Research Note

232

Conservation Biology, Pages 232–237Volume 16, No. 1, February 2002

Bushmeat Exploitation in Tropical Forests:an Intercontinental Comparison

JOHN E. FA,* CARLOS A. PERES,† AND JESSICA MEEUWIG‡

*Durrell Wildlife Conservation Trust, Les Augrès Manor, Trinity, Jersey JE3 5BP, United Kingdom,email jfa@durrel.org/

†School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom

‡Project Seahorse, Department of Biology, McGill University, 1205 Ave. Dr. Penfield, Montreal H3A 1B1, Canada

Abstract:

We calculated extraction and production rates of bushmeat species in two main tropical, moist-for-est regions, the Amazon and Congo basins. Extraction was estimated from the average number of animalsconsumed per person per year from anthropological studies that reported animal kills brought into settle-ments in the regions. We calculated extraction rates (kg/km

2

/year) for 57 and 31 mammalian taxa in theCongo and Amazon, respectively. We then examined the sustainability of these extraction rates by basin andby taxa, using extraction-to-production (

E:P

) mass-balance equations. Production (tonnes/year) was calculatedas the product of

r

max

(the intrinsic rate of natural increase), mammal biomass, and total area of forest ineach region. Species exploitation rates at specific body masses were significantly greater in the Congo than inthe Amazon. The

E:P

ratio for the Congo was 2.4, 30 times the Amazon’s ratio of 0.081. Thus, Congo Basinmammals must annually produce approximately 93% of their body mass to balance current extraction rates,whereas Amazonian mammals must produce only 4% of their body mass. We calculated sustainability levelsderived from Robinson and Redford’s harvest model for each taxa. On a basin-wide level, 60% and none ofthe mammal taxa in the Congo and Amazon basins, respectively, were exploited unsustainably. To evaluatethe effect of error on the estimates of

E:P

, we conducted a sensitivity analysis, which suggests that the mass-balance was most sensitive to error in standing stock but that our results are robust. We estimated that over5 million tons of wild mammal meat feed millions in Neotropical (0.15 million) and Afrotropical (4.9 mil-lion) forests annually. Our Congo basin estimates are four times higher than those calculated for the regionby other workers, and we conclude that the current situation of bushmeat extraction in African rain forests ismore precarious than previously thought.

Explotación de Carne Silvestre en Bosques Tropicales: una Comparación Intercontinental

Resumen:

Calculamos las tasas de extracción y de producción de especies cinegéticas en dos importantes re-giones tropicales de bosque húmedo: las cuencas del Amazonas y del Congo. Se estimó la extracción a partirdel número promedio de animales consumidos por año en estudios antropológicos que registraron animales

cazados y llevados a los poblados en las regiones. Calculamos tasas de extracción (kg/km

2

/año

�

1

) para 57 y31 taxa de mamíferos en el Congo y el Amazonas, respectivamente. Posteriormente examinamos la sustent-abilidad de estas tasas de extracción por cuenca y por taxa, utilizando ecuaciones de balance de masas ex-tracción a producción (

E:P

). La producción (ton/año

�

1

) se calculó como el producto de

r

max

la biomasa demamíferos y la superficie total de bosque en cada región. Las tasas de explotación de especies para masascorporales específicas fueron significativamente mayores en el Congo que en el Amazonas. La proporción

E

:

P

para el Congo 2.4, 30 veces mayor que la proporción 0.081 en el Amazonas. Así, los mamíferos de la Cuencadel Congo anualmente deben producir aproximadamente el 93% de su masa corporal para equilibrar las ac-tuales tasas de extracción, mientras que los mamíferos del Amazonas deben producir solo el 4% de su masacorporal. Para cada taxa calculamos los niveles de sustentabilidad derivados del modelo de extracción deRobinson y Redford. Al nivel de toda la cuenca, 60% y ninguno de los taxa de mamíferos en la cuenca delCongo y del Amazonas respectivamente, fueron explotados no sustentablemente. Para evaluar el efecto del

Paper submitted June 28, 2000; revised manuscript accepted April 18, 2001.

