely to occur as distance increases (Hugh-Jones, 1979). For the Achuar in Ecuador, deviations from the ideal model to adapt to the proximity of other villages occur in the less-frequented and -patrolled distant areas within the 5-km radius hunting territory (Descola, 1994). Third, a village's proximity to the IL border functions as a proxy of proximity to surrounding non-indigenous populations, which influences hunting territory shape in two distinct ways: by reducing one-day hunting territories size and by expanding its maximum distance.

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

Variation in size is likely associated with the same process of avoiding entering areas of neighbors' intensive use. Villages that are close to the borders of Kaxinawa do Baixo Jord~ ao IL (Fig. 2b) and in the untitled Kaxinawa do Seringal Curralinho ILs (Fig. 2c), for example, have reduced hunting territory size to accommodate for
had hunting non-indigenous populations. In Bolivia, the Siriono territories that were shaped differently from circles due to the landuse practices of neighboring colonists (Stearman, 1992). In contrast, where Kaxinawa hunters were not limited by neighbors, their territories extrapolated ILs borders. Jacobina 1, for example, expanded their hunting territory into unoccupied areas in Peru, beyond the national and IL borders, approaching the ideal Kaxinawa one-day hunting territory shape (Fig. 2a). In addition, Novo Segredo is far from the IL border and without nearby villages (Fig. 2a). Forth, I suggest that new hunting territory shapes must adapt to previously existing constraints, including the territories of older villages. Therefore, hunters in older villages have control over larger oneday hunting territories than do those in newer villages, but this difference is not caused by an increase in the distance hunters have to walk to catch their prey. This distance is an alternative explanation to the expected increase in hunting territory size of Neotropical indigenous groups as consequence of hunters' necessity to walk farther to forage on denser game populations, which have been locally depleted by persistent hunting in old villages (Hames & Vickers, 1982). An important aspect of these processes is that the overall Kaxinawa population growth within a village does not linearly influence hunting territories but rather does so as a response to the creation of new villages, as suggested for general Amazonian settlements (Levi et al., 2009). In Ecuador, the Achuar hunting territories adjust their shapes and severely reduce the area (up to 1/20) when the density of indigenous settlements increases although not influenced by environmental aspects (Lopez et al., 2013). In the Guiana rainforest, the Makushi and Wapishana 5.5-km-radius oneday hunting territories varied in shape and size when surrounded by other villages (Read et al., 2010). Among the Kaxinawa, villages are created as result of individual or family emigration due to many sociopolitical and ecological reasons, related or not to internal population growth, which include internal disputes, access to markets, search for better environmental conditions, conflicts with other groups, demarcation of land, and large-scale economic cycles (Iglesias, 2003; Kensinger, 1985). Additionally, I suggest two hypotheses to explain the influence of proximity to IL borders on the variation in the maximum distance of hunting territories. First, Kaxinawa hunting territories are constrained by non-indigenous presence as the villages near the IL borders. Colonists are less flexible in overlapping with Kaxinawa territories where conflicts often occur. Therefore, when attempting to expand the hunting territory, hunters walk further. As observed among the Yuquí in the Bolivian Amazon (Stearman & Redford, 1995) and in Ecuadorian indigenous people (Lopez et al., 2013), Kaxinawa regular patrolling is undertaken during hunts. While patrolling, the hunters forage outside IL borders to search for non-indigenous campsites from where illegal non-indigenous hunters encroach (Iglesias, 2003). Second, game populations are expected to become depleted close to IL borders because colonists, who frequently live in more-populated areas, regularly hunt with dogs and shotgun traps, increasing the pressure on indigenous hunting, which tends to have less impact on wildlife. Thus, hunters close to IL borders would was approximately 1 km farther to forage in more available game communities. These two processes are likely to operate concomitantly. In turn, the lack of influence of neighboring villages on the maximum hunting distance seems to be a consequence of the dispersion of all of the villages lined up alongside the riverbanks, allowing for the long hunts towards the upland forest of the back of ILs.

