Ohl Etal 2007 MatsigenkaHunting ConBio w COVER

The Sustainability of Subsistence Hunting byMatsigenka Native Communities in Manu NationalPark, PeruJULIA OHL-SCHACHERER, GLENN H. SHEPARD JR., HILLARD KAPLAN, CARLOS A. PERES,

TAAL LEVI, AND DOUGLAS W. YUCentre for Ecology, Evolution, and Conservation (CEEC), Schools of Environmental and Biological Sciences, University of EastAnglia, Norwich, Norfolk NR47TJ, United KingdomHuman Evolutionary Ecology, Department of Anthropology, University of New Mexico, Albuquerque, NM 87131, U.S.A.

Abstract: The presence of indigenous people in tropical parks has fueled a debate over whether people inparks are conservation allies or direct threats to biodiversity. A well-known example is the Matsigenka (orMachiguenga) population residing in Manu National Park in Peruvian Amazonia. Because the exploitationof wild meat (or bushmeat), especially large vertebrates, represents the most significant internal threat tobiodiversity in Manu, we analyzed 1 year of participatory monitoring of game offtake in two Matsigenkanative communities within Manu Park (102,397 consumer days and 2,089 prey items). We used the Robinsonand Redford (1991) index to identify five prey species hunted at or above maximum sustainable yield withinthe150-km2 core hunting zones of the two communities: woolly monkey (Lagothrix lagotricha), spider monkey(Ateles chamek), white-lipped peccary (Tayassu pecari), Razor-billed Currasow (Mitu tuberosa), and Spixs Guan(Penelope jacquacu). There was little or no evidence that any of these five species has become depleted, otherthan locally, despite a near doubling of the human population since 1988. Hunterprey profiles have notchanged since 1988, and there has been little change in per capita consumption rates or mean prey weights.The current offtake by the Matsigenka appears to be sustainable, apparently due to sourcesink dynamics.Sourcesink dynamics imply that even with continued human population growth within a settlement, offtakefor each hunted species will eventually reach an asymptote. Thus, stabilizing the Matsigenka population aroundexisting settlements should be a primary policy goal for Manu Park.

Keywords: biodiversity conservation, bushmeat, community-based conservation, human-inhabited protectedareas, indigenous rights, Manu National Park, Peru, protected-area management, sourcesink dynamics, subsistencehunting, wild meat

La Sustentabilidad de la Cacera de Subsistencia de Comunidades Nativas Matsigenka en el Parque Nacional Manu,Peru

Resumen: La presencia de indgenas en parques tropicales ha generado un debate sobre s la gente en losparques son aliados para la conservacion o s son amenazas directas para la biodiversidad. Un ejemplobien conocido es la poblacion Matsigenka (o Machiguenga) que reside en el Parque Nacional Manu en laAmazona peruana. Debido a que la explotacion de carne de vida silvestre, especialmente de vertebradosmayores, representa la amenaza interna mas significativa para la biodiversidad en Manu, analizamos 1ano de monitoreo participativo de la captura de presas en dos comunidades Matsigenka dentro del ParqueManu (102,397 das consumidor y 2,089 presas). Utilizamos el ndice de Robinson y Redford (1991) paraidentificar cinco especies de presas cazadas en o por arriba de la produccion maxima sostenible dentro de laszonas de caza de 150-km2 de las dos comunidades: Lagothrix lagotricha, Ateles chamek, Tayassu pecari, Mitu

email [email protected] submitted September 18, 2006; revised manuscript accepted April 23, 2007.

1174

Conservation Biology Volume 21, No. 5, 11741185C2007 Society for Conservation BiologyDOI: 10.1111/j.1523-1739.2007.00759.x

Ohl-Schacherer et al. PeoplePark Conflicts in Peru 1175

tuberosa y Penelope jacquacu. A pesar de que la poblacion humana casi se ha duplicado desde 1988, hubopoca o ninguna evidencia de que alguna de estas especies ha sido diezmada. Los perfiles cazador-presa nohan cambiado desde 1988, y ha habido poco cambio en las tasas de consumo per capita o el peso promedio delas presas. La captura actual por los Matsigenka parece ser sustentable, debido aparentemente a la dinamicafuente-vertedero. La dinamica fuente-vertedero implica que aun con el crecimiento de la poblacion humana, lacaptura de cada especie eventualmente alcanzara una asntota. Por lo tanto, la estabilizacion de la poblacionMatsigenka alrededor de los asentamientos existentes debera ser una poltica primaria en el Parque Manu.

Palabras Clave: areas protegidas, cacera de subsistencia, carne de vida silvestre, conservacion basada en co-munidades, conservacion de la biodiversidad, derechos indgenas, dinamica fuente-vertedero, manejo de areasprotegidas, Parque Nacional Manu

Introduction

The presence of indigenous peoples in parks in the Ama-zon basin has fueled a debate between those who viewindigenous people as conservation allies and those whosee them as a threat (Redford 1991; Alcorn 1993; Redford& Stearman 1993; Robinson 1993; Peres 1994; Harmon1998; Zimmerman et al. 2001; Shepard 2002; Terborgh &Peres 2002; da Silva et al. 2005; Nepstad et al. 2006). Oneskirmish appeared in the pages of Conservation Biology,triggered by Terborghs (1999) warning that the western-izing and fast-growing Matsigenka indigenous populationin Perus Manu National Park will eventually degrade theparks biological integrity unless someway is found to pro-mote voluntary resettlement outside the park (Redford& Sanderson 2000; Schwartzman et al. 2000; Terborgh2000; Peres & Zimmerman 2001).

For Manu the main biodiversity cost of human occupa-tion is the reduction in large-bodied vertebrate game pop-ulations caused by overhunting (Terborgh 1999; Shepardet al. 2007). Matsigenka agricultural practices by them-selveswill cause little disturbance to the park. Even allow-ing for a 50-year fallow period, suitable soils within 500m of the main settlements can sustain agriculture indefi-nitely for a population of at least 2100 people, five timesthe current population (details inOhl et al. 2007). Thus, inthis park, the reconciliation of biodiversity conservationwith indigenous rights starts with effective game man-agement. To this end we implemented a participatoryprotocol for monitoring game animal consumption.