Conservation BiologyVolume 16, No. 1, February 2002

Fa et al. Bushmeat Exploitation in Tropical Forests

233

Introduction

Vertebrate defaunation of the world’s remaining tropicalforests through overhunting is considered a major causeof biodiversity loss, in some cases more important thandeforestation (Redford 1992). Robinson et al. (1999)have indicated that exploitation of bushmeat by tropicalforest dwellers has increased in recent years, because ofgrowing human populations, greater access to undis-turbed forests, changes in hunting technology, scarcityof alternative protein sources, and the fact that bush-meat is often a preferred food ( Robinson & Bennett2000). Mammals hunted for subsistence or commercialpurposes are particularly affected (Alvard et al. 1997;Robinson et al. 1999; Robinson & Bodmer 1999; Robin-son & Bennett 2000).

Bushmeat hunting is the single most geographicallywidespread form of resource extraction in tropical for-ests and can affect the core of even the largest and leastaccessible nature reserves ( Peres & Terborgh 1995).Game harvests in South America and Africa usually ex-ceed production (Robinson & Bodmer 1999), even inthe case of traditional aborigine societies still using rudi-mentary hunting technology (Alvard et al. 1997). Suchuncontrolled exploitation will bring about marked pop-ulation declines, and eventually the extinction of a num-ber of game species. Coupled with threats from habitatloss, even from historical deforestation (Cowlishaw1999), global extinctions of the most sensitive speciessuch as primates are likely to occur as an accumulationof local disappearances. This may result in long-termchanges in tropical forest dynamics through the loss ofseed dispersers, large granivores, frugivores, and “habi-tat landscapers” such as large forest mammals (Dirzo &Miranda 1991; Chapman & Onderdonk 1998; Wright etal. 2000).

Numerous international conservation organizationsbelieve that tropical forest faunas are seriously endan-gered by current extraction levels of subsistence andcommercial hunting (World Society for the Protectionof Animals 1995; Ape Alliance 1998), but few studieshave quantified this at a regional scale. Such a broaderpicture would help conservationists understand the ex-tent of the problem and serve to highlight differencesand similarities among geographical areas. Moreover, in-tercontinental comparisons of consumption and extrac-

tion rates can shed light on the interaction between eco-system productivity, harvestable game biomass, andhuman consumers.

Methods

We used data on the extraction rates of mammal speciesin the Congo and Amazon basin from a review of hunt-ing studies by Fa and Peres (2001). The average numberof animals consumed per person per year was estimatedfollowing Peres (2000), which builds on previous esti-mates by Redford and Robinson (1987). We used meanbody mass (Fa & Purvis 1997; Peres 2000) to estimateextraction rates ( kg/km

2

/year), for a total of 57 and 31mammalian taxa in the Congo and Amazon, respectively.Data on the number of animals and biomass extractedper species were derived from anthropological studiesthat reported animal kills brought into settlements (23and 14 rural settlements in the Amazon and Congo, re-spectively) during a period

�

1 year. An average extrac-tion rate for each species was derived from all studiesthat reported the species as hunted in the settlement. Al-though the hunting studies incorporated reflect a varietyof habitat types and hunting conditions, our sample maynot represent all regions or forest types in the geograph-ically large study areas. Given the conservation urgencyof this issue, however, we believe results from our com-pilation will emphasize the gross intercontinental differ-ences that exist in wild meat exploitation. To standard-ize the quality and representativeness of the studies weincluded, we carefully chose sites that were studied overlonger periods and that did not concentrate on un-hunted areas. The African databank is based on a smallersample size than the South American one, a direct reflec-tion of the scarcity of data being collected from Africanforests.

The species we considered represent all mammal spe-cies hunted in each region. The main taxa representedare primates (19 African and 11 Amazonian), ungulates(14 and 3), rodents (9 and 6), and carnivores (11 and 4).To assess the sustainability of current extraction ratesfor each major taxon and species, we determined ratiosof extraction to production (

E

:

P

) (Hennemann 1983)for the Congo and Amazon. At the basin-wide level,

r

max

was derived empirically from mean body mass. For dif-

error en las estimaciones de

E :P

, realizamos un análisis de sensibilidad que sugiere que el balance de masafue más sensible al error que en animales en pie pero que nuestros resultados son robustos. Estimamos quemás de 5 millones de toneladas de carne de animales silvestres sirven de alimento anualmente para millonesde personas en bosques neotropicales (0.15 millón) y afrotropicales (4.9 millones). Nuestras estimacionespara la Cuenca del Congo son cuatro veces más altas que las realizadas por otros investigadores en la región,concluimos que la situación actual de la extracción de carne silvestre en selvas africanas es más precaria de

lo que se pensaba anteriormente.

234

Bushmeat Exploitation in Tropical Forests Fa et al.