229

The processes influencing spatial hunting patterns of the Kaxinawa and many other indigenous groups are likely to be restricted to central-place foragers and should not be indiscriminately generalized because other tropical indigenous peoples have different hunting spatial patterns that are influenced by different hunter behaviors, land tenure systems, hunting technologies, and landscape characteristics. The Ashaninka leave their village of residence for months, living in mobile settlements at river beaches where they hunt and fish in areas far away from their villages (Constantino, 2011). Some uncontacted and nomadic groups living along the border of Brazil and Peru, such as the Mascho, seasonally perambulate in groups over a vast hunting area using bow and arrows (Castillo, 2004). The Yanomamo and Ye'kwana in Venezuela follow a rotation system of hunting zones (Hames, 1980). The Waimiri-Atroari in Brazilian Amazonia use trucks to reach further, less-disturbed hunting areas (Souza-Mazurek et al., 2000). In the
use open trials that facilitate flat savanna of Bolivia, the Siriono hunters' movement to hunt in an area of a 10-km radius comprising various land tenure categories (Stearman, 1992). The Makushi and Wapishana villages in the Guiana savanna and forest-savanna ecotone have hunting territories with a maximum mean distance of 12-km (Read et al., 2010). Hunted prey distribution in Kaxinawa hunting territories After eight decades of intensive shotgun hunting in Acrean forests2 (Peres & Palacios, 2007), the high Kaxinawa population density in the hunting territories and intensive hunting practices should have led several sensitive species to extinction (Bennett & Robinson, 2000; Levi et al., 2009). Overall, however, the Kaxinawa still rely mostly on preferred prey that are hunted close to villages more than on other species at any distance. Even some of the most threatened game species in Amazonia were occasionally hunted, such as spider and wooly monkeys, lowland tapir, and yellowfooted tortoise (IUCN, 2012). The Kaxinawa hunting behavior of replacing the high costs of hunting low-dense preferred species for edible high-dense less preferred species prevented huntinginduced extinction (Lu, 2010). Nevertheless, some preferred species population are probably depleted, with lower densities close to villages where there are more human induced disturbances, as indicated by their systematic substitution of less-preferred species in hunter prey profiles (Jerozolimski & Peres, 2003). The woolly and spider monkeys and the caiman were hunted considerably more often outside of oneday hunting territories, probably depleted in an area of app. 9436 km2, and the lowland tapir and tortoise were substantially more hunted in the periphery of one-day hunting territories, probably depleted close to villages. Elsewhere, these species are rarely hunted close to villages because they are very sensitive to human disturbances but are frequently hunted near the edge of catchment areas (Levi et al., 2009; Smith, 2008). Overhunting (Peres, 2001), habitat loss (Michalski & Peres, 2007), fire (Barlow & Peres, 2006), and the daily use of forest and noise (Descola, 1994; Urquiza-Haas, Peres, & Dolman, 2009) reduce and displace prey

2 Hunting with shotguns for meat consumption in the “seringais” of current Kaxinawa ILs, and elsewhere in the state, sustained indigenous and colonists populations since 1880s (Constantino et al., 2008; Iglesias, 2010); nowadays, traditional weapons such as spears and bows and arrows are used only for fishery (Cruz & Ferreira, 2004). Hunting for pelt trade took place all over the state, mainly during the depression of rubber market, but significantly decreased late in the 1970s with the enforcement of new hunting legislation, at least in Kaxinawa ILs areas (Constantino et al., 2008). Illegal commercial hunting for meat remains a problem even in the largest cities in the state but is insignificant in most of the studied Kaxinawa ILs.