Here we present the results from our first year of datacollection. We test the hypothesis that game populationsare being sustained, despite high hunting pressure, by im-migration from unhunted refugia via sourcesink dynam-ics. Spatial prey refuges stabilize predatorprey dynamics(e.g., May 1978; Joshi & Gadgil 1991) and are widely cred-ited with allowing the persistence of game species withinindigenous reserves across the Amazon (e.g., Begazo &Bodmer 1998; Novaro et al. 2000; Peres 2001).

Furthermore, sourcesink dynamics imply a manage-ment tool. For each locally unsustainably exploited gamespecies, the long-term offtake rate should not exceed therate of immigration from the source (e.g., Siren et al.

2004). This implies that itmight be possible to cap the bio-diversity cost of hunting by stabilizing occupation aroundexisting settlements, even as human population growthoccurs within. Nevertheless, such a conservation strat-egy is only viable to the extent that the immigration rateof game is limited. A high rate of dispersal from sourceto sink can cost the source population some of its vi-ability (source risk, Amarasekare 2004) and can evenresult in the extinction of a source population, especiallyif individuals preferentially disperse into empty territory(Gundersen et al. 2001) and/or if quality of the sourcearea is poor, although sufficient to sustain a population(Amarasekare 2004).

Therefore, we also tested for limited immigration ratesby examining whether hunters travel farther to huntspecies that are less resilient to hunting. Lower resilienceshould result in greater local depletion, forcing huntersto travel farther, on average, to make a kill. Nevertheless,high rates of immigration into sinks would tend to erasesuch a distance effect because high rates smooth out dif-ferences in density between source and sink populations(e.g., Siren et al. 2004).

Methods

Study Area

The 1.7-Mha Manu National Park (PNM) covers the wa-tershed of the Manu River, including large stretches oflowland tropical rainforest. Most rainfall (approximately2600 mm) occurs from November to May. Since its cre-ation in 1973, Manu Park has been considered one ofthe worlds most important tropical protected areas (Ter-borgh 1999). It constitutes the core area of a United Na-tions Educational, Scientific and Cultural Organization(UNESCO) biosphere reserve, is located in a biodiver-sity hotspot, and is a World Heritage Site. As of January2005, there were 421Matsigenka people settledmostly intwo state-recognized communities inside the parks core.Three to four hundred more isolated Matsigenka residein remote settlements in the Manu headwaters, and there

Conservation BiologyVolume 21, No. 5, 2007

1176 PeoplePark Conflicts in Peru Ohl-Schacherer et al.

Figure 1. Map of Manu National Park.

are unknown numbers of uncontacted hunter-gatherers(Shepard et al. 2007).

The settled Matsigenka population within the park isdivided among two legally constituted native communi-ties, Tayakomem ( January 2005 population 149, 25 malesfrom ages 16 to 45) and Yomybatom (183, 36). Two satel-lite communities, Maizals (46, 6) and Sarigeminis (35, 6),and a single-family residence, Maronaro (8, 2), approxi-mately double the 1988 population (Fig. 1; see Supple-mentary Material). (Subscripts differentiate the main [m]and satellite [s] communities.) We reviewed the historyof these communities in Shepard et al. (2007), but inshort, Tayakomem was founded by missionaries (expelledin 1973when theparkwas established) in the early 1960s;Yomybatom dates from the late 1970s; Sarigeminis andMaizals from the 1990s, and Maronaro from 2000. OnlyYomybatom and Tayakomem are provisioned with healthposts and schools. Thus, satellite settlements, which areyounger and have fewer hunters, exert less hunting pres-sure than do the main settlements.

TheMatsigenka live in scattered residence groups of nu-clear family households that are bound by kinship or mar-riage and that often share meals. The Matsigenka engagein hunting, fishing, foraging, and swidden agriculture of

manioc, bananas, and diverse minor crops. Hunting andfishing providemost dietary protein. Otherwild foods areonly a small fraction of their diet. Primate hunting hap-pens mostly from February to June, when primates arefat from eating rainy-season fruits. In the dry season theMatsigenka fish extensively with barbasco (Lonchocar-pus sp.) poison. In Tayakomem, many families fish yearround on the river.

These Matsigenka live in an exceptional situation. Parkrules prohibit firearms and commercial activities, whichforces the Matsigenka to maintain a largely traditional pat-tern of bow-and-arrow hunting and subsistence agricul-ture that is disappearing outside PNM. In Tayakomem andYomybatom, some families maintain a second residenceat some distance from the main community, where gameis more abundant and where they can enjoy greater au-tonomy and avoid social conflicts.

Data Collection

GAME OFFTAKE MONITORING

In October 2004we beganmonitoring offtake of game by26 residence groups, including all groups in Yomybatom



(n = 12), Sarigeminis (2), Maizals (2), and Maronaros (1)and 9 of 11 groups in Tayakomem.We analyzed our first 12months of data, a total of 102,397 consumer days, whereconsumers were 3 years old. A successful monitoringsystem in an indigenous population requires attention tocultural nuances, so a detailed protocol is available (seeSupplementary Material).

Each residence group received pictorial monitoringsheets and scales with which to weigh animals (see Sup-plementary Material). The species, weight, and sex ofitems killed were recorded as were hunting techniquesand weapons used (e.g., bow and arrow, dogs), hunt du-ration (many hunters own wristwatches) and date, loca-tion of the kill, and names of hunter and companions.Skulls were saved whenever possible, which caused usto reject the data from one residence group because theofftake records they reported were more than double thenumber of skulls collected, a degree of mismatch foundnowhere else. Other Matsigenka monitors warned us thatthis family was cheating, perhaps hoping for more pay-ment. Thus, the total number of monitored householdswas 25.

