Agroecosystems and Biodiversity
Alison G. Power and Alexander S. Flecker
Section of Ecology and Systematics
Ithaca, New York USA
Until recently, efforts to preserve biodiversity have focused on natural ecosystems, despite the fact that these areas make up only about 5% of the terrestrial environment (Western and Pearl 1989). In contrast, approximately 50% of land is currently under agricultural production and 20% is in commercial forestry (Western and Pearl 1989). Given this pattern, there is increasing recognition that most species interact with agricultural systems, even if their primary habitat is in natural areas. Moreover, a large proportion of the total species of a region are likely to be found in agroecosystems (Pimentel et al. 1992). The management of these agricultural systems can dramatically affect overall levels of biodiversity, as well as the success of particular species.
Agriculture has a significant impact on biodiversity, through a variety of mechanisms: as a political and economic instrument, by means of commodity prices or subsidies; as a production technology using pesticides, fertilizer, and soil disturbance; and as a biological process resulting in habitat fragmentation and species invasions (Carroll 1990). The expansion of agriculture has transformed landscapes into mosaics of managed and unmanaged ecosystems, resulting in habitat loss and fragmentation for many species of flora and fauna. Modern commercial agriculture is dominated by monoculture, and this reduced plant diversity influences the composition and abundance of the associated biota, such as wildlife, pollinators, insect pests, their natural enemies, soil invertebrates, and microorganisms (Matson et al. 1997). Because a less diverse resource base is available, low genetic and species diversity of the crop results in less diversity at higher trophic levels, such as herbivores and predators. In addition, the widespread use of genetically uniform, improved crop varieties can lead to erosion of the genetic diversity inherent in traditional cultivars and land races.
Genetically uniform monocultures are often more vulnerable to pests and diseases and therefore require higher inputs of pesticides. These pesticides kill and injure a variety of nontarget organisms, such as wildlife, pollinators, natural enemies, and decomposer organisms. Monocultural, high-yield production systems also require high rates of nutrient addition, typically through the use of chemical fertilizers. These chemicals can have significant effects on the highly diverse community of soil microorganisms and invertebrates that regulate nutrient cycling in ecosystems (Matson et al. 1997). Through drift and runoff, the impact of agrochemicals such as pesticides and fertilizers can extend far beyond the farm, affecting biological communities in distant freshwater and marine ecosystems.
The loss of biodiversity has a range of negative ecological and societal consequences. More immediately, loss of biodiversity can have significant impacts on ecosystem function within agroecosystems and economic returns from the cropping system. Thus the conservation of biodiversity provides a number of benefits to agriculture. Uncultivated species, including wild relatives of crops, are an important source of germplasm for developing new crops and cultivars. Natural areas adjacent to agricultural systems can provide habitat for pollinators and natural enemies of pests. Within the agroecosystem itself, increasing crop diversity through the use of polycultures can augment the resources available to pollinators and to pest natural enemies such as parasitic wasps, resulting in higher populations of these beneficial organisms (Andow 1991). Minimizing the use of agrochemicals can also result in the conservation of beneficial organisms and the conservation of functional processes such as decomposition and nutrient cycling. Thus, the conservation of biodiversity within the agroecosystem affects plant and soil processes that can, in turn, affect crop productivity (Matson et al. 1997).
Diversity in Agroecosystems
Agroecosystems vary substantially in the attributes that affect biodiversity. Management of agroecosystems for high productivity results often results in low plant species richness, since only species with relatively high productivity are selected. Traditional agricultural systems tend to be significantly more diverse than conventional, commercial cropping systems, although they seldom approach the diversity of surrounding natural systems. However, the genetic diversity of traditional crops is often strikingly high compared to commercial agricultural systems in both the tropics and the temperate zone. Many contemporary crop varieties have been selected for high rates of carbon allocation to particular plant tissues such as seeds or fruits in lieu of allocation to defensive compounds, while in traditional varieties selection for high productivity is likely to be constrained by selection for other useful attributes, such as pest resistance or drought resistance. These varieties are likely to exhibit greater resistance to natural perturbations such as pests or drought, accompanied by lower productivity.