230

P.A.L. Constantino / Applied Geography 56 (2015) 222e231

populations. Disentangling the effects of these drivers is particularly challenging (Fa & Brown, 2009) because they often act synergistically (Michalski & Peres, 2007). A synergetic impact is likely to occur in Kaxinawa ILs, where all of these disturbances occur concentrated alongside the rivers near the villages (Constantino, unpublished data). Kaxinawa disturbances, however, must not be blamed alone for the current game population distribution. Current non-indigenous hunting, encroaching either upon ILs or in their surroundings, may induce a similar or even more pronounced spatial pattern of game depletion, especially close to IL borders (Constantino, 2010). Moreover, many villages were created in which colonists settled and hunted before ILs were titled, suggesting decades of non-indigenous hunting pressure close to Kaxinawa villages (Constantino et al., 2008; Iglesias, 2003). Albeit depleted in Kaxinawa ILs, hunters still pursue these species. A similar pattern is projected in the Matsiguenka reserves in Peru, where hunters are distributed in spread settlements, forming a large contiguous hunting territory that should severely deplete the woolly and spider monkeys. However, these species would be continuously hunted at similar levels after decades because of the low hunting pressure due to the gradual shift to shotgun hunting technology (Levi et al., 2009). Because the Kaxinawa completely substituted their traditional weapons for shotguns decades ago, the existence of source populations in lesshunted areas providing individuals with intensive, large, continuous hunting territories is a more plausible explanation for the persistence of sensitive species in Kaxinawa ILs. This hypothesis was evoked to explain the recovery of extinct wildlife and the existence of variation in the game abundance across Kaxinawa villages that are surrounded by different land-use systems in Kaxinawa do Rio Jord~ ao IL (Constantino et al., 2008). In addition, Descola (1994) suggests that the interstitial area between the Achuar hunting territories functions as a temporary refuge and reproductive reserves of hunted fauna. Source-sink dynamics are alleged to exist in many tropical rainforest hunting systems (Novaro et al., 2001; Ohl-Schacherer et al., 2007), despite the enormous challenge to report it empirically (Hansen, 2011). Conclusion The geopolitics of Kaxinawa society have implications on their hunting territories dynamics, which determine where wildlife are hunted consequently define the impact on wildlife. While hunting territories have diverted from the traditional ideal model, some sensitive game species are probably severely depleted close to villages (not exclusively by Kaxinawa hunting) although safe in distant areas. These findings suggest that Kaxinawa ILs are not achieving their objective of guaranteeing indigenous wellbeing through the use of natural resources, perhaps due to inaccurate identification and demarcation of IL limits. Given that several indigenous people in the Amazon rely on wildlife, behave as Kaxinawa hunters, and have limited space to hunt, the dynamics of their hunting territories and the consequences on the hunted wildlife might be influenced by the same processes that were identified in Kaxinawa ILs. Therefore, the demarcation of ILs for those people and the management of wildlife within these lands should consider indigenous distribution as well as their hunting territory dynamics. Acknowledgments This research was funded by the CLP, the Gordon and Betty Moore Foundation, and the Tropical Conservation and Development Program at the University of Florida. I would like to thank the

~o Pro
-Indio do Acre and the Kaxinawa people and the Comissa Secretaria de Estado de Meio Ambiente do Acre.

Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.apgeog.2014.11.015.

