Shade Coffee: A Disappearing Refuge for Biodiversity Shade coffee plantations can contain as much biodiversity as forest habitats Ivette Perfecto, Robert A. Rice, Russell Greenberg, and Martha E. Van der Voort Within the expanding agri- cultural frontier in the trop- ics, one can find a variety of small, managed forest patches and traditional agricultural systems, which provide a refuge for forest- dwelling organisms. These managed habitats are frequently overlooked as potential areas of biodiversity con- servation (Pimentel et al. 1992). Fur- thermore, the conservation biology literature often refers to forest re- serves as islands in a sea of devasta- tion, in which the sea is formed by agriculture. Although chemically in- tensive monocultural systems may fit well with this perception of low Ivette Perfecto is an assistant professor in the School of Natural Resources and Environment, University of Michigan, Ann Arbor, MI 48109-1115. Her re- search interests include the transforma- tion of agriculture and associated changes in biodiversity. Robert A. Rice works on policy issues at the Smithsonian Migratory Bird Center, National Zoo- logical Park, Washington, DC 20008. His interests include land use changes in Latin America and the environmental consequences of agricultural modern- ization. Russell Greenberg is the direc- tor of the Smithsonian Migratory Bird Center, National Zoological Park, Wash- ington, DC 20008. His research inter- ests include the ecology of migratory birds in tropical ecosystems. Martha E. Van der Voort was a staff researcher at Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC 20008. She is currently studying nontimbef forest products at the De- partment of Wildlife, West Virginia University, P.O. Box 6125, Morgantown, WV 26505. The importance of shade coffee as a refuge for biodiversity may not be in the total land it involves, but in its location in areas that have been particularly hard hit by deforestation biodiversity, many other agro- ecosystems, especially in the tropics, are characterized by high vegetational diversity. One such agroecosystem is coffee, when managed with tradi- tional cultural practices. This popu- lar beverage, used worldwide for cen- turies, constitutes a major source of household income and foreign ex- change for many tropical countries, especially in Latin America. Coffee is traditionally grown un- der a canopy of shade trees. Because of the structural and floristic com- plexity of the shade trees, traditional coffee plantations have relatively high biodiversity. However, coffee plantations increasingly are being transformed into industrial planta- tions with little or no shade (Figure 1). The way that coffee production evolves in the coming decades is likely to have a tremendous impact on its ability to provide a refuge for tropi- cal biodiversity. In this article we discuss the role of shade coffee plan- tations in protecting biodiversity. We focus on northern Latin America, an area encompassing the Caribbean is- lands, Mexico, Central America, and the Andean countries of South America. However, many of the is- sues and conclusions discussed here also apply to coffee-exporting coun- tries throughout the tropics. The economic importance of coffee Coffee was introduced into the New World by the Dutch in 1723 (Wrigley 1988). During the twentieth century it has reached considerable impor- tance in the world market as an ex- port crop. Production has tripled in northern Latin America since World War II, and area under cultivation has nearly doubled (UNFAO Pro- duction Yearbooks). It is hard to overestimate the im- portance that coffee production and exportation has had for northern Latin America. More than 32% of the world's coffee comes from this region, where it is the leading source of foreign exchange. Although cof- fee is produced on only 7.4% of the total arable land, coffee lands at present take up approximately 44% of the area of permanent cropland (UNFAO Production Yearbook 1991). In northern Latin America, coffee plantations cover approxi- mately 2.7 million ha. This total includes roughly 700,000 ha in Mexico, 300,000 ha in the Carib- bean, 750,000 ha in Central 598 BioScience Vol. 46 No. 8 Figure 1. (Lefr) A sun plantation near San Jose, Costa Rica (photo by I. Pcrfecro). (Right) A traditional shade plantation near Tapachula. Chiapas (photo courtesy of M. Van der Voort). America, and 1,000,000 ha in Co- lombia. Coffee cultivation techniques Coffee cultivation systems fall along a continuum, ranging from the "tra- ditional" ro the "modern" (Table l).1 The modern system is character- ized by a reduction in shade, in- creased reliance on new high-yield- ing varieties, and an increase in chemical inputs, pruning, and coffee plant density (Coyner 1960,DeGraaf 19S6). The removal of shade in cof- fee farms helps establish a suite of characteristics of a coffee cultivation system aimed at incteasing yields, at least over the short run. However, with the loss of canopy cover, mod- ern plantations, also known as sun plantations, become more prone to water and soil runoff, threatening the long-term snstainability of the system (Rice 1990). One of the most striking features of the conversion from traditional to modern coffee cultivation is the ra- pidity with which it has occurred. After a largely unsuccessful attempt in the 1950s to modernize coffee growing using new strains and more >W* CM* the term modern here, a system also referred to as ;'?*atl?f^fifarrd technified.Tbk last term, although cumbersome, is used by development agencies and local institution; arid describes quire well thj technical and industrial approach TO what has heretofore been a traditional production sjrsrem. agrochemicals, modernization inten- sified in the 1970s. We estimate that almost half of the area in coffee pro- duction in northern Latin America had been converted by 1990. The speed and extent of conversion, how- ever, have been uneven. The percent- age of land converted in the region varies from as low as 15% in Mexico to more rhan 60% in Colombia (Fig- ure 2). Modernization was initially seen as a way of combating fungal dis- eases, particularly coffee leaf rust [Hemilcia vastatrix). The role that coffee leaf rust played in plantation modernization is significant, because the disease ranks as the most feared obstacle to production in most cof- fee- producing areas (Agrios 1982). Early on, coffee leaf rust provided the hook on which the entire coffee modernization process hung its hat. Phytopathological reasoning main- tained that less shade would allow moisture on coffee plants to dry more readily, therefore reducing fungal germination success. The arrival of rust in Brazil in 1970 and in Central America in 1976 brought to life the agronomic nightmares that had plagued coffee growers and govern- ments in the Old World for genera- tions. However, coffee leaf rust has not been as problematic as predicted, and the major motivation behind modernization has since become in- creased production. Modernized coffee represents a major departure in economic strat- egy for the coffee farmer. Siman found that modern farms out- produced semi-modern (a combina- tion that includes some shade reduc- tion, a change to new coffee varieties, and at least some use of agrochemi- cals) and traditional farms, with yields of 1397, 953, and 317 kg/ha, respectively.2 However, the levels of production had considerably differ- ent costs as well; in absolute terms, the cost (in US dollars) of production for a hectare of modern, semi-mod- ern, and traditional coffee was $1738.94, S1092.00, and $269.47, respectively. The cost to produce 1 kg of coffee was thus $1.24 for modern coffee, $1.14 for semi-modern cof- fee, and SO.85 for traditional coffee. Actual profits, of course, vary with world coffee price fluctuations. The traditional technology, with a much lower use of chemical inputs, repre- sents a passive production system in which the coffee unit receives little attention in the way of labor and/or capital. Traditional produc- tion devotes 2% of its expenditures to chemical inputs, whereas semi- modern and modern production spend 19% and 25% on chemical inputs, respectively. In addition, :J. J. Siman 1991, unpublished manuscript. Tropical Agronomic Center for Research and Teaching, Managua, Nicaragua, September 1996 599 Table 1. Distinguishing characteristics of traditional and modern coffee production technologies. Characteristic Traditional Modern Coffee variety Coffee height Shade cover Shade trees used Density of coffee plants Years to first harvest Plantation life span Agrochemical use Pruning of coffee Labor requirements Soil erosion Tipica, bourbon, maragogipe 3-5 m Moderate to heavy, 60%-90% coverage Tall (15-25 m), mixed forest trees, legumes, fruit trees, bananas 1000-2000/ha 4-6 30+ years None to low Individualized pruning or no pruning Seasonal for harvest or pruning Low Caturra, catuai, Colombia, Guarnica catimor 2-3 m None to moderate, up to 50% coverage Short (5-8 m), legumes; often monocultures 3000-10,000/ha 3-4 12-15 years ?< High, particularly fertilizers, herbicides, fungicides, nematocides Standardized stumping back after first or second year of full production Year-round maintenance with higher demands at harvest High (particularly on slopes) nonharvest labor accounts for the single largest cost in modernized sys- tems because it entails an array of intense cultivation practices such as standardized pruning, fertilization, and insecticide, fungicide, and nema- tocide applications to individual plants. Comparisons have also been made between modernized and organic cof- fee production by Akkerman and Van Baar (1992) and Boyce et al. (1994). They reported that despite lower total income, organic coffee production resulted in a significantly higher net revenue (approximately $350.00/ha), in part because of lower production costs. Furthermore, when externalities generated by environ- mental costs associated with coffee production (e.g., pesticides and/or soil erosion) were incorporated into the analysis, the differences in net revenue between organic and non- organic production increased (Boyce et al. 1994). .So homos Figure 2. A map of northern Latin America with proportional circles depicting relative coffee area by country (figures in thousands of hectares; based on UNFAO Production Yearbook 1990) and approximate area modernized (white portion of circle) and nonmodernized (shaded portion of circle). The modernized area does not include lands designated as "semi-modernized." Question marks indicate no data available for modernized coffee area. Data are from a variety of sources for the early 1990s, except area estimates from Mexico and Nicaragua, which are for the early 1980s (see Rice 1993 for details). Shade coffee and biodiversity conservation?an overview On a geographical scale, the impor- tance of shade coffee as a refuge for biodiversity may not be in the total land it involves, but in its location in areas that have been particularly hard hit by deforestation. In fact, the total land area planted in coffee is moder- ate compared with some other land uses (particularly pasture). However, plantations tend to be localized on relatively high-quality soils and in the mid-elevation (500-2000 m) eco- logical zone. Natural habitats in these zones, which include pine-oak wood- land and premontane tropical forest, are often highly fragmented and de- graded, and few reserves have been established to protect the many or- ganisms endemic to these habitats. The Pacific slope of the central Cor- dillera in Central America has been particularly devastated. In areas where deforestation is high and coffee is still produced on traditional shade plantations, these plantations are likely to be a critical refuge for the forest biota. In fact, coffee plantations may already have served as a critical refuge during a human-caused habitat bottleneck. Brash (1987) suggested that the rela- tively low rate of avian extinction experienced on Puerto Rico during recent periods of deforestation may be due in part to the presence of shade coffee plantations. Similarly, Nir (1988) argued that many rare orchids survived deforestation in Puerto Rico on shaded coffee farms. By the turn of the nineteenth cen- tury, 99% of the original forest cover 600 BioScience Vol. 46 No. 8 on Puerto Rico had been lost, with essentially no second-growth forest replacing it. However, shaded coffee plantations still covered 9% of the island. As the rural economy has been abandoned, forest is returning to much of the island, and the "seed" for its regrowth is often the aban- doned coffee estates (Weaver and Birdsey 1986). The plant diversity of coffee plan- tations results from two distinct pro- cesses. First, a small percentage of plantations ("rustic") are planted in forest cleared of its understory. Therefore, the diversity of the natu- ral forest canopy is preserved in a modified form. In addition, rustic plantations are often unweeded dur- ing periods of low coffee prices, which adds to the maintenance of biodiversity as well. Purata and Meave (1993) found that rustic plan- tations provide the only habitat for several forest trees in the mid-eleva- tion zone of Oaxaca, Mexico. The term rustic is also often conferred on highly diverse indigenous agro- forestry systems that incorporate coffee production. One well-studied example is the te'lom, or managed forests, of the Huastec Maya of Tamaulipas, Mexico. Te'lom forests contain more than 300 plant species and cover a quarter of the Huastec's agricultural land (Alcorn 1984). The incorporation of coffee into such a traditional agroforestry system is not surprising because cacao (chocolate) was cultivated in this manner for 2000 years before the Conquest (Bergmann 1969). In the second and more common process, the shade canopy is of sub- stantially lower diversity than in rus- tic plantations and is maintained through deliberate planting. The overstory species found in traditional coffee plantations vary from country to country and regionally within each country (see Fuentes-Flores 1979 for a classification for Mexican systems), as does the intensity with which shade trees are pruned. Nitrogen-fixing le- gumes such as Ittga spp., Erythrina spp., and Gliricidia sepium form an important component of many cof- fee farms. The latter two genera lose most or all of their leaves during the dry season, which renders planta- tions on which they are grown simi- lar to sun plantations for this period. Table 2. Number of species of beetles, ants, wasps, and spiders in the canopy of shade trees and coffee plants in different types of coffee farms, based on fogging with Pyrethrin-based insecticides.