While annual cropping systems are likely to have low species richness, some traditional systems common to the tropics include a remarkable diversity of plant species. In such systems, species are not selected exclusively for high productivity. Compared to annual cropping systems, these traditional systems of high plant diversity also tend to have higher faunal diversity. Such characteristics make these traditional systems more similar to undisturbed ecosystems in their functional attributes -- for example, they probably have higher functional redundancy than is found in most agricultural or plantation systems of low diversity (Power and Flecker 1996). Compared to highly simplified systems such as monocultures, managed systems with high plant diversity appear to be better buffered against perturbations such as drought or pest epidemics. Examples of relatively diverse traditional agroecosystems include shifting cultivation, traditional rainfed rice systems, home gardens, and traditional shade coffee and cacao systems.
One common traditional agricultural system is shifting cultivation (also known as swidden or slash-and-burn agriculture). In shifting cultivation systems, temporary forest clearings are planted for a few years with annual or short-lived annual crops and then allowed to remain fallow for a period longer than the cultivation period (NRC 1993). While swidden plots can be quite diverse, plant species richness probably rarely approaches the richness of the surrounding forest. Although direct comparisons of the species richness of swiddens and forest are rare, the few existing studies indicate lower species richness in the swiddens compared to the forest. For example, Jivaroan swidden plots had substantially lower species richness than nearby tropical forest in the Peruvian Amazon (Boster 1983). Swidden plots were heavily dominated by a single crop, manioc (Manihot esculenta), the diversity of manioc cultivars was extremely high. In some swidden systems, therefore, genetic diversity within staple crops may play a bigger role than species diversity.
Despite low species richness in the early stages of swidden, over the entire cycle the swidden system as a whole may be quite diverse. For example, Christanty et al. (1986) reported that the swidden kebun-talun system of Java contained 112 plant species, largely because of a long period of perennial production in a managed fallow. In many swidden systems, species richness increases dramatically from the initial stage dominated by annual crops. Hart (1980) and Ewel (1986) have suggested that such systems may be designed as analogs of natural forest systems, in that they tend to mimic successional stages of the forest in structure and presumably in function.
Obviously, swidden cultivation systems have the potential to have a profound impact on biodiversity of flora and fauna in tropical forests, depending on the intensity of the disturbance and the size and distribution of the clearings. Despite lower species richness within the initial stages of the swidden, the impact of traditional swidden systems was probably limited since clearings were small (<1 ha) and distributed at low densities throughout a vast matrix of undisturbed forest. Furthermore, many useful trees were left in the system and the soil was seldom turned over at planting, so the intensity of disturbance was relatively low. In swidden systems on Seram, Indonesia, moth diversity was low in the recently cleared and cultivated areas, higher in the young fallows, and nearly equivalent to undisturbed forest in the older fallows of regenerating forest (Holloway 1991). The representation of moth families was similar to that in undisturbed forest in all phases of the swidden cycle.
Studies of biodiversity in shifting cultivation systems thus suggest that these systems can support biodiversity. However, as the human population which depends on shifting cultivation has grown, whether through population growth per se, through economic policies that marginalize the rural poor, or through government relocation programs, the impacts of shifting cultivation have clearly intensified. In much of the humid tropics, fallows have shortened and the extent of clearing has increased dramatically. These trends have dismaying implications for the continued agricultural productivity of shifting cultivation systems and for the maintenance of biodiversity in the landscape as a whole.
Traditional Rice Systems
Paddy rice systems are notable in having extremely low crop diversity, yet these systems are quite diverse in two respects. First, like many traditional systems, traditional paddy rice systems are characterized by extreme genetic diversity in the form of hundreds of traditional rice varieties many of which are still utilized by small farmers. This diversity undoubtedly protects these systems from the ravages of pests and pathogens (Roger et al. 1991), and therefore reduces the need for pesticides. Second, traditional rice paddies contain a striking number of aquatic species. The tropics contain some of the most extensive freshwater ecosystems in the world, and these systems interact continuously with natural areas and agroecosystems such as rice paddies (Allan and Flecker 1993).