Conservation BiologyVolume 16, No. 1, February 2002

ferent taxa, where sufficient life-history data were avail-able,

r

max

was calculated by the method of Cole (1954);otherwise,

r

max

was estimated empirically from bodymass of primates (Robinson & Redford 1986

a

,

b

) andother taxa (Hennemann 1983). Production (tons/year)was calculated as the product of

r

max

, mammal biomass,and total area of forest in each region.

Our mass-balance estimates were based on production(

P

) derived from body size (

W

a

) and

r

max

and on field-derived extraction rates (

E

) and biomass estimates. Allthese variables include some error. To evaluate the ef-fect of error on the estimates of

E

:

P

, we conducted asensitivity analysis by substituting lower and upper val-ues for

W

,

r

max

and

B

into the mass-balance equation andevaluating the effect on the

E

:

P

ratio (Table 1). Theselower and upper limits are a function of either the 95%confidence limits of the corresponding regression equa-tion or are based on variation in reported estimates of

B

.

Results and Discussion

Rates of exploitation of species at specific body masseswere significantly greater in the Congo than in the Ama-zon (Table 2). In the Congo, 177.7

�

358.4 kg/km

2

/year(

n

�

24) of meat were contributed by taxa of

�

10 kgand 35.4

�

72.2 kg /km

2

/year (

n

�

33) of meat werecontributed by taxa of

�

10 kg. For the Amazon, taxa of

�

10 kg contributed 3.69

�

3.9 kg /km/year (

n

�

7 ),whereas taxa of

�

10 kg contributed 0.6

�

0.9 kg/km/year (

n

�

24). Extraction rates increased with body sizeon both continents, reflecting hunters’ preference forlarge-bodied animals. Because

r

max

and potential produc-tion naturally declines with increased body mass, theprobability of sustainable harvests steeply declines forlarger-bodied species.

The

E

:

P

ratio for the Congo was 2.4, 30 times the Am-azon’s ratio of 0.081 (Table 3). Consequently, Congo Ba-sin mammals must annually produce approximately 93%of their body mass to balance current extraction rates,

whereas Amazonian mammals must produce only 4% oftheir body mass. Of the Congo mammal taxa, 60% wereharvested unsustainably: most ungulates (93%), primatesand carnivores (63%), and rodents (11%) (Fig. 1). Herewe define unsustainable harvesting as extraction ex-ceeding 20% of production, which is considered sustain-able for long-lived taxa (Robinson & Redford 1991). At abasin-wide scale, all Amazonian taxa appear to be ex-ploited at levels below 20% production.

Results of our sensitivity analysis suggest that

E

:

P

wasaffected by variation in

W

a

, provided that standing stock(

B

) remained constant (Table 1). Doubling or halving

W

a

led to a 1.2-fold change in mean

E

:

P

in both theCongo and the Amazon. The effects of error in

r

max

werecomparable to those in

W

a

. By using the 95% lower andupper confidence limits for

r

max

we calculated a 1.24-and 1.2-fold change in

E

:

P

for Congo and Amazon, re-spectively (Table 1). The mass-balance estimates weremore sensitive to error

B

, with a 1.37- and 1.93-foldchange in

E

:

P

ratio for the Congo and the Amazon, re-spectively, at the 95% confidence limits. The relativelylarge change for the Amazon reflects the imprecision in

B

associated with the small sample size available for itsestimation (

n

�

3). Combined,

r

max

and

B

led to 1.77-fold (Congo) and 2.44 (Amazon) -fold changes in

E

:

P

ra-tios for the two basins (Table 3), which corresponds to avariation in

E

:

P

from 1.48 and 4.31 in the Congo and0.04 to 0.23 in the Amazon.

The sensitivity analysis suggests that our results are ro-bust. The high

E

:

P

values in the Congo are the result oflow biomass and low average

r

max

(i.e., a community of afew large animals of low productivity). This is an unlikelycombination (Fa & Peres 2001), so the high

E

:

P

ratio isunlikely. The lower limit of 1.48, however, indicates thatat best extraction is almost 50% higher than annual pro-duction, a grave conservation concern. In the Amazon,

E

:

P

ratios remain relatively low. In the worst-case scenario,extraction is 20% of production; this value is thought tobe at the upper level of sustainability, suggesting that ex-ploitation rates in the Amazon should not increase.