In Yomybato and Tayakome the investigators walkedhunting trails with GPS units (Garmin 12XL and 60, withexternal antennas (Garmin International, Kansas City,Missouri) to georeference landscape features, crossingstreams, salt licks, and secondary residences. We cate-gorized kills into one of four distance categories basedon the locations of the georeferenced features, time andlocation information on the data sheets, and informationfrom detailed interviews conducted during regular visitsto collect the sheets (see Supplementary Material). Theinnermost distance category was a polygon around allthe houses of each community plus a 500-m buffer. Thehunters defined this area as close to the house/field,with a one-way walking time of30minutes. The seconddistance category extended to a radius of 3 km (one-waywalking time of1.5 hours). The third category extendedto 5.5 km (2.5 hours walking time), and the fourth cov-ered all forays >5.5 km beyond the core hunting zone,including stays at secondary residences. The core hunt-ing zones (categories 1 through 3) covered 151 km2 forYomybato and 152 km2 for Tayakome. Interpretable loca-tion information was available for 94% of kills in Yomy-bato and 90% of kills in Tayakome. Incomplete mappingin the other settlements did not permit accurate distancecategorization.

HUNTER-FORAY MONITORING

We used a second pictorial monitoring sheet to obtainadditional information (see Supplementary Material): theduration and location of each hunting foray and the timesat which different game animals were seen, pursued, orhunted. For each animal seen, hunters registeredwhetherthey shot arrows, their arrows hit the target, and the an-

imal was both killed and retrieved. For wounded but un-retrieved animals, hunters were asked to judge by thewounds whether that animal will survive or will die,to be eaten by vultures. In all, 619 forays by 56 differenthunters from eight residence groups were recorded fromNovember 2004 to December 2005.

Data Analyses

We used the following strategy to test whether sourcesink dynamics were sustaining offtake. First, we used theRobinson and Redford (1991) production model to iden-tify a set of species that we could be confident was beingexploited at more than the maximum sustainable yield(MSY) of the core hunting zone. Second, we tested thosespecies for depletion, comparing Yomybatoms currentofftake with historical data sets and comparing all cur-rent offtake from the main settlements with the smaller,more recent satellite settlements. If we could not detectdepletion, we inferred that immigration was maintaininglocal game populations.

SUSTAINABILITY

Wecalculated theMSYbymultiplying the population den-sity at maximum production (K= 0.6, the carrying capac-ity) by the net intrinsic rate of population increase (max= exp[rmax]). Of this production, Pmax, a fraction, F, wastaken as safe to exploit, where F was 0.2, 0.4, and 0.6 forspecies with long, medium, and short life spans, respec-tively:

MSY = Pmax F = 0.6K (max 1)F ; (1)K and max are notoriously difficult to estimate. We fol-lowed Robinson and Redford (1991) and used their val-ues for K, most of which are based on estimates fromPNM, updating parameters for which new informationis available. We also used their method, Coles equation,to estimate rmax. The calculated MSY values are overesti-mates (Milner-Gulland & Akcakaya 2001). Thus, speciesfor which offtake is greater than or equal to MSY surelyindicates locally unsustainable exploitation.

We calculated MSY for all species for which rmax couldbe estimated. Tayakomem offtake was multiplied by 1.4to correct for the unmonitored consumer days of non-participating households. We used results from the foraymonitoring to augment offtake with two measures of col-lateral mortality, when available: retrieved+will die only,and the more conservative retrieved+will die+will sur-vive.

DEPLETION

Optimal foraging theory predicts a more diverse offtakeprofile when preferred game species are depleted, be-cause hunters are forced to accept less-preferred prey



(Bodmer et al. 1997; Fa et al. 2000; Hill et al. 2003; Jero-zolimski & Peres 2003; Rowcliffe et al. 2003). Thus, todetect depletion, we compared the species diversity ofcurrent offtake in the main communities with historicaldata sets and with the satellite settlements.

There are three historical data sets for Yomybatom: (1)October 1, 1988,May 15, 1989 (the largest) (Alvard et al.1997), on five residence groups monitored for 6481 con-sumer days, (2) January 19992000 on three households(da Silva et al. 2005), and (3) October 2001May 2002 onfive households for 1580 consumer days (Ohl 2004). Thefirst and third data sets were based on regular visits, andrecently captured prey items were weighed. The seconddata set was based on skulls saved. To match seasonal-ity, our current data from Yomybatom were limited to 25October 200431 May 2005.

An important limitation of the historical data is thatsampling effort was not as exhaustive and free of prey sizebias as in our current study, in which we recorded offtakedown to small birds for all hunters. Thus, we used onlyprimate offtake, which included both favored and unfa-vored species andwhich for cultural reasons (da Silva et al.2005) was better represented in skull collections than un-gulates. For the other species, and for Tayakomem, wherewe had no historical data, we conducted a spatial compar-ison between the larger communities and their satellitesettlements, which were smaller and younger (see StudyArea) and should have had less-depleted game popula-tions. We used offtake data from 20 October 2004 to 31May 2005 to match the dates of Maizalss participation (asabove).

We used the software package PAST (Hammer et al.2001) to calculate a measure of diversity, Simpsons in-dex, 1! = 1ni=1 p2i , where pi is the frequency ofspecies i. Diversities were compared in a pairwise man-ner with a two-sample Monte Carlo bootstrap with 1000replicates. Simpsons index has been recommended overShannons index for both theoretical and statistical rea-sons (Keylock 2005), but results were similar with bothindices (not shown).

PER CAPITA CONSUMPTION AND MEAN PREY WEIGHTS

Per capita consumption and mean prey weights shoulddecrease when game is depleted. To compare species-by-species consumption between the 19881989 and cur-rent data sets (Yomybatom, OctoberMay), we treatedeach residence group as a single data point. Mean preyweights from our study were substituted when histori-cal records lacked weights. To minimize type I statisticalerror caused by tablewide comparisons, the significanceof a difference in means was calculated only when con-sumption in 19881999 was greater than in 20042005.We used POPTOOLS (Hood 2006) to compare the meanswith a two-sample Monte Carlo permutation test with9999 replicates (i.e., sampling without replacement, ow-

ing to low sample sizes and nonnormal distributions oferror residuals). For the spatial comparison we used aone-way analysis of variance (ANOVA) with a Tukey HSDpost hoc test to compare ln-transformed prey weights be-tween the main settlements and their satellites.