* Species Type of farm Beetles Ants Non-ant Hymenoptera Spiders Shade trees Erythrina Traditional 126 30 103 NA* poeppigiana Erythrina fusca Traditional 110 27 61 NA Annona sp. Traditional NA 10 63 NA E. poeppigiana Technified with shade 48 5 46 NA Coffee plants Coffea arabica* Traditional 39 14 34 44 Technified with 29 9 31 NA shade Technified without 29 8 30 29 shade *I. Perfecto, 1996, manuscript in preparation. ?Data not yet available. *Coffee based on ten plants per treatment. Throughout northern Latin America, it is common to find banana (Musa spp.), citrus (e.g., tangerine, orange, grapefruit, and lemons), or other fruit trees (e.g., avocados, mamey, man- goes, and zapotes) mixed in with the coffee, filling out the multistrata sys- tems in which coffee itself forms the shrub layer. Some fuel wood- and timber-producing trees, including Cedrela mexicana, Cordia alliodora, and Swietenia macrophylla, are also found in diverse coffee farms. Tradi- tional planted coffee farms com- monly have more than 40 tree spe- cies. Larger plantations tend to be less diverse, planted with one or a few species of native legumes, which are often heavily pruned. In many plantations, exotic trees are used, particularly Grevillea robusta, which grows well at higher eleva- tions and survives low temperatures. The high structural complexity of the traditional coffee plantation is a result of the various vegetative lay- ers in the agroecosystem. This struc- tural complexity offers living and nesting sites for a variety of organ- isms. In addition to increasing pri- mary structural diversity of foliage layers, the canopy of plantations can support secondary structures com- prised of epiphytes, parasites (e.g., mistletoes?Lorarithaceae), mosses, and lichens, which in turn support a community of arthropods, amphib- ians, and other creatures. The canopy also affects the microclimate of the coffee understory. Sun coffee planta- tions lack the protection provided by canopy trees from the impact of rain and wind, and they also lack the input of canopy leaf litter (Beer 1988). Therefore, even structurally equivalent layers of shade and sun plantations are dramatically differ- ent habitats. Finally, shade trees pro- vide a high diversity of food items for herbivores, frugivores, and nectarivores. Where there is a diver- sity of canopy species, differences in the timing of fruit and flower pro- duction are likely to reduce pheno- logical gaps (periods when no fruit or nectar resource for a particular taxa is available). Arthropod diversity Studies that have compared arthro- pod diversity in coffee plantations with that in forests have reported either similar or higher diversity in plantations. In Sulawesi, Stork and Brendell (1990) found the number of arthropod species in coffee planta- tions to be almost double that of mid-elevation forests. In a compara- tive study in Puerto Rico, Torres (1984) reported a more diverse ant fauna in a coffee plantation than in an upland tropical forest in the same region. Similar high ant diversity has been reported for cacao (Theobroma cacao) plantations, which are struc- turally similar to coffee plantations but are typically found at lower el- evations (Majer 1978, Room 1971, 1975). Studies of arthropod assemblages in the canopy of shaded plantations September 1996 601 attest to the high diversity of arthropods in these systems (Stork and Brendell 1990). The most dra- matic finding to date is that of Per- fecto et al.3 from a shaded plantation in Heredia, Costa Rica (Table 2). By fogging with pyrethrin-based insec- ticides, in a manner similar to some tropical forest studies (Erwin and Scott 1980, Stork and Brendell 1990), Perfecto and colleagues sampled the arthropods in the canopy of four shade trees and ten coffee bushes. Ants, other hymenopterans, beetles, and spiders were sorted into morpho- species. In the canopy of a single Erythrina poeppigiana they recorded 30 species of ants, 103 species of other hymenopterans, and 126 spe- cies of beetles. A second tree yielded 27 species of ants, 61 species of other hymenopterans, and 110 species of beetles. Although the two sampled trees were less than 200 m apart, the overlap of species was only 14% for beetles and 18% for ants. These pre- liminary results suggest that shaded plantations can have a local species diversity within the same order of magnitude as undisturbed forest. For example, Wilson (1987) reported 62 and 47 ant species from two trees each in upland rain forest in Peru, and Adis et al. (1984) reported 38 ant species in one Dipterix alata and two Eschweilera cf. odora in an up- land rain forest in Brazil. Birds and other vertebrates With the possibility of deforestation causing declines in several species of birds that migrate from North America to northern Latin America (Askins et al. 1990), many studies have focused on the status of over- wintering populations in different agricultural and natural habitats. Coffee plantations have often been singled out for their ability to sup- port numbers of forest migrants, those species most likely to be af- fected by conversion of forest to farm- land. Wunderle and Waide (1993) conducted a regional survey of the Greater Antilles and concluded that shade coffee plantations support high densities of certain species that de- 3I. Perfecto et al., 1995, unpublished mansucript. University of Michigan, Ann Arbor, MI. pend on closed canopy forest. A more detailed study in the Dominican Re- public supports this finding (Wun- derle and Latta in press). In addi- tion, Wunderle and Latta (1994) found that individuals of several migratory species in shade coffee plantations survived the winter at a rate comparable with those in natu- ral forest habitats. Greenberg et al. (1995) classified the migratory avi- fauna of eastern Chiapas, Mexico, as forest specialists, forest generalists, and scrub/open species, and then de- termined that shade coffee planta- tions support a high number of species of forest migrants (both gen- eralists and specialists) compared with other habitats in the region, and often at higher densities than natural forests. Shade coffee plantations may also be an important dry season refuge, providing fruit and nectar for birds when insect populations are other- wise dwindling. Several of the com- monly planted shade trees are native species that produce flower crops favored by omnivorous birds. It is likely that the movements of latitu- dinal and altitudinal migrants (Vannini 1994) are timed to take advantage of the asynchrony of flower crops available in shade plan- tations. Greenberg et al. (in press) examined the seasonality of bird use of coffee plantations and other habi- tats in eastern Chiapas through the repeated censusing of transects. Of the 23 habitats censused in eastern Chiapas, only the two plantation types (rustic and planted with Inga) showed an overall significant increase in bird numbers over the winter. Both the number of individual birds and the number, of species nearly doubled, a pattern that held for resi- dent tropical species as well as mi- grants. This increase was specific to omnivorous species; insectivores showed stable or slightly declining populations through the winter. In the 7?g<3-dominated plantations, the increase consisted largely of flower- feeding species such as Baltimore orioles (Icterus galbula) and Tennes- see warblers (Vermivora peregrina). Shade coffee plantations are par- ticularly well represented by canopy omnivores and nectarivores. During the dry season in Chiapas, more than 45% of the individual birds were in these guilds (Greenberg et al. in press), a figure that is significantly higher than for other forest habitats in the region. Shade coffee also sup- ports a high concentration of nectarivores or partial nectarivores in the Dominican Republic (Wun- derle and Latta in press). The high concentration of euphonias (small tanagers that eat mistletoe berries) in taller, less-pruned plantations sug- gests that parasitic plants support additional diversity as well. In general, migratory birds seem to fare better than resident birds in shade plantations, perhaps because migratory birds have less stringent habitat requirements than those spe- cies committed to breeding in the region. Resident birds may be af- fected by a variety