Table 1. Sensitivity analysis of the effects of error on estimates of average body size (

W

a

),

r

max

(the intrinsic rate of natural increase), and standing stock (

B

).*

Basin Variable Mean Lower limit Upper limit Mean

E:P

Lower

E :P

Upper E :P Error

Congo Wa 17.1 8.5 35 2.39 2 2.91 1.21rmax 0.39 0.31 0.48 2.39 1.95 2.98 1.24B 2848 1979 3716 2.39 1.86 3.46 1.37combined 17.1 2.39 1.38 4.31 1.77

Amazon Wa 11.6 5.75 23 0.08 0.07 0.1 1.20rmax 0.44 0.35 0.53 0.08 0.07 0.1 1.20B 1109 500 1718 0.08 0.05 0.18 1.93combined 11.6 0.08 0.04 0.23 2.44

*Lower and upper limits for Wa are 0.5 and 2 times the estimate respectively. Lower and upper limits for rmax and B are the 95% confidence lim-its based on the regression model (rmax) or the sample (B). Mean lower and upper E:P are based on the mean value and lower and upper limitsif the other variables are held constant. For the combined model, lower and upper limits are based on the lower or upper limits for both rmaxand B simultaneously. Error indicates the x-fold difference from the expected E:P.

Conservation BiologyVolume 16, No. 1, February 2002

Fa et al. Bushmeat Exploitation in Tropical Forests 235

We estimated that over 5 million tonnes of wild mam-mal meat feed millions in Neotropical and Afrotropicalforests annually. Our calculated extraction figures aresignificantly lower in Neotropical than in Afrotropicalforests. Furthermore, our Congo basin estimates werefour times higher than those calculated by Wilkie andCarpenter (1999). But the per-capita per-day meat con-sumption (calculated from extraction in tonnes/yearfrom Table 3, divided by a total human population sizeof 33.35 million inhabitants, and multiplied by 0.70 toconvert undressed to dressed meat) was 282.3 g/person/day. This figure is not unrealistic because it falls withinthe reported range of consumption for Africa (151.8 �72.3 g/person/day, range 50–290, n � 15) (Noss 1995;DelVingt 1997; Wilkie & Carpenter 1999) and for the

Neotropics (205.9 � 145.9 g/person/day, range 7–582,n � 54) ( Jerozolimski 1998).

The key question here is how much our use of extrac-tion rates biases the mass-balance extrapolation. If themodel were constrained to refer to E:P on a square-kilo-meter basis, results would simply show that, where ani-mals are exploited, the mass-balance result is the ratiobetween E and P. But if one scales up to a larger geo-graphical area, in our case whole river basins, the appro-priateness depends on the proportion of the region sub-ject to exploitation. Similarly, the degree to whichunexploited areas serve as refugia for sources of biomassthat move into already hunted sink areas would also af-fect the extrapolations (Novaro et al. 2000). In theCongo, where much of the basin is subject to hunting

Table 2. Exploitation (E [kg/ km2/year]) of mammal species arranged according to log body mass for the Congo and Amazon basins.

Basin Log10 body mass Mean SD Minimum Maximum n

Amazon1.0 3.08 4.14 0.01 10.23 72.0 0.18 0.43 0.01 1.40 103.0 0.86 0.88 0.03 2.32 74.0 0.48 0.64 0.02 1.42 55.0 3.73 3.73 3.73 2

Total 1.18 2.32 0.01 10.23 31Congo

1.0 0.12 0.16 0.01 0.44 62.0 25.49 65.50 0.11 272.84 193.0 75.75 91.24 1.03 262.13 94.0 168.04 368.03 0.70 1348.89 145.0 192.66 364.37 1.31 1130.50 9

Total 92.16 241.87 0.01 1348.89 57

Table 3. Calculation of ratios of exploitation (E [kg/km2/year]) to production (P [kg/km2/year]) ratios for the Congo and Amazon basins based on specific production (rmax/year).