COMPARING DEPLETION TESTS

The depletion tests we used have different advantagesand disadvantages. Prey profile comparisons are justifiedunder optimal foraging theory (Rowcliffe et al. 2003).Nevertheless, although large changes in prey diversity areinterpretable (e.g., Fa et al. 2000; Jerozolimski & Peres2003), it is still not clear how powerful this approach isfor detecting the onset of game depletion or for detect-ing the depletion of individual species that have nevermade up a large numerical proportion of offtake (such astapirs) or, conversely, whether small but statistically sig-nificant changes in a diversity index are ecologically rele-vant. Moreover, comparisons crucially assume that substi-tutes for preferred game are included in the comparison.For example, if hunters switch from game to fish, but fishare not included in the analysis, then game diversity couldremain constant, even as game offtake declines. Addition-ally, small species tend not to be well represented in skullcollections (e.g., Bodmer 1994; da Silva et al. 2005), andembarrassment might lead some hunters to underreportkills of small species, which could obscure increases inofftake diversity. During our regular visits we reinforcedthe importance of reporting all kills. Because every res-idence group reported small species (sometimes exclu-sively), we are confident our data set is representative.Finally, it is important to complement diversity indiceswith measures of prey mass. Depletion is indicated onlyif an increase in offtake diversity is accompanied by adecrease in mean prey mass, indicating that hunters areaccepting less profitable prey.

Consumption rates are a more direct way of infer-ring depletion. Nevertheless, alternative reasons must bechecked separately. Reduced consumption of game couldresult from the introduction of substitute activities, suchas employment, or substitute protein sources, such aslivestock. Alternatively, because we were measuring percapita consumption, growth in the consumer popula-tion might have outstripped growth in offtake becauseof limited immigration from the source. Conversely, evenif game stocks are depleted, consumption rates couldbe maintained temporarily by the introduction of betterhunting technology, such as shotguns. In our case, littlehas changed economically since 1988 other than popula-tion growth.

We conducted both spatial and temporal tests of deple-tion. Nevertheless, spatial comparisons were less easilyinterpreted because we could not control for the possi-ble effects of habitat heterogeneity and/or hunter skilldifferences on the patterns of offtake.



Table 1. Offtake rates within the core Tayakomem and Yomybatom hunting zones (151 and 152 km2, respectively), October 2004 October 2005,compared with the Robinson and Redford maximum sustainable offtake (MSY).a

Tayakomem Yomybatom

retrieved + retrieved +Offtake rate (individuals/km2/year) retrieved retrieved + will die + retrieved retrieved + will die +and sourceb MSY offtake will die will survive offtake will die will survive

Ateles chamek, black-faced spider monkey (1) 0.38 0.29 0.33 0.58c 0.41c 0.48c 0.82c

Lagothrix lagotricha, common woolly monkey (2) 0.19 0.14 0.16 0.26c 0.44c 0.52c 0.84c

Mitu tuberosa, Razor-billed Curassow (3) 0.09 0.44c 0.57c 0.71c 0.19c 0.24c 0.30c

Penelope jacquacu, Spixs Guan (3) 0.14 0.12 0.14c 0.16c 0.14c 0.16c 0.19c

Tayassu pecari, white-lipped peccary (4) 0.16 0.71c 0.89c 1.68c 0.60c 0.75c 1.42c

Agouti paca, paca (2) 0.40 0.07 0.23 Alouatta seniculus, red howler monkey (2) 0.67 0.13 0.16 0.27 0.12 0.15 0.25Callicebus moloch, dusky titi monkey (2) 0.74 0.04 0.04 0.05 0.05 0.05 0.06Cebus albifrons, white-fronted capuchin monkey (2) 0.78 0.04 0.04 0.06 0.05 0.05 0.07C. apella, brown capuchin monkey (2) 0.72 0.11 0.16 0.34 0.07 0.09 0.20Dasyprocta variegata, brown agouti (2) 2.50 0.25 0.13 Geochelone denticulata, yellow-footed tortoise (5) 1.32 0.25 n/a n/a 0.22 n/a n/aMazama americana, red-brocket deer (6) 0.3 0.01 0.03 Myoprocta pratti, green acouchy (7) 4.07 0.06 0.36 Nasua nasua, South American coati (5) 0.35 0.07 0.10 0.13 0.07 0.09 0.12Ortalis guttata, Speckled Chachalaca (3) 0.54 0.09 0.02 Pipile cumanensis, Blue-throated Piping-Guan (3) 0.88 0.28 0.32 0.34 0.16 0.18 0.19Saimiri sciureus, common squirrel monkey (2) 2.04 0.03 0.02 Tayassu tajacu, collared peccary (4) 0.58 0.19 0.24 0.34 0.13 0.16 0.23

aThe first five species have offtake rates that exceed the MSY. Tayakomes measured offtake is multiplied by 1.4 to account for three residencegroups that did not participate in the study (see Data Analysis). Two estimates of collateral mortality are added (see also Data Analysis).Retrieved offtake includes only prey that were killed and retrieved by the hunter. The + will die adds wounded but escaped animals thathunters judged would die later in the forest. The + will survive adds wounded and escaped animals that hunters judged would not die laterin the forest.bValues for the calculation of the intrinsic rate of population increase and of densities taken from (1) McFarland-Symington (1987a, 1987b)and Robinson & Redford (1986); (2) Robinson & Bennett (2000); (3) Begazo & Bodmer (1998); (4) Gottdenker & Bodmer (1998) and Robinson& Redford (1986); (5) Peres & Nascimento (2006); (6) Robinson & Redford (1986); (7) Bodmer et al. (1997).cOfftake levels MSY.

MEAN KILL DISTANCES

For 19 of the species in Table 1 (omitting tapirs becauseof low sample size), we calculated the mean ordinal dis-tance of all kills with the four distance categories andfit a general linear model (GLM) with the explanatoryvariables MSY and settlement. Residuals were normallydistributed.