Heckman (1979) carried out a remarkably thorough survey of biodiversity in a traditional rice field in northeastern Thailand. During his year-long study, Heckman found some 590 species in a single field including 166 species of algae, 83 species of ciliates, 146 arthropod species, 18 fish species, and 10 species of reptiles and amphibians. Although we consider rice paddies to be monocultures, the diversity of aquatic plants was quite high, including 10 species of dicots and 25 species of monocots. In a comparison of biodiversity from Heckman's traditional rice field to other conventional rice fields, Roger et al. (1991) found that biodiversity has decreased in rice fields since the mid-1970s, due to crop intensification and the widespread application of pesticides.
Traditional rice systems represent sufficiently complex ecosystems that, historically, a variety of organisms were harvested from them. These included an assortment of plants, snails, crabs, prawns, water bugs, fish, amphibians, and reptiles. Some floodplain regions in Asia and Africa have long been used for fish-rice polyculture (Coche 1967, Huat and Tan 1990, Catling 1992). Until relatively recently, much rice-fish polyculture was based on collecting wild fishes (and crustaceans) that naturally ventured into rice fields during inundation. At the end of the rice growing season, fishes that populated and reproduced were captured prior to rice harvest. In general, the captural system of fish-rice polyculture involved very low inputs towards fish production (Huat and Tan 1990). Until the advent of widespread use of agrochemicals in rice paddies, capture fisheries associated with rice fields were an important source of protein and income. For example, during the 1950's the captural system employed in some four million ha of irrigated rice fields in Indonesia, out of a total of 4.5 million ha under rice production (Ardiwinata 1957; cited in Huat and Tan 1990). Thus traditional rice systems support significant biodiversity, some of which is exploited by the human communities that manage the rice fields.
When highly managed, mixed gardens of trees, shrubs and herbaceous species are located close to the house, they are known as home gardens, kitchen gardens or dooryard gardens. These systems typically provide food, fodder, firewood, medicines, ornamental plants and construction materials, as well as domestic animals, to the household. Home garden systems are intensively managed and exhibit high taxonomic and structural diversity. Species richness in these systems ranges from relatively modest levels to very high diversity. In Mexico, for example, home gardens have been reported to contain from 33 to 55 species per garden (Allison 1983) up to a total of 404 species in the home gardens of a single community (Herrera Castro 1991). One study of Costa Rican home gardens found 56-83 species per garden (Gliessman 1990). Brownrigg (1985) reports the use of over 600 plant species in Indonesian home gardens, although no individual garden would contain this many species.
Shaded Coffee and Cacao Systems
In recent years, it has become clear that traditional coffee and cacao plantations can also contain high diversity of flora and fauna (Perfecto et al. 1996). Coffee and cacao are traditionally grown under a canopy of shade trees that may be remnants of the original forest or may have been deliberately planted. Such systems exhibit a high degree of habitat heterogeneity , and they appear to serve as good surrogates of natural forest for many faunal species (Terborgh 1989, Perfecto et al. 1996). These systems are long-lived, remaining productive for many decades. In contrast to these traditional systems, studies have shown substantially lower diversity of ants and birds in coffee grown without shade trees, i.e., in "sun" coffee (Perfecto et al. 1996).
Despite increasing interest in the potential of traditional cacao systems for supporting biodiversity, there has been much less research on biodiversity in cacao than in coffee. In the absence of more data on cacao, we can use the information gained through biological studies in coffee to develop hypotheses about the potential role of traditional cacao plantations in biodiversity conservation (Greenberg, this proceedings). Below we describe on-going studies designed to examine this issue in cacao plantations in the Dominican Republic.
