Ecotoxicology has focused almost exclusively on countries and ecosystems in temperate zones. Tropical ecosystems, which combined contain as much as 75% of the global biodiversity, have been neglected. Tropical ecosystems are under increasing threat of development and habitat degradation from population growth and urbanization, agricultural expansion, deforestation, and mining. Some of these activities also lead to the release of toxic substances into the environment. Little research in ecotoxicology has been carried out in tropical environments. Techniques and procedures developed for temperate environments are often applied, even though physical and chemical environmental parameters in the tropics can be very different. Most research has focused on water quality and aquatic toxicology. The regulatory environment also varies among countries. We present a review of the literature on tropical ecotoxicology, with an emphasis on Latin America and the Caribbean. We also address priority areas for immediate research in the tropics. These include large-scale agricultural activities, especially banana, pineapple, and soybean farming, and gold mining with the associated heavy use of mercury. We outline the special issues that must be addressed as the field of tropical ecotoxicology progresses.
Tropical terrestrial and freshwater ecosystems cover only 25.7% of the land area on the surface of the Earth , but this area generates nearly 60% of the primary productivity of the planet  and contains approx. two-thirds of all the known species of vascular plants . The number of species of most major taxa increases with decreasing latitude  (Table 1); thus, tropical ecosystems harbor the bulk of the world's species. Some estimate that more than two-thirds of the world's flora and fauna is found in the tropics .
Concern over the amount of environmental degradation in tropical ecosystems has increased dramatically over the past decade. The 1992 Earth Summit in Rio de Janeiro, Brazil, further catalyzed interest in the tropics, and the 1994 Summit of the Americas in Miami, Florida, USA, addressed hemispheric concerns over environmental problems in nations with tropical environments. A document prepared for the Miami summit presented examples of innovative approaches to sustainable development and conservation in the tropics .
Innovative approaches are indeed necessary because tropical ecosystems are disappearing at an alarming rate (Table 2). Average annual deforestation between 1981 and 1985 in the Amazon Basin alone totaled 32,500 km2 year, an area equivalent to the U.S. states of Massachusetts and Connecticut combined . Moreover, the annual deforestation rates of the countries of the Amazon Basin (Table 2) are far less than those observed in other nations such as Thailand (2.7.%), Costa Rica (4.0%), and Nigeria (5.0%) .
Although deforestation is the most direct threat to tropical biodiversity, tropical ecosystems are currently threatened by other human activities as well. Little research has been done on the impact of contaminants on tropical ecosystems, yet, considering research conducted in temperate regions, these compounds have potentially large effects . In addition, tropical ecosystems are more poorly studied than temperate ecosystems, and the physical and chemical variables that affect biotic processes are different. We present an introduction to tropical ecosystems with a brief review of important environmental impacts and contaminant issues. We close this report with recommendations for applying the principles of environmental toxicology to the different ecological, sociological, and economic situations that are encountered in developing tropical countries.
The extent and nature of the tropics
Tropical environments differ ecologically from temperate zone habitats in physical, chemical, and biological attributes. Tropical environments occur between approx. 25° north and 25° south of the equator and are characterized by warm temperatures with little or no seasonality and heavy precipitation during at least part of the year  (Fig. 1). The tropics can be broadly classified into six major habitat types: tropical rain forests, tropical dry forests, tropical savannas, tropical wetlands, tropical freshwater systems, and tropical marine environments  (Table 3).
Throughout the tropics, combinations of temperature and precipitation are encountered that do not exist in temperate regions. Our understanding of the fate and transport of environmental contaminants in northern latitudes has little applicability in the tropics . In addition, biological diversity in the tropics is substantially higher than in temperate zones, and the number of species potentially affected by any given compound is also greater. Many taxa are poorly known, so little or nothing has been published on even the most fundamental aspects of their biology. New species are continually being described, and many are rare. Finally, given the high biodiversity of the tropics, the degree of interactions among species is likely to be much higher, and thus the possible ramifications of the indirect effects of a contaminant are likely to be more complex. Tropical ecotoxicology is not merely an extension of methods and techniques developed in temperate latitudes. The mere cost of trying to screen the diverse tropical biota for possible effects of an environmental pollutant dictates that new approaches will have to be developed.
Table Table 1.. Trends in latitudinal diversity for swallowtail butterfliesa
a Figures are the numbers of species in each band of latitude. The pattern exhibited here is typical of most groups of plants, invertebrates, and vertebrates .