Results

Total Offtake

Over 1 year 99 hunters recorded 2,089 kills, for a total of15,875 kg of undressed prey weight (see SupplementaryMaterial), the live market weight of about 30 U.S. beefcattle. The average prey weight was 6.4 kg (0.23 SE),with a range of


Prey Profile Comparisons

The 19881989 offtake of woolly and spider monkeys inYomybatom approached or exceeded the Robinson andRedford MSY, respectively, as calculated by Alvard et al.(1997), even without incorporating collateral mortality.We inferred therefore that woolly and spider monkeyshave been hunted unsustainably in the Yomybatom hunt-ing zone for at least 17 years (Ohl 2004; Table 1).

Nonetheless, Yomybatoms current offtake profile forprimates was not significantly different from any of thethree historical profiles (Fig. 2a). Woolly and spider mon-keys made up over 80% of primate offtake in all data sets.Moreover, the primate species offtake in Yomybatom inthis study was significantly less diverse than in its satellitesettlement, Sarigeminis ( p = 0.011, see SupplementaryMaterial). In the spatial comparison between Yomybatomand Sarigeminis, for which we used the entire database,pooled by major game categories, we also failed to de-tect a significant difference in prey profile diversity (Fig.2b). Large primates and ungulates together made up 41%of the Yomybatom offtake versus 32% of the Sarigeminisofftake.

On the other hand, overall prey diversity in Tayakomem(1! = 0.81) was significantly higher than in its satel-

Figure 2. (a) Relativeabundances of primates killedover four time periods inYomybatom. All pairwisecomparisons between data setsare not significantly different atp 0.70. (b) Relativeabundances of prey killed byresidents of the main and satellitesettlements (Yo, Yomybatom vs.Sa, Sarigeminis, p = 0.868; Ta,Tayakomem vs. Ma, Maizals, p 0.05), which is consis-tent with the more diverse prey profile at Tayakomem(Fig. 2b).



Per Capita Consumption

In Yomybatom all but 1 of the 15 study species for whichwe had historical consumption data exhibited per capitameat consumption rates in 20042005 that were eitherhigher than or not significantly lower than those in 19881989 (Fig. 3). The only significant decrease was for thecollared peccary (Tayassu tajacu), despite an offtakeless than the MSY (Table 1) and locally high abundance.The clear explanation for the decrease is substitution bythe now-abundant and larger-bodied white-lipped pec-cary,whichwere inexplicably rare throughoutManu from1978 to 1990 (Silman et al. 2003). Only two kills wererecorded in 19881989. For Razor-billed Currasows onlyone kill was recorded in 19881989, so we could not cal-culate a useful historical consumption rate. Nonetheless,because 21 birds of this species were killed between Oc-tober 2004 and May 2005, we inferred that depletion hasnot yet occurred.

Hunting Distances by Species

Forty-one percent of all prey were hunted in the firstdistance category (i.e., 500 m from the communityperimeter), corresponding to just 6% of the hunting zone(Fig. 4a). More generally, the distance distribution of off-take was divided between those species hunted mostlywithin the first distance category (00.5 km; Fig. 4b)and those hunted mostly in the second category (0.53 km; Fig. 4c). This second group included four of thefive species flagged as being unsustainably hunted (Table1). A GLM showed that more vulnerable species (lowerMSY) were hunted at greater distances (Fig. 4d).

Figure 3. Average dailymeat consumption perperson and species inYomybatom, October to May19881989 versus20042005, with 95%confidence intervals.

Discussion

We deliberately used the overly permissive Robinson andRedford index to identify five species for which we couldbe confident that local offtake is currently unsustainable(Table 1). These included woolly and spider monkeys,which are among the five vertebrate species most vulner-able to hunting in Amazonia (Peres 2000). Nonetheless, inYomybatom, the five species were not depleted, despitedecades of hunting. Prey profiles, per capita consumptionrates, and mean prey weights did not indicate depletionin either the temporal or spatial comparisons (Fig. 2, 3,S8), with the easily explained exception of the peccaryspecies.

Our results for Tayakomem were less clear-cut becausewe lacked historical data. Tayakomem prey profiles weresignificantly more diverse than those of Maizals, and preyweights were significantly lower, suggesting depletion.Nevertheless, as can occur with any spatial comparison,Maizals was far enough away (22 km, Fig. 1) that we couldnot rule out habitat differences andwas small enough thata few skilled hunters could alter the offtake compositionappreciably. Moreover, hunters in Tayakomem still tookmany large primates (see SupplementaryMaterial), and allgame specieswerehuntedmore frequentlywithin 3 kmofthe community. Therefore, evidence for game depletionin Tayakomem was weak.

Finally, for most of the species that had offtake amountsbelow MSY (Table 1), historical data (Fig. 2a, Fig. 3) in-dicated that these species were also taken at low ratesin the past, supporting the interpretation that currentlow offtake is better attributed to some combination oflow hunter preference and intrinsically low abundance



Ani

mal

ski

lled

(n)

0

100

200

300

400

500

600

700

0-0.5 0.5-3 3-5.5 > 5.5

Hunting area (km)

Overnight tripOne-day trip

0

50

100

150

200

250

300

0-0.5 0.5-3 3-5.5 > 5.5

Hunting area (km)

Ani

mal

ski

lled

(n)

Other Birds RodentsSmall Primates TortoiseTayassu tajacu Penelope jacquacuCarnivores

Overnight trip

One-day trip

10

20

30

40

50

60

70

80

90

100

0-0.5 0.5-3 3-5.5 > 5.5

Hunting area (km)

Ani

mal

ski

lled

(n)

Tayassu pecariAteles chamekMitu tuberosaLagothrix lagotrichaAlouatta seniculusMazama americanaOvernight trip

One-day trip

0

10

20

30

40

50

60

70

80

90

100

0-0.5 0.5-3 3-5.5 > 5.5

Hunting area (km)

Ani

mal

ski

lled

(n)

Tayassu pecariAteles chamekMitu tuberosaLagothrix lagotrichaAlouatta seniculusMazama americanaOvernight trip

One-day trip

0.0

0.5

1.0

1.5

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

MSY

Mea

nki

ll di

stan

ce c

ateg

ory

(km

)