of local ecological and landscape factors. The small size of the average coffee plantation makes it susceptible to fragmenta- tion, whose effects are known to be severe in tropical areas (Lovejoy et al. 1986). Structural modifications that remove foraging and nest sites for some species probably account for the loss of many forest special- ists. In addition, larger birds, such as cracids, parrots, and raptors, may be susceptible to hunting pressures. Nevertheless, Aguillar-Ortiz (1982), Corredor (1989), and Greenberg et al. (in press) found that the species richness of birds in coffee plantations with a structurally and floristically diverse canopy compares well with other natural forest habi- tats with which many species are shared. Greenberg et al. (in press) also showed that diversity of birds in coffee plantations (and other forest types) is considerably higher than in other agricultural habitats. In part, the high diversity of shade planta- tions results from the number of edge and second-growth species that oc- cur along with a smaller number of true forest birds (Corredor 1989, Greenberg et al. in press). Greenberg et al. (in press) showed that in Chiapas, the avifaunal similarity between pine-oak woodland and planted and rustic coffee is high (75%-80%). Similarity with meso- philous forest, however, was low, and several of the more specialized species found commonly in premon- tane forest, such as the spectacled foliage-gleaner (Anabacerthia varie- 602 BioScience Vol. 46 No. 8 gaticeps), were never recorded in cof- fee plantations. The bird diversity of heavily pruned shade plantations dominated by a single canopy spe- cies, a common plantation type, was only two-thirds of that of the more forestlike coffee plantations.4 In par- ticular, forest frugivores, bark-glean- ers, and understory species were poorly represented in the more mono- specific and heavily trimmed shade plantations. Shade coffee plantations support a high diversity of other vertebrate groups as well as birds. Estrada et al. (1993) found that, compared with other agricultural habitats, a high diversity and abundance of bats use various shade plantations with di- verse canopies. However, diversity was considerably lower than in low- land tropical forest. As with birds, a large proportion of individuals and species were partially frugivorous and nectarivorous, feeding on the flow- ering and fruiting trees of the canopy. Estrada et al. (1993) argue that the mobility of bats (like some birds) allows them to forage over shade plantations and other forest patches scattered over a large area. A strong relationship between the presence of a structurally diverse canopy and a high diversity of small terrestrial and scansorial mammals was found by Gallina et al. (1992) in Veracruz, Mexico. In this study, species depen- dent on canopy trees made up more than half of the fauna of the planta- tions, and more than 40% of the species were omnivores that com- monly fed on fruit. A similar high abundance and di- versity, as well as proportion of omnivores, was found for nonflying mammals in the Las Tuxtlas region of Mexico (Estrada et al. 1993, 1994). This observation is not sur- prising, considering that many of the trees managed in coffee plantations produce fruit that is eaten not only by humans but also by other mam- mals. Gallina et al. (1992) also re- ported that in addition to the many omnivorous species, some more spe- cialized mammals, such as small cats and otters, can be found in shade 4R. Greenberg, P. Bichier, A. Cruz, and R. Reitsma, 1996, manuscript in review. Smithsonian Migratory Bird Center, Wash- ington, DC. plantations in Veracruz. Estrada et al. (1994) did not find such mam- mals in plantations in Las Tuxtlas but did regularly observe howler monkeys (Allouata palliata). And working in Guatemala, Seib (1986) reported that mixed-shade planta- tions can support up to 50% of the original forest snake fauna. Biodiversity and the impact of coffee conversion The few direct comparisons between sun and shade plantations focus on the ground or coffee strata, and gen- erally they show a decrease in diver- sity with the conversion from shade to sun types. Perfecto et al.5 (Table 2) showed that arthropod diversity was lower in the ground strata of mono- specific shade farm and shadeless coffee monoculture than in that of shaded canopy. Although Hanson (1991) reported a high hymenopteran diversity in coffee monoculture (80 species of parasitoids) in the same region of Costa Rica, species rich- ness was still lower than on tradi- tional plantations. Perfecto and Snelling (in press) surveyed ant spe- cies diversity using bait transects on 16 coffee farms and found a positive correlation between species diver- sity and vegetational complexity. Once again, the highest diversity was found in the traditional farm and the lowest in sun coffee plantations. Per- fecto and Vandermeer (1994) and Perfecto (1994) suggested that both direct (e.g., loss of nesting sites) and indirect (e.g., changes in competitive interactions) mechanisms are respon- sible for reductions within the ant community in the coffee monocul- ture. Working in Mexico, Nestel and colleagues (Nestel and Dickschen 1990, Nestel et al. 1993) reported a reduction in the diversity of ant and macrocoleopteran assemblages in sun coffee plantations as compared with shaded coffee plantations. Studies restricted to the coffee lay- ers are likely to greatly underesti- mate the difference in overall diver- sity between plantation types. The elimination of trees, with their foli- age, flower, fruit mesocarp, and extrafloral nectaries, results in a dra- 5See footnote 3. matic reduction in hymenopterans.6 Aside from the loss of food provided by the trees, the habitat structure becomes simplified through loss of canopy foliage layers, tree trunks, and associated epiphytes. Canopy trees provide a host of poorly known microhabitats. The effect of canopy loss is likely to be severe for trees with a specialized canopy fauna. For example, the high diversity of arbo- real beetles, with more than 100 spe- cies in a single tree, is undoubtedly lost in systems that lack canopy trees (Table 2). Preliminary studies7 sug- gest that a large percentage of the ants found in the canopy of shade trees are also exclusively arboreal. For example, in a shaded plantation in Costa Rica, an average of 72% of the ants were found exclusively in trees. The loss of the shade in coffee plantations also means the loss of resources for many species in the detritivore food chain, particularly saproxilic and leaf litter arthropods. Shaded plantations in Costa Rica produce between 5000 and 20,000 kg ha"1 ? yr1 of leaf litter and prun- ing residues (Beer 1988)?values that fall within the range for tropical for- ests (Vitousek 1984). Shaded planta- tions, particularly rustic ones, con- tain old and dead trees that provide habitats for a diverse saproxilic ar- thropod community. In Mexico, a single coffee and cacao plantation was reported to contain 78 families of saproxilic invertebrates, with 93 % belonging to the orders Coleoptera, Diptera, Hymenoptera, or Collem- bolla (Moron and Lopez-Mendez 1985). Although studies directly com- paring saproxilic communities be- tween shade and nonshaded planta- tions are apparently lacking, the reduction in decaying wood and leaf litter suggests that these assemblages are greatly reduced along with shade elimination. Changes in the coffee stratum it- self due to loss of microclimate buff- ering are profound. The conversion of coffee plantations invariably re- sults in an increase in the amount of solar radiation reaching the ground, with concomitant increases in tem- perature and wind speed, direct im- *See footnote 3. 