Variable Congo Amazon

Area of forest (km2)a 1,856,207 3,751,160Rural population (no. of inhabitants)a 24,354,700 8,158,897Mean body mass Wa (kg)b 17.1 11.6rmax/yearc 0.39 0.44Exploited mammal standing biomass (B)d 2,848 kg/km2/year 5,286,478 tonnes/year 1,109 kg/km2/year 4,160,036 tonnes/yearProduction (P)e 1,111 kg/km2/year 2,061,726 tonnes/year 488 kg/km2/year 1,830,416 tonnes/yearExtraction (E)f 2,645 kg/km2/year 4,909,668 tonnes/year 39.5 kg/km2/year 148,171 tonnes/yearE:Pg 2.38 0.081Required rmax/year to balance current Eh 0.93 0.036aFrom Wilkie and Carpenter (1999) and Peres (2000).bWeighted average body mass of mammals was based on their densities (Fa & Purvis [1997]), which thereby accounts for varying contribu-tions of size classes to the overall mean.cEstimated from Hennemann (1983).dCalculated average crude mammalian biomass in African forests is 2848.0 (95% confidence limits [CI] � 1979–3716 kg/km2, n � 9) and1108.7 (95% CI � 500–1718 kg/km2, n � 3) (Fa & Peres 2001), and � ( Wa � density � area of occupancy)/c.f., where c.f. is the conversion fac-tor from ecological densities (Fa & Purvis 1997) to density.eCalculated as B � rmax and B � rmax � area of forest.fSum of taxa-specific extraction rates, Et , from Fa and Peres (2001) and � ( Et � area of occupancy). Around 60–70% of game extracted fromAfrican forests is commercial supply to another 9 million people in urban centers (Wilkie & Carpenter 1999).gExtraction divided by production.hSpecific production necessary to balance current levels of exploitation, calculated as E / B.

236 Bushmeat Exploitation in Tropical Forests Fa et al.

Conservation BiologyVolume 16, No. 1, February 2002

and where human population density is high, we can as-sume that our estimates of extraction rates reflect thoseof the entire region. Few areas of the Congo are not sub-ject to exploitation (Ape Alliance 1998; Wilkie & Car-penter 1999). In the Amazon, however, where vast for-est tracts remain unhunted, our E is likely to be high andthus to represent a worse-case scenario. Thus, we arguethat the E:P ratios for the Congo are probably relativelyaccurate on both a square-kilometer basis and on a basin-wide scale. In the Amazon, the E: P ratio is probablymost appropriate in hunted areas (on a square-kilometerlevel) and less accurate on a basin scale.

Our extrapolations from local data to entire basinsmust be viewed with some caution. The picture thatemerges here is a reflection of the current state of evi-dence. Notwithstanding our data sources, we believethat the overall picture of extraction to productionemerging from our study is realistic and of heuristicvalue. An important result is that although our E:P esti-mates suggest that Amazonian game species do not ap-pear to be harvested unsustainably at a large spatialscale, they more often than not are overexploited at thescale of village catchments. Therefore, we do not advo-cate complacency, because the main reason for Neotrop-ical forests being less vulnerable at present is their lowerhuman densities. As we have seen in the Congo Basin,

hunting levels of forest mammals reflect the high humanpopulation in the region. Coupled with acceleratinghabitat loss (Cowlishaw 1999; Musters et al. 2000), thecurrent bushmeat hunting scenario foretells cata-strophic results for African forest wildlife. The urgentneed for research into ways of mitigating the problem ofdefaunation at a continental scale cannot be overempha-sised. Our study, although ambitious, should assist inquantifying the extent of the problem.

Acknowledgments

We thank J. G. Robinson, J. F. Oates, G. Cowlishaw, A.Noss, G. M. Mace, and an anonymous reviewer for com-ments and suggestions on the paper.

Literature Cited

Alvard, M., J. G. Robinson, K. H. Redford, and H. Kaplan 1997. The sus-tainability of subsistence hunting in the Neotropics. ConservationBiology 11:977–982.

Ape Alliance. 1998. The African bushmeat trade: a recipe for extinc-tion. Fauna and Flora International, Cambridge, United Kingdom.

Chapman, C. A., and D. A. Onderdonk. 1998. Forests without pri-mates: primate/plant codependency. American Journal of Primatol-ogy 45:127–141.

Cole, L. C. 1954. The population consequences of life history phenom-ena. Quarterly Review of Biology 29:103–137.

Cowlishaw, G. 1999. Predicting the pattern of decline of African pri-mate diversity: an extinction debt from historical deforestation.Conservation Biology 13:1183–1193.

DelVingt, W. 1997. La chasse villageoise: synthèse régionale desétudes réalisées durant la première phase du programme ECOFACau Cameroun, au Congo et en Republique Centrafricaine. Conser-vation et Utilisation Rationelle des Ecosystèmes Forestiers d’Af-rique Centrale, Brussels.