References
gio-Econo ^mico do Estado do Acre e Fase II. Secretaria Acre. (2006). Zoneamento Ecolo de Estado de Meio Ambiente do Acre. Governo do Acre. Alvard, M. S., Robinson, J. G., Redford, K. H., & Kaplan, H. (1997). The sustainability of subsistence hunting in the Neotropics. Conservation Biology, 11(4), 977e982. Azevedo-Ramos, C., Amaral, B. D., Nepstad, D. C., Soares-Filho, B., & Nasi, R. (2006). Integrating ecosystem management, protected areas, and mammal conservation in the Brazilian Amazon. Ecology and Society, 11.
gico das Terras Indígenas do Bant, A., & Pessoa, M. (2008). Levantamento etnoecolo
: Kaxinawa da Praia do Carapana ~, Kampa do Igarap
complexo Bacia do Rio Jurua e Primavera e Kulina do Igarap
e do Pau. Brasília: FUNAI/PPTAL/GTZ. Barlow, J., & Peres, C. A. (2006). Effects of single and recurrent wildfires on fruit production and large vertebrate abundance in a central Amazonian forest. Biodiversity and Conservation, 15, 985e1012. Bennett, E. L., & Robinson, J. G. (2000). Hunting for sustainability: the start of a synthesis. In J. G. Robinson, & E. L. Bennett (Eds.), Hunting for sustainability in tropical forest (pp. 499e519). New York: Columbia University Press. Bodmer, R. E., & Robinson, J. G. (2004). Evaluating the sustainability of hunting in the Neotropics. In K. Silvius, R. E. Bodmer, & J. M. V. Fragoso (Eds.), People in nature: Wildlife conservation in South and Central America (pp. 299e323). New York: Columbia. ~o da República Federativa do Brasil. Brasil. Brasil. (1988). Constituiça
lise do manejo florestal de “baixo impacto” e da caça de Calouro, A. M. (2005). Ana subsist^ encia sobre uma comunidade de primatas na Floresta Estadual de Antimary ~o Carlos. (Acre, Brasil) (Ph.D. thesis). S~ ao Carlos, Brasil: Universidade de Sa Castillo, B. H. (2004). Indigenous peoples in isolation in the Peruvian Amazon: Their struggle for survival and freedom. Copenhagen: International Work Group for Indigenous Affairs. Constantino, P. A. L. (2010). Disentangling deforestation and indigenous hunting effects on wildlife in Brazilian Amazonia. Gainesville, USA: University of Florida (M.Sc. thesis).
rio Asheninka do rio Envira. Constantino, P. A. L. (2011). Etnozoneamento do Territo
rio da Oficina I. Secretaria de Estado de Meio Ambiente do Acre. Governo Relato do Acre. Constantino, P. A. L. (2014). Deforestation and hunting effects on wildlife variation across Amazonian Indigenous Lands. Biological Conservation (unpublished data). Constantino, P. A. L., Carlos, H. S. A., Ramalho, E. E., Rostant, L., Marinelli, C. E., Teles, D., et al. (2012). Empowering local people through community-based resource monitoring: a comparison between Brazil and Namibia. Ecology and Society, 17(4). Constantino, P. A. L., Fortini, L. B., Kaxinawa, F. R. S., Kaxinawa, A. M., Kaxinawa, E. S., Kaxinawa, A. P., et al. (2008). Indigenous collaborative research for wildlife management in Amazonia: the case of the Kaxinaw
a, Acre, Brazil. Biological Conservation, 141(11), 2718e2729.
rio, F. M., Kaxinawa, E., & Constantino, P. A. L., Tavares, R. A., Kaxinawa, J. L., Maca Kaxinawa, A. S. (2012). Mapeamento e monitoramento participativo da caça na
da Praia do Carapana ~ Indigenous Land, Acre, Amazo ^nia Brasileira. In Kaxinawa ~es geoA. Paese, A. Uezu, M. L. Lorini, & A. Cunha (Eds.), Sistema de informaço
ficas e a conservaça ~o da biodiversidade. Sa ~o Paulo, Brasil: Oficina do Texto. gra
, ShaCruz, T. A., & Ferreira, P. R. N. (2004). Retrato cultural dos Katukina, Kaxinawa ~o de Cultura e Comunicaç~ nenawa, Jaminawa e Manchineri. Fundaça ao Elias Mansour. Governo do Acre. 131 p. Descola, P. (1994). In the society of nature: A native ecology in Amazonia. In Cambridge studies in social and cultural anthropology. Cambridge: Cambridge University Press. Fa, J. E., & Brown, D. (2009). Impacts of hunting on mammals in African tropical moist forests: a review and synthesis. Mammal Review, 39, 231e264. Hames, R. B. (1980). Game depletion and hunting zone rotation among the Ye'Kwana and Yanomamo of Amazonas, Venezuela. In R. B. Hames, & K. M. Kensinger (Eds.), Working papers on South American Indians: Studies in hunting and fishing in the Neotropics (pp. 31e66). Bennington: Bennington College. Number 2. Hames, R. B., & Vickers, W. T. (1982). Optimal diet breadth theory as a model to explain variability in Amazonian hunting. American Ethnologist, 9(2), 358e378. Hansen, A. (2011). Contribution of source-sink theory to protected area science. In J. Liu, V. Hull, & J. A. Wiens (Eds.), Sources, sinks, and sustainability (pp. 339e360). Cambridge: Cambridge University Press. Hayashi, S., Souza, C., Jr., Sales, M., & Verissimo, A. (2009). Transpar^ encia florestal da ^nia Legal (Dezembro 2009 e Janeiro 2010). Bele
m: IMAZON. Amazo Hugh-Jones, C. (1979). From the milk river: Spatial and temporal processes in northwest Amazonia. Cambridge: Cambridge University Press.