7See footnote 3. September 1996 603 pact of precipitation, and a decrease in relative humidity (Beer 1987). In sun coffee plantations, fluctuations of both temperature and humidity become more extreme. Perfecto and Vandermeer8 demonstrated that by experimentally increasing shade, the diversity of ground-foraging ants in coffee plantations increased, at least partly as a result of changes in mi- croclimate. The modernization of coffee plan- tations frequently includes a sub- stantial increase in agrochemicals (De Graaf 1986). Insecticides are known to decrease biological diversity in agroecosystems (Jepson 1989). Fun- gicide applications are also more common in modern plantations than in traditional ones. Certain fungi- cides also are known to have insecti- cidal activity and can have a detri- mental impact on insect diversity (S otherton and Moreby 1988). More- over, it is likely that fungicide appli- cations adversely affect the decom- position of leaf litter in modern systems. The shade of traditional plantations reduces weed growth (Nestel and Altieri 1992), so intense herbicide applications are necessary to reduce the ground cover of forbs and grasses in sun plantations. The removal of shade has also been shown to disrupt the natural nitrogen cycle associated with litter decomposition and with the actions of nitrogen- fixing bacteria associated with com- monly planted leguminous shade trees, therefore requiring the addi- tion of chemical fertilizers (Babbar 1993). These compounds can pol- lute local water supplies. Even the most cursory observa- tion in sun plantations shows them to be almost devoid of birds. Borrero (1986) first noted the dramatic de- crease in bird diversity in planta- tions in Colombia. In part, birds respond to the same loss of food resources, structural complexity, and microclimate buffering that is re- sponsible for changes in arthropod assemblages. Most observers have noted that the high abundance and diversity of birds in coffee planta- tions isassociated primarily with the canopy trees. Not surprisingly, then, *I. Perfecto and J. Vandermeer, 1995, unpub- lished manuscript. University of Michigan, Ann Arbor, MI. Figure 3. Annual population indices of the Baltimore oriole from 1978 to 1994 based on 1351 Breeding Bird Survey routes throughout North America (cour- tesy of US National Biological Service). Wunderle and Latta (in press) found a reduction in overall diversity and a significant shift from forest to matorral (shrubby second growth) species when comparing monotypic Inga shade and sun plantations in the Dominican Republic. Greenberg et al.,' working in Gua- temala, determined that the density and diversity of birds in sun coffee plantations is approximately half that in traditional coffee plantations. Furthermore, sun coffee plantations support few individual birds and bird species than adjacent areas of matorral, and many common mator- ral species avoid sun plantations. This finding probably reflects the low density of arthropods associated with coffee plants and the high de- gree of weed control associated with sun plantations. Birds make relatively little use of coffee flowers or berries, and consequently the nectarivorous and frugivorous species so prevalent in shade coffee plantations largely disappear. Gallina et al. (1992) esti- mated that approximately half of the species diversity of nonflying mam- mals is lost due to coffee conversion. An even higher percentage of reptile and amphibian diversity appears to be lost (Seib.1986). Migratory birds often occur in a range of habitats, so the impact on them of coffee modernization is dif- ficult to assess. Three largely nectarivorous migratory songbird 'See footnote 4. species are probably the most spe- cialized migratory species on shaded plantations (although smaller popu- lations may occur in urban gardens): the Baltimore oriole and the Tennes- see warbler in Mesoamerica, and the Cape May warbler {Dendroica tigrina) in the Antilles.10 Data from the Breeding Bird Survey of the US National Biological Service indicates that all three species have experi- enced sharp and statistically signifi- cant population declines from 1980 to 1994, corresponding to the period of intense coffee modernization. The annual estimated declines are 2.2%, 5.7%, and 4.2% for the oriole, Ten- nessee warbler, and Cape May war- bler, respectively (Figure 3). The three species experienced significantly ex- panding populations during the pre- vious 16 years, which would also be predicted from the continuing in- crease of area under coffee cultiva- tion. Other factors may be respon- sible for population declines, for example, habitat fragmentation, parasitism, and long-term cycles in prey abundance (Askins et al. 1990). The two warblers, for example, are boreal forest "spruce budworm" spe- cies and may be tracking long-term caterpillar cycles on the breeding grounds. However, the sharp decline in orioles is less likely to be caused by these breeding season factors because this species is not known to respond to insect cycles, it breeds successfully in edge situations, and it is rarely parasitized by cowbirds. Predator-prey interactions and coffee pests Coffee, at least when grown in the Western Hemisphere, is well known for its lack of insect pests. Although many herbivores can potentially dam- age coffee plants, only a few are economically important (Le-Pelley 1973). Coffee's resistance to herbi- vores may lie in the fact that it is a chemically well-defended plant (Frischknecht et al. 1986), with young leaves containing high quan- tities of alkaloids. In addition, older leaves are tough. Furthermore, there may simply be no native species in 10J. Wunderle, 1995, personal communica- tion. Institute for Tropical Forestry, US For- est Service, Palmer, Puerto Rico. 604 BioScience Vol. 46 No. 8 Latin America that have evolved mechanisms to overcome coffee's defenses?a common phenomenon in plant introductions. In addition, it has been argued that the structurally complex and floristically diverse traditional cof- fee plantation supports a high den- sity and diversity of predators and parasitoids, which are ultimately re- sponsible for the reduced number of insect pests in traditional plantations (Ibarra-Nunez 1990). The few com- parative studies in coffee plantations support this assertion (Benitez and Perfecto 1990, Nestel and Dickschen 1990, Perfecto and Snelling in press, Perfecto and Vandermeer 1994). For example, of the arthropod taxa sampled by Perfecto et al.,11 ants, other hymenopterans, and spiders were all more diverse in shaded plan- tations than in sun plantations. Robinson and Robinson (1974) esti- mated spider abundance in shade plantations in Papua New Guinea and suggested that the spiders have considerable insecticidal effects. Web-building spiders, for instance, consume 40 million insects per hect- are per year (Robinson and Robinson 1974). Perfecto etal.12 reported 34% more spiders in the coffee bushes in a traditional plantation than in a coffee monoculture. Ants, which show high diversity in traditional plantations, are effective predators as well (Carroll and Risch 1989). These observations suggest that the elimination of shade may ultimately result in increased pest problems as well. Diversity and economic risk reduction Biological diversity can provide im- portant economic returns to coffee growers. Because of the larger num- ber of products derived, the diverse plant community within a traditional farm fits much better into the risk- averse mentality of many small farm- ers (Reeves and Lilieholm 1993). Although much coffee is grown