Biodiversity in Dominican Cacao Plantations: A Case Study
In a study of biodiversity in northeastern Dominican Republic, we were interested in determining the extent to which agricultural systems may support faunal biodiversity (Power and Flecker, unpublished data). We compared biodiversity in patches of undisturbed forest and three of the most common agricultural systems in the region: traditional shaded cacao plantations, African oil palm plantations, and pastures. A minimum of two replicate sites for each system was monitored each season. We selected birds, lizards, and ants as indicators of biodiversity, as organisms that are relatively common, easily detected, restricted to one or a few habitats, and sensitive to habitat disturbance. In this study, birds and arthropods were monitored in two wet seasons and two dry seasons, while lizards were monitored during one wet season and two dry seasons. Birds were monitored through point counts, lizards were surveyed using transects, and ants were sampled using pitfall traps.
Both birds (Stotz et al. 1996) and ants (Andersen 1990) have been used to assess quickly and accurately the ecological characteristics of many terrestrial communities. Birds can be particularly good biological indicators, due to their conspicuous behavior, rapid and reliable identification, ease of sampling, stable taxonomy, diversity and ecological specialization, and high sensitivity to disturbance (Stotz et al. 1996). Ants are excellent bio-indicators because they are ubiquitously diverse and abundant, they are functionally important at all trophic levels, they are highly sensitive to environmental variables, and they respond rapidly to environmental change (Andersen 1990). Ants are both numerically and ecologically dominant in tropical forest systems, where the non-overlapping territories of dominant ants form a three-dimensional mosaic that significantly impacts the distribution and abundance of other arthropods (Majer 1993) . This "ant mosaic" significantly influences the diversity and species composition of neotropical forest biota, including plants and vertebrates as well as invertebrates (Gilbert 1980). In agroecosystems, the ant community has the potential to play a highly significant role in controlling herbivorous insects.
In comparisons between agricultural systems, we found relatively high species diversity for all taxa in cacao plantations, although response to other agroecosystems varied among the three taxa discussed here (Power and Flecker, unpublished data). Birds were nearly as diverse in cacao as in undisturbed forest, and more abundant in cacao than in any other habitat (Fig. 1). Oil palm plantations were remarkably depauperate of birds, with pastures intermediate in terms of both diversity and abundance of birds. Overlap among bird species was higher between cacao and forest than between cacao and other agroecosystems. Robbins et al. (1992) also found that cacao supported particularly high species richness and abundance of birds, particularly neotropical migrants. Some migrants, such as Magnolia Warblers, were more abundant in groves of cacao, citrus and mango than in natural forest (Robbins et al. 1992). Despite large numbers of birds in agricultural systems, these habitats were inadequate habitats for many resident birds and some migrant species.
In our study, lizards were exceptionally diverse and abundant in cacao plantations, showing significantly higher diversity in cacao than in undisturbed forest (Fig. 2). Lizard diversity and abundance was quite low in pastures, and intermediate in oil palm plantations. Similar to birds, the overlap of lizard species was highest between cacao and forest. Heinen (1992) compared the herpetofauna in old abandoned (25 yrs) and recently abandoned (5 yrs) cacao plantations, to primary forest at La Selva, Costa Rica. In contrast to our results, Heinen (1992) found that species diversity was higher in primary forest than in recently abandoned cacao plantations. However, abundance and biomass of reptile and amphibians were greatest in the recently abandoned site, and lowest in the primary forest. Heinen suggested that primary forest may be an important refuge for rare species and a source pool for colonization into disturbed sites. Despite the high species richness of lizards in Dominican cacao plantations, we also found that some rare species (e.g., Sphaerodactylus cochranae) were restricted to undisturbed natural ecosystems and never encountered in cacao plantations. In general, high herpetofauna abundance in cacao may be due to greater densities of arthropod prey, supported by the high litter fall of cacao Heinen (1992).
In our study, ground-foraging ants were extremely diverse in both cacao and pastures, and significantly more abundant in pastures than in any other habitat (Fig. 3). For ants, species overlap was high among forest, cacao and oil palm and relatively low between pasture and any other habitat. Torres (1984) reported complete overlap of the ant species in humid tropical forest and traditional coffee plantations in Puerto Rico, but greater species richness in grassland and agricultural land than in forest or coffee.