2.0

2.5

3.0

3.5Yomybatom, Tayakomem

ba

dc

Figure 4. Kill distance distributions by (a) all game (b) game categories (see Methods) that are hunted most oftenin the first distance category (close to the house). (c) Game species hunted more often in the farther distancecategories. (d) Mean kill distances compared with the maximum sustainable yield (MSY) (general linear model,mean kill distance category = 1.6 + 0.30 0.24ln(MSY) for Yomybatom, R2 = 25.1%, F2,35 = 13.6, p = 0.006).The significance of the ln(MSY) term was p = 0.007. Log transformation improved fit but was not necessary forsignificance. The effect of settlement was marginally nonsignificant (p = 0.067), in part due to an influential datapoint caused by three squirrel monkeys, out of six total, killed in Tayakomems fourth distance category (circled).The interaction effect was not significant (p = 0.39). Kills that took place during stays at distant secondary housesand during overnight camping trips, usually next to animal clay licks, are coded as overnight trips.

than to depletion from a previously high abundance. Wedid not have historical consumption data for red brocketdeer (Mazama americana), South American coatis (Na-sua nasua), or Speckled Chachalacas (Ortalis guttata),but Matisgenka hunters did not identify these species asfavored prey. The exception was the collared peccary,which appeared tohavebeendiscarded in favor of the cur-rently more abundant white-lipped peccary, which wasdeemed sustainably exploited in 19881989 (Alvard et al.1997).

Evidence for Source-Sink Dynamics

Jerozolimski and Peres (2003) surveyed 31 hunted Neo-tropical sites and found that prey diversity generally dou-bles in settlements older than 18 years, indicating de-pletion. Yomybatom has been occupied continuously foralmost 30 years, and Tayakomem for more than 40, yetno such changes were noted. The parks prohibition offirearms may be a factor, but the presence of vast, essen-tially unhunted, game refuges may be largely responsible.



Sourcesink dynamics are often invoked in hunting stud-ies (Novaro et al. 2000; Peres 2001; Siren et al. 2004; No-varo et al. 2005; Peres & Nascimento 2006), and manage-ment recommendations regularly include the establish-ment of reserve areas (e.g., Fragoso et al. 2000; Bodmer& Robinson 2004).

Within the hunting literature the basis for demonstrat-ing a sourcesink dynamic typically rests on showing theexistence of a mortality sink, which is inferred when off-take exceeds the calculated MSY of the focal area (Ta-ble 1). If threatened species continue to be taken, es-pecially at high levels, and if a candidate source area isnearby, then immigration is inferred. This protocol makestwo assumptions that our results support. (1) Huntingmortality in the sink has not been counterbalanced fullyby higher, density-dependent population growth. Ouruse of the Robinson and Redford index, which overesti-mates the realmaximumsustainable yield (Milner-Gulland& Akcakaya 2001), to identify unsustainably exploitedspecies makes it unlikely that density dependence hascounterbalanced hunting mortality. (2) Hunters have notincreased their hunting zone. For 1-day forays this is al-most a given because hunters cannot increase their max-imum walking range beyond a 10-km radius. Thus, inYomybatom the recorded outer limit of 1-day forays wasa salt lick about 10 km from the village center, the samespot thatwas georeferenced in a 1996 study ofMatsigenkalandscape ecology (Shepard et al. 2001) and the samedistance as used by Alvard et al. (1997) in their analy-sis of the 19881989 offtake data. In Tayakomem we didnot have comparable historical data, but over both set-tlements, our current offtake data indicated that 90% ofkills were obtained during 1-day forays (Fig. 4), and evenfor the vulnerable large primate species, almost 70% ofofftake occurred within the core hunting zone. Most ofthe rest of the offtake was taken during temporary staysat secondary houses, which, along with the satellite set-tlements, Sarigeminis, Maronaro, and Maizals, representincreases in the area used for hunting.

Source Risk

Prey refuges and sourcesink dynamics can form part ofthe basis for a game management plan, but only if dis-persal from the source to the sink is limited so that em-igration will not threaten the viability of source popula-tions (Amarasekare 2004).We found thatmore-vulnerablespecies (lower MSY) were generally hunted farther awayfrom the settlements (Fig. 4bd). This result probablymeans that immigration is slow enough that it is un-able to restore sink populations completely (Siren et al.2004). As a result low-MSY species are more depletedand hunters must search longer to kill them. Nonethe-less, following Amarasekare (2004), a full analysis ofsource risk requires assessments of source quality and

of the nature of density dependence as it affects dispersalbehavior.

Policy Implications

Parks are held to a higher conservation standard thanare indigenous territories; a park has the added role ofprotecting vulnerable vertebrate species in populationslarge enough to maintain their biodiversity maintenancefunctions, a role that is not obviously compatible withthe presence of a growing human population (Terborgh1999). Nevertheless, in this case Peruvian law grants resi-dence rights to preexisting indigenous populations in na-tional parks as long as their presence does not interferewith conservation objectives (Shepard et al. 2007). Nev-ertheless, our results support the idea that sourcesinkdynamics help maintain populations of game species inthe face of hunting by theMatsigenka inhabitants of ManuPark; thus, Matsigenka presence is currently compatiblewith conservation.

What about the future? It appears inevitable that humanpopulation growth will eventually threaten the viabilityof some game species in Manu Park, particularly largeprimates. Sourcesink dynamics can provide the basis of asolution for this challenge because total offtake of a givenspecies in a hunting zone may not exceed its immigrationrate (e.g., Siren et al. 2004), independent of the numberof consumers. Thus, settlement spread (which starts asthe establishment of temporary secondary houses andeventually results in permanent satellite settlements) isproximately more important than population growth perse, although the latter drives the former.

A key research goal is therefore to estimate immigrationrates for vulnerable species and to estimate the degree towhich dispersal draws down source populations. Theseestimates, combined with projections of growth in thenumber and spatial distribution of settlements, will allowprojections of the total source area drawn upon by thehunting sinks. Even without quantitative estimates, sta-bilizing the Matsigenka population around existing set-tlements should be a primary goal for Manu Park; thus,another research goal should be to understand the socio-economic factors that promote settlement stability.