on a relatively small number of large es- tates, in most coffee-producing coun- tries the average coffee plantation is small. The size distribution of hold- 11 See footnote 3. '-See footnote 3. ings varies considerably from coun- try to country. For example, Mexico is dominated by small holdings on private and ejido land?91% of the holdings are less than 5 ha. In Co- lombia this value is only 49%, and 5% of the holdings are greater than 100 ha. For small farmers, committing oneself to total dependence on coffee puts one at great risk, not only with the vagaries of local weather and pest outbreaks, but with the often dramatic and unpredictable fluctua- tions of the global market. However, the traditional coffee farm sustains the grower beyond simply generat- ing an income at harvest time be- cause of the noncoffee products as- sociated with the shade treees. For example, overstory species provide fruits, fuelwood, and construction materials for household consump- tion, as well as a potential source of income derived from the local mar- ket. Honey production is a common rural industry in northern Latin America. Chazaro (1982) found at least 90 species of bee-pollinated plants in the shade coffee planta- tions near Xalapa, Veracruz. Wood from natural and human prunings provides a steady supply of fuel and, in the case of larger shade species, construction materials for the home and household furnishings. Selling surplus wood also brings in added income. Where precious hardwoods are mixed within the shade trees, single trees can be sold to local saw- mills or other buyers when times are tough economically. In situations in which a more managed coffee plan- tation system is possible, Somarriba (1990,1992) has shown that timber production and harvesting based on C. alliadora as a shade species can occur. Production of various fruit provides a household with a con- tinuous supply of nutritious prod- ucts for consumption and for the local market. A single, well-tended avocado tree, for instance, can yield between 2000 and 3000 fruits per season. Aside from household use, such produce can fetch $0.18 per fruit. For the two months of the avocado harvest, a producer can gain as much as $360." This single tree's 13R. Rice, field notes and personal observa- tions. harvest represents an equivalent of 100 work days at minimum farm wage. A shaded coffee farm displays two distinct types of biodiversity, man- aged and natural. The choices made by growers to use a variety of shade species yields an array of useful prod- ucts and at the same time provides cover with varying degrees of struc- tural diversity. This intentionally managed biodiversity by growers al- lows for higher levels of natural biodiversity as well in these shaded systems than is found in sun coffee. The issue now becomes how best to make use of the relationship between managed and natural biodiversity. Promoting biodiversity on coffee farms Defining an environmentally friendly coffee. Research to address this issue is in its infancy; however, we believe that the broad aspects of a biologi- cally diverse coffee farm can be out- lined. Clearly the presence of a shade canopy is essential. Furthermore, the greater the structural and floristic diversity of this canopy, the greater the likelihood that resources will be provided for a greater array of or- ganisms. A greater variety of animal- pollinated and -dispersed plants will support the diverse guild of omnivo- rous species that populate traditional shade plantations throughout the year. The canopy needs to provide sufficient coverage throughout the year to buffer the microclimate of the understory from rain and desic- cating winds. Tree species selection and pruning practices should have minimal impact on the epiphytic plants, mosses, and lichens as well as on dead trunks and limbs that pro- vide homes for so many canopy spe- cies. On-farm presence of shrub veg- etation along arroyos or on steep slopes will protect streams from ero- sion and provide an additional ha- ven for understory species unable to cope with a coffee monoculture. Sun drying, or using more energy-effi- cient technology, will reduce the need for harvesting trees to provide fuel for coffee dryers used in bean pro- cessing. Finally, reduced or no use of pesticides, herbicides, and chemical fertilizers should be promoted. The September 1996 605 use of natural or managed mulch will additionally foster a rich soil flora and fauna. Clearly, some of these recommendations, such as maintenance of a diverse shade struc- ture and protection of epiphytes, are inconsistent with what are consid- ered to be the most productive or expedient agronomic practices (Beer 1987, Boyce et al. 1994). The chal- lenge will be to develop cultivation systems that are a workable compro- mise between what is good for the farmer and what will truly benefit biological diversity. Foreign assistance. Because coffee plantations are managed primarily for export commodity production, the motivating force behind their existence is a powerful international market. Therefore, their continued ecologically sustainable management is likely to require the use of nontra- ditional policy tools. One possible approach is to influence the institu- tional programs providing assistance for rural development in the region. A principal institutional link for the projects involving modernization of the coffee sector of the region is the US Agency for International Devel- opment (USAID), although for some countries, such as Colombia and Mexico, USAID has played no role. (Influential national institutions in these two countries precluded any US involvement in their coffee sec- tors.) Working sometimes through its own (now-defunct) Regional Of- fice on Central America and Panama and sometimes with regional institu- tions like the Inter-American Insti- tute for Cooperation on Agriculture (IICA) of the Tropical Agronomic Center for Research and Teaching, both located in Costa Rica, USAID has played a major role in promoting modernized coffee. Since 1978, at least eight projects totaling US $81 million have targeted small coffee producers for modernization through reduced shade, high-yielding variet- ies, and increased chemical applica- tion (Rice and Ward 1996). Coffee modernization continues in at least three USAID-sponsored projects in the region through 1997 (Haiti, El Salvador, and Guatemala; Rice and Ward 1996). At a minimum, USAID should re- duce or eliminate its role in the cof- fee modernization process and aug- ment projects that promote prod- uction of organic and other environ- mentally sustainable coffee. It should be noted that PROMECAFE (a Cen- tral American USAID-supported pro- gram aimed at coffee modernization) held a workshop in February 1995 that sought to explore a sustainable coffee sector. Subsequently, this IICA-sponsored endeavor has begun to question the strictly modern model. Moreover, recent develop- ments and advances in organic cof- fee production in El Salvador point to USAID's burgeoning interest in alternatives to modernization. Marketing environmentally friendly coffee. Perhaps market forces can be harnessed to provide economic in- centives to farmers producing "envi- ronmentally friendly" coffee. Fortu- nately, because coffee functions as a segmented market, the possibility of providing market incentives for en- vironmental coffee is better than for many export commodities. An in- creasing number of consumers are prepared to pay premium prices for so-called specialty coffees?the fast- est growing segment of the coffee market. If shade-grown coffee can be marketed in this context, thereby providing higher prices to the pro- ducer, this