In general, ants tend to dominate the abundance and species richness of the arthropod fauna of many tropical ecosystems, including cacao plantations. Ants constitute 10-33% of the arboreal arthropod biomass in Brazilian cacao plantations (Majer et al. 1994) and up to 70% of the biomass (Majer 1976) and 89% of the total insect numbers (Leston 1973) in Ghanaian cacao plantations. Arboreal sampling indicates that ant species richness is also high in cacao plantations, with 67 species found in Ghanaian cacao (Room 1971) and 88 species in Papua New Guinea (Room 1975b). Estimates of species richness increase substantially when ground-foraging ants are included in the samples, rising to 128 species in Ghanaian cacao plantations (Room 1971). Since our data include only ground-foraging ants, this is undoubtedly an underestimate of the species richness of Dominican cacao plantations.
Despite the high species richness of ants in cacao in many studies, comparisons with primary forest usually show higher species richness in forest, which we did not find. Roth et al. (1994) reported significantly lower diversity of ground-foraging ants in monocultural banana plantations and in active cacao plantations than in either undisturbed humid forest or cacao plantations that had been abandoned for 25 years. Among the active cacao plantations, the plantation that included the most diverse shade tree assemblage also had the highest ant diversity. There was high species overlap between undisturbed forest and abandoned cacao plantations, but relatively little overlap with ants in either of the active agricultural systems, banana or cacao (Roth et al. 1994). Studies of ground-foraging ants in seven habitats in Papua New Guinea indicated that ant species richness was highest in primary forest and in 15 to 25 year old rubber plantations with a diverse understory of saplings and herbs (Room 1975a). Species richness was somewhat lower in cacao and coffee plantations, which contained a variable diversity of shade trees and understory herbs, and lower still in grasslands with low plant species diversity. Lowest ant species richness occurred in oil palm plantations, which were effectively monocultures since understory herbs were weeded frequently. This is consistent with our results indicating higher diversity in pasture than in oil palm. Moreover, as in our studies, Room (1975a) found little species overlap between forest systems and grassland.
Within agricultural systems, management practices may significantly affect ant diversity, as demonstrated by Perfecto and Snelling (1995) in their study of ants in five types of coffee production systems in Costa Rica, ranging from traditional, low-input coffee plantations with a high diversity of shade trees and some annual crops, to intensive, high-input monocultural coffee systems containing no shade trees or annual crops. In this study, species diversity of ground-foraging ants was correlated with the amount of shade in the five types of coffee systems, which ranged from 99% shade in traditional coffee systems to 32% shade in intensively managed coffee monocultures. The amount of shade in the different systems was also correlated with the gradient of vegetational and structural diversity. Systems with many shade trees appear to provide a more diverse resource base for insect prey, including a diverse layer of decaying wood and leaf litter, as well as more moderate temperature and moisture regimes than systems with little shade. Similarly, cacao systems with many shade trees are also likely to provide a broad range of resources and to display relatively high levels of biodiversity.
These studies suggest that traditional cacao plantations, like traditional coffee plantations, may be reasonably effective forest surrogates for some faunal groups (e.g., Terborgh 1989, Robbins et al. 1992, Perfecto et al. 1996). The extensive habitat heterogeneity in traditional cacao systems is probably one of the major reasons why they serve as good surrogates of natural forest compared to oil palms and pasture (Terborgh 1989). In this area of the Dominican Republic, cacao is grown under cover of various species of large shade trees, and the understory often contains a diverse assemblage of herbaceous plants. Factors such as humidity, light, and ground cover in cacao groves make these systems much more similar to a natural forest than are pastures and oil palm plantations. In fact, from aerial photographs of this region, it is hard to distinguish cacao groves from natural forest. One of the limitations on cacao groves serving as reservoirs of biodiversity is that they may not represent large enough forest tracts to support some vertebrate species. However, this is a problem associated with fragment size rather than the suitability of cacao per se.