The results of such an interdisciplinary research pro-gramhave applicability across theNeotropics. Indigenousterritories account for 52%of all reserves by acreage in thenine Amazonian countries, and overall cover 100 millionha or 21% of forested area in the Brazilian Amazon (Peres1994).Moreover, 70%of Amazonian parks already containpeople (Terborgh & Peres 2002), and most protected ar-eas being created today explicitly include people, withthe notable example of Brazilian Amazonia, where a vastnetwork of national forests and extractive and sustainabledevelopment reserves are legally occupied by nontribalforest dwellers subsisting partially or entirely on game



vertebrate meat. Managing protected areas so that hu-man inhabitants exact the minimal cost to biodiversityand even contribute to the defense of protected areasrepresents one of the largest opportunities for conserva-tion in the Neotropics (Zimmerman et al. 2001; Nepstadet al. 2006; Shepard et al. 2007).

Acknowledgments

We thank our Matsigenka friends for their cooperationand Perus Instituto de Recursos Naturales (INRENA) forgranting us research permission. Logistical and intellec-tual support were given by N. Gibson, C. Huamantupa, Q.Meyer, S. Miller, F. Puygrenier, J. Terborgh, W. Townsend,and S. Zent. This work was funded by the LeverhulmeTrust.

Supplementary Material

The following are available as part of the on-line articlefrom http://www.blackwell-synergy.com/: detailed par-ticipatory monitoring protocol (S1), list of all recordedkills (S2), timeline of Matsigenka population growth (S3),photograph of a hunter monitoring station (S4), offtakemonitoring sheet (S5), map of hunting trails and distancecategories in Yomybatom (S6), hunter-foray monitoringsheet (S7), collateral mortality statistics (S8), spatial com-parison of prey profiles (S9), and frequency distributionof prey weights (S10). The authors are responsible forthe content and functionality of these materials. Queries(other than absence of the material) should be directedto the corresponding author.

Literature Cited

Alcorn, J. 1993. Indigenous peoples and conservation. ConservationBiology 7:424426.

Alvard, M. S., J. G. Robinson, K. H. Redford, and H. Kaplan. 1997. Thesustainability of subsistence hunting in the Neotropics. Conserva-tion Biology 11:977982.

Amarasekare, P. 2004. The role of density-dependent dispersal insourcesink dynamics. Journal of Theoretical Biology 226:159168.

Begazo, A. J., and R. E. Bodmer. 1998. Use and conservation of Cracidae(Aves: Galliformes) in the Peruvian Amazon. Oryx 32:301309.

Bodmer, R. E. 1994. Managing wildlife with local communities: the caseof the Reserva Comunal Tamshiyacu-Tahuayo. Pages 113134 in D.Western, M. Wright, and S. Strum, editors. Natural connections. Is-land Press, Washington, D.C.

Bodmer, R. E., J. F. Eisenberg, and K. H. Redford. 1997. Hunting andthe likelihood of extinction of Amazonian mammals. ConservationBiology 11:460466.

Bodmer, R. E., and J. G. Robinson. 2004. Evaluating the sustainabilityof hunting in the Neotropics. Pages 217225 in K. M. Silvius, R.E. Bodmer, and J. M. V. Fragoso, editors. People in naturewildlifeconservation in South and Central America. Columbia UniversityPress, New York.

Fa, J. E., J. E. G. Yuste, and R. Castelo. 2000. Bushmeat markets onBioko Island as a measure of hunting pressure. Conservation Biology14:16021613.

Fragoso J. M. V., K. M. Silvius, and M. Villa-Lobos. 2000. Wildlife Manage-ment at the Rio das Mortes Xavante Reserve, MT, Brazil: integratingindigenous culture and scientific method for conservation. Volume4. World Wildlife Fund-Brazil, Brasilia.

Gottdenker, N., and R. E. Bodmer. 1998. Reproduction and productiv-ity of white-lipped and collared peccaries in the Peruvian Amazon.Journal of Zoology 245:423430.

Gundersen, G., E. Johannesen, H. P. Andreassen, and R. A. Ims. 2001.Source-sink dynamics: how sinks affect demography of sources. Ecol-ogy Letters 4:1421.

Hammer, O., D. Harper, and P. Ryan. 2001. PAST: paleontological statisti-cal software package for education and data analysis. PalaeontologiaElectronica 4: http://palaeo-electronica.org/2001 1/toc.htm

Harmon, D. 1998. The other extinction crisis: declining cultural di-versity and its implications for protected area management. Pages352359 in N. W. P. Munro and J. H. M. Willison, editors. Linkingprotected areas with working landscapes, conserving biodiversity.Science and Management of Protected Areas Association, Wolfville,Nova Scotia.

Hill, K., G. McMillan, and R. Farina. 2003. Hunting-related changes ingame encounter rates from 1994 to 2001 in the Mbaracayu Reserve,Paraguay. Conservation Biology 17:13121323.

Hood, G. M. 2006. PopTools. Software for the analysis of ecologicalmodels. Version 2.7. Pest Animal Control Co-operative ResearchCentre, Canberra, Australia. Available fromhttp://www.cse.csiro.au/poptools (accessed July 2006).

Jerozolimski, A., and C. A. Peres. 2003. Bringing home the biggest ba-con: a cross-site analysis of the structure of hunter-kill profiles inNeotropical forests. Biological Conservation 111:415425.

Joshi, N. V., and M. Gadgil. 1991. On the role of refugia in promoting theprudent use of biological resources. Theoretical Population Ecology40:211229.

Keylock, C. 2005. Simpson diversity and the Shannon-Wiener index asspecial cases of a generalized entropy. Oikos 109:203207.

May, R. M. 1978. Host-parasitoid models in patchy environments: a phe-nomenological model. Journal of Animal Ecology 47:833843.

McFarland-Symington, M. 1987a. Sex ratio and maternal rank in wildspider monkeys: when daughters disperse. Behavioral Ecology andSociobiology 20:421425.