could compensate for lower levels of production. Perhaps the closest that most US consumers can come to purchasing coffee with a high probability of coming from shaded plantations is to purchase "certified organic." The organic cof- fee sector, still a minuscule fraction of the total US coffee market (ap- proximately 0.5 % ),14 has grown con- siderably. Promotion of organic pro- duction has been embraced by grassroots institutions working with small cooperatives, particularly in Mexico, Costa Rica, and Nicaragua. Many producers have enthusiasti- cally embraced organic production because it brings higher prices. How- ever, the certification process can be time consuming, expensive, and bu- reaucratic. Farmer incentives. Without addi- tional long-term support, it may be 14M. Rozyne, 1994, personal communication. Equal Exhange Coffee Co., Stoughton, MA. difficult for the majority of small, traditional coffee farmers to com- pete with more industrial produc- tion units. Probably the least ex- plored set of policy tools with potential to influence coffee cultiva- tion techniques is incentives that could be provided to traditional farms through tax easements, access to credit, and technical and market- ing assistance. The rationale for such incentives would be that farmers employing traditional shade tech- niques are providing a long-term stewardship service of protecting topsoils, pure water supplies, and worker safety. This approach is at least partly addressed under pro- grams participating in the Interna- tional Coffee Register. The register certifies "fair trade" practices, in which roasters provide small farm- ers and producer cooperatives with, among other things, access to credit, prices above production cost indexed to world prices, and technical assis- tance to increase productivity using recently developed organic tech- niques and to diversify commodities produced. Currently, banks often tie access to credit, which is critical for farmers to bring coffee to market during periods of both low and high prices, to certain technological pack- ages that include the use of agro- chemicals, rather than to more eco- logically sustainable technologies. Internalization of environmental costs. Finally, growing coffee in modern plantations outcompetes more traditional systems in part be- cause associated environmental costs are paid by the state or people in other sectors of the economy, rather than by the coffee producer. These costs include the cleanup of polluted water supplies or the development of alternative sources of water, produc- tion declines associated with long- term pesticide use or soil erosion, the treatment of workers exposed to pes- ticides, and the loss of fish produc- tion in streams suffering sedimenta- tion. Establishing policies to ensure that some of the environmental costs are borne by the local producers would encourage more environmen- tally benign coffee production. One example could be an environmental or health fund supported by taxes on pesticides. 606 BioScience Vol. 46 No. 8 Conclusions Coffee lands within northern Latin American coffee-producing countries are undergoing fundamental changes. For the landscapes involved in this transformation, these changes trans- late into a reduced vegetative cover, lowered species diversity of the plant community and its associated fauna, and the application of agrochemi- cals onto lands that previously re- ceived little or no such inputs. Al- ready, 1.1 million ha of coffee within the countries in northern Latin America qualify as modernized. The total potential area that could be modernized is just more than twice that, at 2.7 million ha. What little work has been done on the environ- mental impact of the landscape modi- fications suggests that, unless steps are taken, many of these coffee zones, characterized by high rainfall and broken terrain, are likely to suffer environmental degradation in the coming years. This degradation is likely to include a severe loss of bio- logical diversity in areas where cof- fee plantations currently provide the last refuges. Actions that might re- verse this loss include working with small farmers to market ecologically sustainable coffee and reduce the support for technification in favor of policies that reward land steward- ship. Acknowledgments We thank L. Petit, M. Marvin, P. Rosset, J. Vandermeer, D. Weiden- field, J. Wunderle, and an anony- mous reviewer for commenting on the manuscript. Figures were pre- pared by Ellen Seefeldt and John Sterling. Support for research on coffee was provided to Ivette Per- fecto by National Science Founda- tion grant RII-9012484, and to Russell Greenberg by the National Geographic Society, Scholarly Stud- ies Program of the Smithsonian In- stitution, and the US AID Projecto Ambiental de Centro America pro- gram. References cited Adis J, Lubin YD, Montgomery GG. 1984. Arthropods from the canopy of inundated and Terra Firme forests near Manaus, Bra- zil with critical considerations on the pyre- thrum-fogging techniques. Studies on Neo- tropical Fauna and Environment 19: 223- 236. Agrios GN. 1982. Plant pathology. Orlando (FL): Academic Press. Aguilar-Ortiz F. 1982. Estudio ecologico de las aves del cafetal. Pages 103-128 in Avila- Jimienez E, ed. Estudios ecologia en el agroecosistema cafetalesa. Xalapa (Mexico): Instituto Nacional de Investigaciones Sobre Recursos Bioticos. Akkerman A, Van Baar X. 1992. El cafe organico: la sostenible de ungramo de oro. San Jose (Costa Rica): Centro de Estudio para el Desarrollo Rural, Universidad Libre de Amsterdam. Alcorn J. 1984. Development policy, forests, and peasant farms: reflections on Huastec- managed forests' contribution to commer- cial production and resource conservation. Economic Botany 38: 389-406. Askins RA, Lynch JR, Greenberg R. 1990. Population declines in migratory birds in eastern North America. Current Ornithol- ogy 7: 1-57. Babbar LL. 1993. Nitrogen cycling in shade and unshaded coffee plantations in the Cen- tral Valley of Costa Rica. [Ph.D. thesis.] University of Michigan, Ann Arbor, MI. Beer J. 1987. Advantages, disadvantages and desirable characteristics of shade trees for coffee, cacao and tea. Agroforestry Systems 5:3-13. . 1988. Litter production and nutrient cycling in coffee (Coffea arabica) or cacao (Theobroma cacao) plantations with shade trees. Agroforestry Systems 7:103-114. Benitez J, Perfecto 1.1990. Efecto de diferentes tipos de manejo de cafe sobre las comuni- dades de hormigas. Agroecologia Neotropi- cal 1: 11-15. Bergmann JF. 1969. The distribution of pre- Columbian America. Annals of the Associa- tion of American Geographers 59: 85-96. Borrero H. 1986. La substitucion de cafetales de sombrio por caturrales y su effecto negative sobre la fauna de vertebrados. Caldasia 15: 725-732. Boyce JK, Fernandez Gonzales A, Furst E, Segura Bonilla 0.1994. Cafe y desarrollo sostenible: del cultivo agroquimico a la produccion organica en Costa Rica. Heredia (Costa Rica): Editorial Fundacion UNA. Brash AR. 1987. The history of avian extinc- tions and forest conversion on Puerto Rico. Biological Conservation 39: 97-111. Carroll CR, Risch SJ. 1989. An evaluation of ants as possible candidates for biological control in tropical annual systems. Pages 30-46 in Gleisman SR, ed. Agroecology. New York: Springer-Verlag. Chazaro M. 1982. Flora epicola de la zona cafetales de Coatepec, Veracruz. Pages 95- 102 in Jimenez E, ed. Estudios ecologicas en al agrosistema cafetalero. Xalapa (Mexico): Instituto Nacional de Investigaciones Sobre Recursos Bioticos. Corredor G. 1989. Estudio comparative entre la avifauna de un bosque natural y un cafetal tradicional en el Quindio. [Professional the- sis.] University del Valle, Cali, Colombia. Coyner MS. 1960. Agriculture and trade in Nicaragua. Washington (DC): Foreign Ag- riculture Service. De Graaf R. 1986. The economics of coffee. Wageningen (the Netherlands): Publika- tienen Documentatie Dienst Landbouw Onderzoek. Estrada A, Coates-Estrada R, Merrit D Jr. 1993. Bat species richness and abundance in tropical rain forest fragments and in agri- cultural habitats at Los Tuxtlas, Mexico. Ecography 16: 309-318. . 