Lessons Learned from Traditional Agroecosystems
Research on biodiversity in traditional agroecosystems and comparisons between traditional and conventional agroecosystems have led to a number of clear guidelines for designing agricultural systems that support high levels of biodiversity. For example, we know that:
? Higher diversity (genetic, taxonomic, structural, resource) within the cropping system leads to higher diversity in associated biota.
? Lower use of pesticides leads to higher diversity in associated biota.
? Increased biodiversity leads to more effective pest control and pollination.
?Increased biodiversity leads to tighter nutrient recycling.
As we accumulate more information about the specific relationship between biodiversity conservation and productivity in cacao production systems, these guidelines can be used to improve the sustainability and conservation value of cacao systems.
Allan, J. D. and A. S. Flecker. 1993. Biodiversity conservation in running waters. BioScience 43: 32-43.
Allison, J. 1983. An ecological analysis of home gardens (huertos familiares) in two Mexican villages. M.A. Thesis, University of California, Santa Cruz.
Andersen, A. N. 1990. The use of ant communities to evaluate change in Australian terrestrial ecosystems: a review and a recipe. Proceedings of the Ecological Society of Australia 16: 347-357.
Andow, D. A. 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36: 561-586.
Ardiwinata, R. O. 1957. Fish culture on paddy fields in Indonesia. Proceedings of the Indo-Pacific Fisheries Council 7: 119-154.
Boster, J. 1983. A comparison of the diversity of Jivaroan gardens with that of the tropical forest. Human Ecology 11: 47-68.
Brownrigg, L. A. 1985. Home gardening in international development. League for International Food Education, United States Agency for International Development., Washington, D.C.
Carroll, C. R. 1990. The interface between natural areas and agroecosystems. Pages 365-383 in: Agroecology (C. R. Carroll, J. H. Vandermeer and P. Rosset, Ed.). McGraw-Hill Publishing Co., New York.
Catling, D. 1992. Rice in deep water. IRRI, Los Banos, Philippines.
Christanty, L., O.E. Abdoellah, G.G. Marten, and J. Iskandar. 1986. Traditional agroforestry in West Java: The pekarangan (homegarden) and kebun-talun (annual-perennial rotation) cropping systems. Pages 132-158 in: Traditional agriculture in Southeast Asia. A human ecology perspective (G.G. Marten, ed.). Westview Press, Boulder.
Coche, A. G. 1967. Fish culture in rice fields a world-wide synthesis. Hydrobiologia 30: 11-44.
Ewel, J. J. 1986. Designing agricultural ecosystems for the humid tropics. Annual Review of Ecology & Systematics 17: 245-271.
Gilbert, L. E. 1980. Food web organization and conservation of neotropical diversity. in: Conservation Biology. An Evolutionary-Ecological Perspective (M. E. Soule and B. A. Wilcox, Ed.). Sinauer Associates, Inc., Sunderland, Massachusetts.
Gliessman, S. R. 1990. Understanding the basis of sustainability for agriculture in the tropics: experiences in Latin America. Pages 378-390 in: Sustainable Agricultural Systems (C. A. Edwards, R. Lal, P. Madden, R. H. Miller and G. House, Ed.). Soil and Water Conservation Society, Ankeny, Iowa.
Hart, R. D. 1980. A natural ecosystem analog approach to the design of a successional crop system for tropical forest environments. Biotropica : 73-82.
Heckman, C. W. 1979. Rice field ecology in northeastern Thailand. Dr W. Junk bv Publishers, Boston. 228 pp.
Heinen, J.T. 1992. Comparisons of the leaf litter herpetofauna in abandoned cacao plantations and primary rain forest in Costa Rica: some implications for faunal restoration. Biotropica 24: 431-439.
Herrera Castro, N. 1991. Los huertos familiares Mayas en el oriente de Yucatan. M.S. Thesis, Universidad Nacional Autonoma de Mexico, Mexico, D.F.
Holloway, J. D. 1991. Aspects of the biogeography and ecology of the Seram moth fauna. Pages 37-62 in: The Natural History of Seram (I. Edwards and J. Proctor, eds.). Intercept, Andover, MD.