North of the equator
South of the equator
Environmental problems in the tropics
Deforestation and agricultural expansion. Loss of habitat is by far the most serious environmental problem in both temperate and tropical habitats. Per hectare impact is higher in the tropics, however, because of the higher alpha diversity of tropical systems. The causes of deforestation vary from continent to continent and from country to country. Deforestation for the commercial production of timber is a major factor in southeast Asia, whereas slash-and-burn agriculture and the production of fuel wood is more severe in Africa and Latin America . Commercial production in Africa peaked in the 1950 s and has since declined, but timber exploitation in Latin America is currently on the rise and is expected to surpass that in Asia by the end of the century .
The population growth rates of many tropical nations, especially those in Africa, remain high, with annual rates of more than 3% common [13,14] (Table 4). A 3% annual population growth rate means that the country will double in population in slightly over 23 years. This places great pressure on a nation to expand agricultural capacity. As a consequence, enormous tracts of land have been claimed for agricultural use in the tropics. The percentage of land used for agriculture has consistently increased throughout the tropics while remaining constant or declining slightly in developed countries . Agribusiness has also expanded in tropical countries as foreign investors are lured to develop agricultural exports. One consequence of expanding agribusiness has been the emigration of peasants from rural areas to urban areas; thus, many tropical countries now are among the most urbanized nations on earth (Table 4).
Table Table 2.. Data on forest cover, deforestation, and protected areas for the countries of the Amazon Basin 
Forest area (1,000km2)
Deforestation (1,000 km2/year)
Reserves (% protected)
Agricultural exports help to improve the trade balance of developing countries without the high cost required to build industrial infrastructure. Foreign currency earned through exports is often desperately needed to pay the interest costs on foreign debts, which have escalated greatly since the 1970s (Table 4). The economies of tropical countries have deteriorated over the past several decades. Although per capita gross national product has increased in most countries of the Amazon Basin, gross national product growth has slowed or is declining. Inflation, on the other hand, has shown explosive growth (Table 4). Most developing nations (with exceptions like Brazil and Mexico) continue to depend heavily on the exportation of raw agricultural products. The expansion of agriculture and the concomitant environmental problems will continue as reliance on exportation to mitigate economic distress continues.
Population growth and infrastructure expansion. Countries that are doubling their populations every 25 to 35 years must double all aspects of their infrastructure (schools, universities, hospitals and doctors, energy production, raw material output) at the same rate to maintain their standard of living. This places tremendous pressure on developing countries to expand mining for internal consumption and exportation, to dam rivers to increase electrical power generation, to accept foreign development projects that may have large-scale environmental impacts, and to avoid passing or implementing environmental regulations because the social cost of these regulations in the short term is too high. This dooms developing countries to continue to decline in all measurable aspects of quality of life (Appendix), including environmental quality.
Table Table 3.. General aspects of tropical ecosystems 
Rain forests Equatorial
Between 15°N and 15°S of the equator
Little seasonality in temperature or precipitation; rainfall from 3,000 mm in continental forests to 10,000 in montane forests
Equator to as far as 25°N and 25°S, along the eastern coasts of continents
Seasonal at higher latitudes; rainfall comparable to that in equatorial forests
Between 12 and 20°N and between 12 and 20°S latitude
All months warm with mild seasonality but extreme seasonality in precipitation; 1,500 mm of rainfall or more, concentrated in 6 months or less
Tropical dry forests
Rain shadows on the west side of equatorial mountains (e.g., Guana-caste region of Costa Rica), N or S of interior savannas (e.g., South American Chaco), and in areas under certain edaphic, altitude or relief conditions (e.g., east African dry woodlands)
Seasonality increases with increasing latitude, though climate is generally hot; precipitation is seasonal and can be less than 500 mm annually
Tropical environments present special challenges for research and conservation because of their ecological characteristics. Conservation, however, is a social process, and the socioeconomic conditions of a culture will greatly affect conservation practices. A review of research on the use and effects of toxic substances in the tropics must be evaluated in this context. The following review is intended to provide an overview of studies that evaluate the effects of toxic substances on tropical ecosystems. Coupled with the other reports in this annual review issue, this report will present a clear picture of the status and needs of the field of tropical ecotoxicology. At the end of this report we address a strategy for pursuing the development of the field in the future.
LITERATURE REVIEW OF CHEMICAL USE AND IMPACTS IN THE TROPICS
Toxic substances are produced and used in the tropics in agriculture, mining, crop storage, and the prevention of human disease and disease vectors. Although adverse effects can be associated with large-scale chemical use, demand has not diminished . Rather, industrial companies have built production and development facilities in tropical countries, occasionally superseding the laws, regulations, and logistics surrounding production in temperate nations. The use of toxic substances has increased with the growth and development of less developed nations in the past 20 years, and their effects on human and wildlife health have become a concern.