1994. Non flying mammals and land- scape changes in the tropical rain forest region of Los Tuxtlas. Ecography 17: 229- 241. Erwin TL, Scott JC. 1980. Seasonal and size patterns, trophic structure, and richness of Coleoptera in the tropical arboreal ecosys- tem: The fauna of the tree Leuhea seemannii Triana and Planch in the Canal Zone of Panama. Coleopterist Bulletin 34:305-322. Frischknecht PM, Dufek JV, Baumann TW. 1986. Purine alkaloid formation in buds and developing leaflets of Coffea arabica. Phytochemistry 5: 613-617. Fuentes-Flores R. 1979. Coffee production sys- tems in Mexico. Pages 60-72 in de Salas G, ed. Agroforestry systems in Latin America. Turrialba (Costa Rica): The Tropical Agro- nomic Center for Research and Teaching. Gallina SE, Mandujaro S, Gonzales-Romero A: 1992. Importancia de los cafetales mixtos para la conservation de la biodiversidad de Mamiferos. Boletin Soc. Ver. Zool. 2: 11- 17. Greenberg R, Salgado-Ortiz J, Warkentin I, Bichier P. 1995. Managed forest patches and the conservation of migratory birds in Chiapas, Mexico. Pages 178-190 in Wilson M, Sader S, Santana E, eds. The conserva- tion of migratory birds in Mexico. Techni- cal Publication. Orono (ME): University of Maine, School of Natural Resources. Greenberg R, Bichier P, Sterling J. In press. Bird populations and planted shade coffee plan- tations of eastern chiapas. Biotropica. Hanson P. 1991. Los parasitoides asociados al cafeto en Costa Rica. Manejo Integrados de Plagas (Costa Rica) 20-21: 8-10. Ibarra-Nunez G. 1990. Los artopodos asociados a cafetos en un cafetal mixto del Soconusco, Chiapas, Mexico. Variedad y abundancia. Folia Entomologica Mexicana 79:207-231. Jepson PC, ed. 1989. Pesticides and non-target invertebrates. Andover: Intercept. Le-Pelley RH. 1973. Coffee insects. Annual Review of Entomology 18:121-142. Lovejoy TE, et al. 1986. Edge and other effects of isolation on Amazon forest fragments. Pages 275-285 in Soule ME, ed. Conserva- tion biology: the science of scarcity and diversity. Sunderland (MA): Sinauer. Majer JD. 1978. The maintenance of the ant mosaic in Ghana cocoa farms. Journal of Applied Ecology 13:123-144. Moron MA, Lopez-Mendez JA. 1985. Analisis de la entomofauna necrofila de un cafetal en el Soconusco, Chiapas, Mexico. Folia Entomologica Mexicana 63:47-59. Nestel D, Altieri MA. 1992. The weed commu- nity of Mexican coffee agroecosystems: ef- fect of management upon plant biomass and species composition. Acta Oecologia 13: 715-726. Nestel D, Dickschen F. 1990. The foraging kinetics of ground ant communities in dif- ferent Mexican coffee agroecosystems. Oecologia 84: 58-63. Nestel D, Dickschen F, Altieri MA. 1993. Di- versity patterns of soil macro-coleoptera in September 1996 607 Mexican shaded and unshaded coffee agroecosystems: an indication of habitat perturbation. Biodiversity and Conserva- tion 2: 70-78. Nir MA. 1988. The survivors: orchids on a Puerto Rican coffee finca. American Orchid Society Bulletin 57: 989-995. Perfecto 1.1994. Foraging behavior as a deter- minant of asymmetric competitive interac- tions between two ant species in a tropical agroecosystem. Oecologia 98: 184-192. Perfecto I, Snelling R. In press. Biodiversity and tropical ecosystem transformation: ant di- versity in the coffee agroecosystem in Costa Rica. Ecological Applications. Perfecto I, Vandermeer JH. 1994. Understand- ing biodiversity loss in agroecosystems: re- duction of ant diversity resulting from trans- formation of the coffee ecosystem in Costa Rica. Entomology (Trends in Agriculture) 2: 7-13. Pimentel D, Stachow U, Takacs DA, Brubaker HW, Dumas AR, Meaney JJ, O'Neil JAS, Onsi DE, Corzilius DB. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience 42: 354-362. Purata S, Meave J. 1993. Agroecosystems as an alternative for biodiversity conservation oi forest remnants in fragmented landscapes. Page 9 in Forest remnants in the tropical landscapes: benefits and policy implications. Symposium abstracts. Washington (DC): Smithsonian Migratory Bird Center. Reeves LH, Lilieholm RJ. 1993. Reducing fi- nancial risk in agrof orestry planning: a case study in Costa Rica. Agroforestry Systems 21:169-175. Rice RA. 1990. Transforming agriculture: The case of coffee leaf rust and coffee renovation in southern Nicaragua. [Ph.D. dissertation.] University of California, Berkeley, CA. _. 1993. New technology in coffee pro- duction: examining landscape transforma- tion and international aid in northern Latin America. Report to the Smithsonian Migra- tory Bird Center, Washington, DC. Rice R, Ward J. 1996. Coffee, conservation, and commerce in the Western Hemisphere. Washington (DC): Smithsonian Migratory Bird Center. Robinson M, Robinson B. 1974. A census of web-building spiders in coffee plantations at Wau, New Guinea, and an assessment of their insecticidal effects. Tropical Ecology 15:95-107. Room PM. 1971. The relative distribution of ant species in Guana's cocoa farms. Journal of Animal Ecology 40: 735-751. . 1975. Relative distribution of ant species in cocoa plantations in Papua New Guinea. Journal of Applied Ecology 12:47-61. Seib R. 1986. Feeding ecology and organization of neotropical snake faunas. [Ph.D. disser- tation.] University of California, Berkeley, CA. Somarriba E. 1990. Sustainable timber produc- tion from uneven-aged shade stands for Cordia alliodora in small coffee farms. Agro- forestry Systems 10: 253-263. . 1992. Timber harvest, damage to crop plants and yield reduction in two Costa Rican coffee plantations with Cordia alliadora shade trees. Agroforestry Systems 18: 69-82. Sotherton NW, Moreby SJ. 1988. The effect of foliar fungicides on beneficial arthropods in wheatfields. Entomophaga 33: 87-99. Stork NE, Brendell MJDI1990. Variation in the insect fauna of Salawesi trees with season, altitude, and forest type. Pages 173-194 in Knight WJ, Holloway JD, eds. Insects and the rain forests of South East Asia (Wallacaea). London (UK): The Royal En- tomological Society of London. Torres JA. 1984. Diversity and distribution of ant communities in Puerto Rico. Biotropica 16: 296-303. [UNFAO Production Yearbook] United Na- tions Food and Agriculture Production Year- books. Rome (Italy): United Nations. Vannini JP. 1994. Nearctic migrants in coffee plantations and forest fragments of south- western Guatemala. Bird Conservation In- ternational 4:209-232. Vitousek PM. 1984. Litterfall, nutrient cycling and nutrient limitation in tropical forests. Ecology 65: 285-298. Weaver PL, Birdsey A. 1986. Tree succession and management opportunities in coffee shade stands. Turrialba 36: 47-58. Wilson EO. 1987. The arboreal ant fauna of Peruvian Amazon forests: a first assessment. Biotropica 19: 245-251. Wrigley G. 1988. Coffee. New York: John Wiley & Sons. Wunderle JM, Latta S. 1994. Overwinter turn- over of Nearctic migrants wintering in small coffee plantations in Dominican Republic. Journal fuer Ornithologie 135:477. . In press. Avian abundance in sun and shade coffee plantations and remnant pine forest in the Cordillera Central, Dominican Republic. Ornitologia Neotropical. Wunderle JM, Waide RB. 1993. Distribution of overwintering nearctic migrants in the Ba- hamas and Greater Antilles. Condor 95: 904-933.