Huat, K. K. and E. S. P. Tan. 1990. Review of fish-rice culture in Southeast Asia. 24 pp. International Center for Living Aquatic Resources Management, Manila, Philippines.
Leston, D. 1973. The ant mosaic - tropical tree crops and the limiting of pests and diseases. Pest Abstracts and News Summaries 19: 311-341.
Majer, J. D. 1976. The ant mosaic in Ghana cocoa farms: further structural considerations. Journal of Applied Ecology 13: 145-156.
Majer, J. D. 1993. Comparison of the arboreal ant mosaic in Ghana, Brazil, Papua New Guinea and Australia: its structure and influence on arthropod diversity. Pages 136-141 in: Hymenoptera And Biodiversity (J. Lasalle and I. D. Gauld, eds.). C.A.B. International: Wallingford, England, UK; Tucson, Arizona, USA.
Majer, J. D., J. H. C. Delabie and M. R. B. Smith. 1994. Arboreal ant community patterns in Brazilian cocoa farms. Biotropica 26: 73-83.
Matson, P. A., W. J. Parton, A. G. Power, and M. J. Swift. 1997. Agricultural intensification and ecosystem properties. Science 277: 504-509.
National Research Council (NRC). 1993. Sustainable agriculture and the environment in the humid tropics. National Academy Press, Washington, D.C.
Perfecto, I., R. A. Rice, R. Greenberg, and M. E. van der Voort. 1996. Shade coffee: a disappearing refuge for biodiversity. BioScience 46: 598-608.
Perfecto, I. and R. Snelling. 1995. Biodiversity and the transformation of a tropical agroecosystem: ants in coffee plantations. Ecological Applications 5: 1084-1097.
Pimentel, D., U. Stachow, D. A. Takacs, H. W. Brubaker, A. R. Dumas, J. J. Meaney, J. A. S. O'Neil, D. E. Onsi and D. B. Corzilius. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience 42: 354-362.
Power, A. G. and A. S. Flecker. 1996. The role of biodiversity in tropical managed ecosystems. Pages 173-194 in Biodiversity and Ecosystem Processes in Tropical Forests (G. H. Orians, R. Dirzo, and J. H. Cushman, editors). Springer-Verlag, New York.
Robbins, C. S., B. A. Dowell, D. K. Dawson, J. A. Colon, R. Estrada, A. Sutton, R. Sutton and D. Weyer. 1992. Comparison of Neotropical migrant landbird populations wintering in tropical forest, isolated fragments, and agricultural habitats. Smithsonian Institution Press, Washington, D.C. 609 pp.
Roger, P. A., K. L. Heong and P. S. Teng. 1991. Biodiversity and sustainability of wetland rice production: role and potential of microorganisms and invertebrates. Pages 117-136 in: The biodiversity of microorganisms and invertebrates: its role in sustainable agriculture (D. L. Hawksworth, Ed.). CAB International, U.K.
Room, P. M. 1971. The relative distribution of ant species in Ghana's cocoa farms. Journal of Animal Ecology 40: 735-751.
Room, P. M. 1975a. Diversity and organization of the ground foraging ant faunas of forest, grassland and tree crops in Papua New Guinea. Australian Journal of Zoology 23: 71-89.
Room, P. M. 1975b. Relative distributions of ant species in cocoa plantations in Papua New Guinea. Journal of Applied Ecology 12: 47-61.
Roth, D. S., I. Perfecto and B. Rathcke. 1994. The effects of management systems on ground-foraging ant diversity in Costa Rica. Ecological Applications 4: 423-436.
Stotz, D. F., J. W. Fitzpatrick, T. A. Parker III, and D. K. Moskovits. 1996. Neotropical Birds. Ecology and Conservation. University of Chicago Press, Chicago.
Terborgh, J. 1989. Where have all the birds gone? Princeton University Press. 207 pp.
Torres, J. A. 1984. Diversity and distribution of ant communities in Puerto Rico. Biotropica 16: 296-303.
Western, D. and M. C. Pearl. 1989. Conservation for the twenty-first century. Oxford University Press, New York.