P.A.L. Constantino / Applied Geography 56 (2015) 222e231 IBGE Instituto Brasileiro de Geografia e Estatística. (2010). Censo 2010: Unidades da ~o e Terras Indígenas. Downloaded from http://ibge.gov.br/home/ Federaça estatistica/populacao/censo2010/default_resultados_universo.shtm. in 08.04.13. ~o” da Terra Indígena Kaxinawa do Rio Iglesias, M. P. (1992). A incessante “produça Jord~ ao: reconhecimento e regularizaç~ ao pelo Estado Brasileiro e o processo de mobilizaç~ ao e organizaç~ ao do grupo indígiena para a conquista e o ger
rio. In GT Polítia Indigenista (p. 36). Caxambu: Orgaenciamento de seu territo ~o Anual da ANPOCS. nized by XVI Reunia
rio: estrate
gias Iglesias, M. P. (2003). Ocupando e comprando para construir o territo
para o reconhecimento e a regulariaça ~o de novas terras indígenas no Kaxinawa Município do Jord~ ao, Estado do Acre. In GT Povos Indígenas (p. 50). Caxambu: Organized by XVI Reuni~ ao Anual da ANPOCS.
de Felizardo: correrias trabalho e civilizaça ~o no Alto Iglesias, M. P. (2010). Os Kaxinawa
. Brasília: Paralelo 15. 432p. Jurua IUCN. (2012). IUCN red list of threatened species. Version 2012.1 www.iucnredlist. org Accessed on 23.07.12. 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, 415e425. Kensinger, K. M. (1975). Studying the Cashinahua. In J. P. Dwyer (Ed.), The Cashinahua of Eastern Peru (Vol. 1, pp. 9e86). Providence: The Haffenreffer Museum of Anthropology, Brown University. Kensinger, K. M. (1983). On meat and hunting. Current Anthropology, 24, 128e129. Kensinger, K. M. (1985). Hunting and male domination in Cashinahua society. In S. Kent (Ed.), New directions in archaeologyFarmer as hunters: The implication of sedentism (pp. 18e26). Cambridge: Cambridge University Press. Kensinger, K. M. (1995). How real people ought to live: The Cashinahua of Eastern Peru. Long Grove: Waveland Press, Inc. Levi, T., Shepard, G. H., Jr., Ohl-Schacherer, J., Peres, C. A., & Yu, D. (2009). Modelling the long-term sustainability of indigenous hunting in the Manu National Park, Peru: landscape-scale management implications for Amazonia. Journal of Applied Ecology, 46, 804e814. Levi, T., Lu, F., Yu, D. W., & Mangel, M. (2011). The behavioral and diet breadth of central-place foragers: an application to human hunters and Neotropical game management. Evolutionary Ecology Research, 13, 171e185. Little, P. E. (2001). Amazonia: Territorial struggles on perennial frontiers. Baltimore: The Johns Hopkins University Press. Lopez, S., Beard, R., & Sierra, R. (2013). Landscape change in western Amazonia. Geographical Review, 103(1), 37e58. Lu, F. (2010). Patterns of indigenous resilience in the Amazon: a case study of Huaorani hunting in Ecuador. Journal of Ecological Anthropology, 14(1), 5e21. Michalski, F., & Peres, C. A. (2007). Disturbance-mediated mammal persistence and abundanceearea relationships in Amazonian forest fragments. Conservation Biology, 21, 1626e1640. Milner-Gulland, E. J., Bennett, E. L., &, SCB 2002 Annual Meeting Wild Meat Group. (2003). Wild meat: the bigger picture. Trends in Ecology and Evolution, 18, 351e357. Nasi, R., Brown, D., Wilkie, D., Bennett, E., Tutin, C., van Tol, G., et al. (2008). Conservation and use of wildlife-based resources: The bushmeat crisis. Montreal: Secretariat of the Convention on Biological Diversity and Bogor: Center for International Forestry Research (CIFOR). Navarro, J. G. G. (2004). Aprovechamiento de la fauna silvestre en comunidades cashinaua del río Curanja y Purus. Informe T
ecnico I. Lima: World Wildlife Foundation.