McFarland-Symington, M. 1987b. Ecological and social correlates ofparty size in the black spider monkey, Ateles paniscus chamek. PhDdissertation. Princeton University, Princeton, New Jersey.

Milner-Gulland, E. J., and H. R. Akcakaya. 2001. Sustainability indicesfor exploited populations. Trends in Ecology & Evolution 16:686692.

Nepstad, D., et al. 2006. Inhibition of Amazon deforestation and fire byparks and indigenous lands. Conservation Biology 20:6573.

Novaro, A. J., K. H. Redford, and R. Bodmer. 2000. Effect of huntingin sourcesink systems in the Neotropics. Conservation Biology14:713721.

Novaro, A. J., M. C. Funes, and R. S. Walker. 2005. An empirical testof sourcesink dynamics induced by hunting. Journal of AppliedEcology 42:910920.

Ohl, J. 2004. El Ecoturismo como oportunidad para un desarrollosostenible? La economa de los Matsiguenkas en el Parque Nacionaldel Manu, Peru. Deutsche Gesellschaft fur Technische Zusammenar-beit, Eschborn, Germany.

Ohl, J., A. Wezel, G. H. Shepard Jr., and D. W. Yu. 2007. Swidden agri-culture in a human-inhabited protected area: the Matsigenka nativecommunities of Manu National Park, Peru. Environment, Develop-ment, and Sustainability: DOI: 10.1007/S10668-007-9086-3.

Peres, C. A. 1994. Indigenous reserves and nature conservation in Ama-zonian forests. Conservation Biology 8:586588.

Peres, C. A. 2000. Evaluating the impact and sustainability of subsis-tence hunting at multiple Amazonian forest sites. Pages 3157 in J.G. Robinson and E. L. Bennett, editors. Hunting for sustainability intropical forests. Columbia University Press, New York.



Peres, C. A. 2001. Synergistic effects of subsistence hunting and habitatfragmentation on Amazonian forest vertebrates. Conservation Biol-ogy 15:14901504.

Peres, C. A., and H. S. Nascimento. 2006. Impact of game hunting by theKayapo of south-eastern Amazonia: implications for wildlife conser-vation in tropical forest indigenous reserves. Biodiversity and Con-servation 15:26272653.

Peres, C. A., and B. Zimmerman. 2001. Perils in parks or parks in peril?Reconciling conservation in Amazonian reserves with and withoutuse. Conservation Biology 15:793797.

Redford, K. H. 1991. The ecologically noble savage. Orion 9:2429.

Redford, K. H., and A. M. Stearman. 1993. Forest-dwelling native Ama-zonians and the conservation of biodiversity: interests in commonor in collision? Conservation Biology 7:248255.

Redford, K. H., and S. E. Sanderson. 2000. Extracting humans from na-ture. Conservation Biology 14:13621364.

Robinson, J. G. 1993. The limits to caring: sustainable living and the lossof biodiversity. Conservation Biology 7:2028.

Robinson, J. G., and K. H. Redford. 1986. Body size, diet, and popula-tion density of Neotropical forestmammals. The AmericanNaturalist128:665680.

Robinson, J. G., and K. H. Redford. 1991. Sustainable harvest of Neotrop-ical wildlife. Pages 415429 in J. G. Robinson and K. H. Redford,editors. Neotropical wildlife use and conservation. University ofChicago Press, Chicago.

Robinson, D. J., and E. L. Bennett. 2000. Carrying capacity limits tosustainable hunting in tropical forests. Pages 1330 in E. L. Bennettand D. J. Robinson, editors. Hunting for sustainability in tropicalforests. Columbia University Press, New York.

Rowcliffe, J. M., G. Cowlishaw, and J. Long. 2003. A model of humanhunting impacts in multi-prey communities. Journal of Applied Ecol-ogy 40:872889.

Schwartzman, S., A. Moreira, and D. Nepstad. 2000. Rethinking tropical

forest conservation: perils in parks. Conservation Biology 14:13511357.

Shepard, G. H., Jr. 2002. Primates in Matsigenka subsistence and world-view. Pages 101136 in A. Fuentes and L. Wolfe, editors. Primatesface to face: the conservation implications of human and nonhumanprimate interconnections. Cambridge University Press, Cambridge,United Kingdom.

Shepard, G. H., Jr., D. W. Yu, M. Lizarralde, and M. Italiano. 2001. Rain-forest habitat classification among the Matsigenka of the PeruvianAmazon. Journal of Ethnobiology 21:138.

Shepard, G. H., Jr., K. Rummenhoeller, J. Ohl, and D. W. Yu. 2007. Trou-ble in paradise: indigenous populations, anthropological policies,and biodiversity conservation in Manu National Park, Peru. Journalof Sustainable Forestry: in press.

Silman, M. R., J. W. Terborgh, and R. A. Kiltie. 2003. Population regula-tion of a dominant rain-forest tree by a major seed predator. Ecology84:431438.

da Silva, M. N. F., G. H. Shepard Jr., and D. W. Yu. 2005. Conservationimplications of primate hunting practices among the Matsigenka ofManu National Park, Peru. Neotropical Primates 13:3136.

Siren, A., P. Hamback, and J. Machoa. 2004. Including spatial heterogene-ity and animal dispersal when evaluating hunting: a model analysisand an empirical assessment in an Amazonian community. Conser-vation Biology 18:13151329.

Terborgh, J. 1999. Requiem for nature. Island Press, Washington, D.C.Terborgh, J. 2000. The fate of tropical forests: a matter of stewardship.

Conservation Biology 14:13581361.Terborgh, J., and C. A. Peres. 2002. The problem of people in parks.

Pages 307319 in J. Terborgh, C. van Schaik, L. Davenport, and M.Rao, editors. Making parks work. Island Press, Washington, D.C.

Zimmerman, B., C. A. Peres, J. R. Malcolm, and T. Turner. 2001. Conser-vation and development alliances with the Kayapo of south-easternAmazonia, a tropical forest indigenous people. Environmental Con-servation 28:1022.


Documents

Ohl Etal 2007 MatsigenkaHunting ConBio w COVER