Dirzo, R., and A. Miranda. 1991. Altered patterns of herbivory and di-versity in the forest understorey: a case study of the possible conse-quences of contemporary defaunation. Pages 273–287 in P. W.Price, G. W. Lewinsohn, G. Fernandes, and W. W. Benson, editors.Plant-animal interactions: evolutionary ecology in tropical and tem-perate regions. Wiley, New York.

Fa, J. E., and C. A. Peres. 2001. Game vertebrate extraction in Africanand Neotropical forests: an intercontinental comparison. Pages203–241 in J. D. Reynolds, G. M. Mace, J. G. Robinson, and K. H.Redford, editors. Conservation of exploited species. CambridgeUniversity Press, Cambridge, United Kingdom.

Fa, J. E., and A. Purvis. 1997. Body size, diet and population density inAfrotropical forest mammals: a comparison with Neotropical sites.Journal of Animal Ecology 66:98–112.

Hennemann, W. W. 1983. Relationship among body-mass, metabolic-rate and the intrinsic rate of natural increase in mammals. Oecolo-gia 56:104–108.

Jerozolimski, A. 1998. Effect of subsistence hunting on the structure ofmammal assemblages in Neotropical forests. M.S. thesis. Universityof East Anglia, Norwich, United Kingdom.

Musters, C. J. M., H. J. de Graaf, and W. J. ter Keurs. 2000. Can pro-tected areas be expanded in Africa? Science 287:1759–1760.

Noss, A. J. 1995. Duikers, cables and nets: a cultural ecology of hunting ina central African forest. Ph.D. thesis. University of Florida, Gainesville.

Novaro, A. J., K. H. Redford, and R. E. Bodmer. 2000. Effect of hunting insource-sink systems in the Neotropics. Conservation Biology 14:713–721.

Figure 1. Log total extraction ( E; kg/year) versus log total production ( P; kg/year) of bushmeat taxa in the Congo (open symbols) and Amazon (solid symbols) basins. Solid line, extraction equals production; dashed line, exploitation levels at 20% of production considered sustainable for long-lived taxa (Robinson & Redford 1991).

Conservation BiologyVolume 16, No. 1, February 2002

Fa et al. Bushmeat Exploitation in Tropical Forests 237

Peres, C. A. 2000. Effects of subsistence hunting on vertebrate communitystructure in Amazonian forests. Conservation Biology 14:240–253.

Peres, C. A., and J. Terborgh. 1995. Amazonian nature-reserves: ananalysis of the defensibility status of existing conservation unitsand design criteria for the future. Conservation Biology 9:34–46.

Redford, K. 1992. The empty forest. BioScience 42:412–422.Redford, K. H., and J. G. Robinson. 1987. The game of choice: patterns

of Indian and colonist hunting in the Neotropics. American Anthro-pologist 89:650–667.

Robinson, J. G., and E. L. Bennett, editors. 2000. Hunting for sustain-ability in tropical forests. Columbia University Press, Columbia.

Robinson, J. G., and R. E. Bodmer. 1999. Towards wildlife manage-ment in tropical forests. Journal of Wildlife Management 63: 1–13.

Robinson, J. G., and K. H. Redford. 1986a. Intrinsic rate of natural in-crease in Neotropical forest mammals: relationship to phylogenyand diet. Oecologia 68:516–520.

Robinson, J. G., and K. H. Redford. 1986b. Body size, diet and popula-

tion density of Neotropical forest mammals. The American Natural-ist 128:665–680.

Robinson, J. G., and K. H. Redford. 1991. Sustainable harvest of Neo-tropical forest animals. Pages 415–429 in J. G. Robinson and K. H.Redford, editors. Neotropical wildlife use and conservation. Chi-cago University Press, Chicago.

Robinson, J. G., K. H. Redford, and E. L. Bennett. 1999. Wildlife har-vest in logged tropical forests. Science 284:595–596.

Wilkie, D. S., and J. F. Carpenter. 1999. Bushmeat hunting in theCongo Basin: an assessment of impacts and options for mitigation.Biodiversity and Conservation 8:927–955.

World Society for the Protection of Animals. 1995. Slaughter of theapes. World Society for the Protection of Animals, London.

Wright, S. J., H. Zeballos, I. Dominguez, M. M. Gallardo, M. C. Moreno,and R. Ibáñez. 2000. Poachers alter mammal abundance, seed dis-persal and seed predation in a Neotropical forest. Conservation Bi-ology 14:227–239.