231

Nepstad, D., Schwartzman, S., Bamberger, B., Santilli, M., Ray, D., Schlesinger, P., et al. (2006). Inhibition of Amazon deforestation and fire by parks and Indigenous Lands. Conservation Biology, 20, 65e73. Novaro, A. J., Redford, K. H., & Bodmer, R. E. (2001). Effects of hunting in source-sink systems in the Neotropics. Conservation Biology, 14(3), 713e721. Ohl-Schacherer, J., Shepard, G. H., Jr., Kaplan, H., Peres, C. A., Levi, T., & Yu, D. (2007). The sustainability of subsistence hunting by Matsigenka native communities in Manu National Park, Peru. Conservation Biology, 21, 1174e1185. Parry, L., Barlow, J., & Peres, C. A. (2009). Allocation of hunting effort by Amazonian smallholders: implications for conserving wildlife in mixed-use landscapes. Biological Conservation, 142, 1777e1786. Peres, C. A. (2001). Synergetic effects of subsistence hunting and habitat fragmentation on Amazonian forest vertebrates. Conservation Biology, 15(6), 1490e1505.
of Peres, C. A., & Nascimento, H. (2006). Impact of game hunting by the Kayapo southeastern Amazonia: implications for wildlife conservation in tropical forest indigenous reserves. Biodiversity and Conservation, 15, 2627e2653. Peres, C. A., & Palacios, E. (2007). Basin-wide effects of game harvest on vertebrate population densities in Amazonian forests: implications for animal-mediated seed dispersal. Biotropica, 39, 304e315. Ramos, R. M. (2005). Estrat
egia de caça e uso da fauna na Reserva Extrativista do Alto
, AC (M.Sc. thesis). S~ ~o Paulo. Jurua ao Paulo, Brasil: Universidade de Sa Read, J. M., Fragoso, J. M. V., Silvius, K. M., Luzar, J., Overman, H., Cummings, A., et al. (2010). Space, place, and hunting patterns among indigenous people of the Guyanese Rupununi region. Journal of Latin American Geography, 9(3), 213e243. Robinson, J. G., & Redford, K. H. (1994). Measuring the sustainability of hunting in tropical forests. Oryx, 28(4), 249e256. Smith, D. A. (2008). The spatial pattern of indigenous wildlife use in western Panama: implications for conservation management. Biological Conservation, 141, 925e937. Sombroek, W. (2001). Spatial and temporal patterns of Amazon rainfall: consequences for the planning of agricultural occupation and the protection of primary forests. Ambio, 7, 388e396. ^ncio, S., & Souza-Mazurek, R. R., Pedrinho, T., Feliciano, X., Hil
ario, W., Gero Marcelo, E. (2000). Subsistence hunting among the Waimiri Atroari Indians in Central Amazonia, Brazil. Biodiversity and Conservation, 9, 576e596. Stearman, A. M. (1990). The effects of settler incursion on fish and game resources of the Yuquí, a native Amazonian Society of Eastern Bolivia. Human Organization, 49(4), 373e385. Stearman, A. M. (1992). Neotropical indigenous hunters and their neighbors:
, Chimane, and Yuquí hunting on the Bolivian frontier. In K. H. Redford, & Siriono C. Padoch (Eds.), Conservation of tropical forests: Working from traditional resource use (pp. 108e128). New York: Columbia University Press. Stearman, A. M., & Redford, K. H. (1995). Game management and cultural survival: the Yuqui ethnodevelopment project in lowland Bolivia. Oryx, 29(1), 29e34. Stocks, A. (2005). Too much for too few: problems of indigenous land rights in Latin America. Annual Review of Anthropology, 34, 85e104. Urquiza-Haas, T., Peres, C. A., & Dolman, P. M. (2009). Regional scale effects of human density and forest disturbance on large-bodied vertebrates throughout the
n Peninsula, Mexico. Biological Conservation, 142, 134e148. Yucata Vickers, W. T. (1985). An analysis of Amazonian hunting yields as a function of settlement age. In R. B. Hames, & K. M. Kensinger (Eds.), Working papers on South American Indians: Studies in hunting and fishing in the Neotropics (pp. 7e30). Bennington: Bennington College. Number 2.