When investigating the effects of chemical contaminants in tropical ecosystems, habitat modifications, such as those made for silviculture, agriculture, and mining, should also be considered. Extensive habitat changes lead to alterations in structural biodiversity. A variety of sensitive species appropriate for toxicity testing could have already been eliminated. Furthermore, wildlife and habitat managers have limited access to data on the ecology, relative abundance, and distributions of species in a given area. Consequently, there is little scientific evidence when species are affected by chemical exposures, further precluding the capacity for management decisions by governments of affected tropical nations. Wildlife concerns related to environmental contaminants are not simply the problems of local countries but are international in scope. Few case studies exist demonstrating international cooperation in management of contaminant effects in wildlife. Several recent studies, however, have documented and suggested solutions for effects of chemicals on Neotropical migratory birds [16–18].
The controversial issue of human and wildlife health hazards associated with the use of persistent or acutely lethal chemicals in tropical regions  has brought the topic of widescale pesticide use under the scrutiny of government and nongovernment agencies, scientists, and the lay public . For example, throughout the 1970s and 1980s, worker protection in tropical agriculture was not a strong concern . Public pressure, developed from ideas such as those suggested by Matthews and Clayphon , stimulated educational extension programs, field safety tests, and government control initiatives to ensure worker safety both in the field and in homes adjacent to aerially sprayed fields [22–24]. The field of tropical ecotoxicology needs similar progress addressing the effects of contaminant on terrestrial and aquatic organisms. The following section presents examples of emerging ecotoxicological issues. Special emphasis is given to neotropical migratory birds (NTMBs), those 340 species that nest in the United States and Canada and migrate to the neotropics for the nonbreeding season.
Heavy metals and mining
Heavy metals enter the ecosystem via natural geological processes, mining, and industrial activity. Regulations controlling runoff and disposal of wastewater laden with heavy metals are not always followed during operations, and heavy metals have accumulated in human and wildlife tissues. Effects are subchronic to acutely lethal and have been shown to produce reproductive dysfunction and learning disabilities. Where mercury has been released into the aquatic system as a result of unregulated gold mining, subsequent contamination of invertebrates, fish, and birds has been detected; biomagnification of mercury was documented from gastropod mollusks (Ampullaria sp.) to accipiters (Rostrhamus sociabilis) and from invertebrates and fish to waterbirds and humans [25,26]. Mercury contamination of fish used as a local food source or as an exportable commodity is both an ecological and socioeconomic issue. Alho and Vieira  found that 50% of the fish from the Cuibá River and 35% from the Bento Gomes River in the Brazilian Pantanal had mercury concentrations exceeding international standards for consumption. Pfeiffer et al.  observed high concentrations of mercury (> 70 mg/g) in the hair of individuals consuming fish along a tributary of the Amazon where mercury was being used in gold mining activity; no data are available on mercury poisoning in these populations.
Table Table 4.. Population, economic, and environmental data on the countries of the Amazon Basina
b Estimates in these categories are not all from the same years within the range.
Mercury levels in relation to the distribution and diets of seabirds were analyzed in Perú by Gochfeld . Seabirds feed in the open ocean for most of the year and are presumed to be less exposed to many contaminants than shorebirds or terrestrial birds. Geographic location and trophic level were more related to mercury accumulation than taxon. Levels of mercury in tissues were low in all cases but nevertheless indicated the widespread distribution of this metal. In a later study comparing geographic factors, no differences in lead or cadmium concentrations were found between terms nesting in tropical regions and those resting in temperate regions .
Persistent organochlorines in terrestrial, aquatic, and marine environments
Organochlorine (OC) insecticide residues in tropical wildlife have been investigated more frequently than other contaminant burdens. Avian reproductive effects and mortality in North America stimulated this interest. The proximity of wildlife habitat to agricultural areas, especially with expansion into nondeveloped areas, appears to have influenced patterns of exposure. Residues of the OCs, such as DDE (dichlorodiphenyldichloroethylene), DDT (dichlorodiphenyltrichloroethane), dieldrin (IR, 4S, 4aS, 5R, 6R, 7S, 8S, 8aR-1,2,3,4,10,10-hexachloro-1,4,4a,5,6,7,8,8a-octahydro-6,7-epoxy-1,4,5,8-dimethanonaphthalene), HCB (hexachlorobenzene), DDD (1,1-dichloro-2,2–bis (p-chlorophenylethane), lindane (1,2,3,4,5,6-hexachlorocyclohexane), endrin (1,2,3,4,10,10-hexachloro-6,7-epoxy-1,4,4a,5,6,7,8,8a-octahydro-endo-endo-1,4,5,8-dimethanonaphthalene), and aldrin (IR, 4S, 4aS, 5S, 8R, 8aR-1,2,3,4,10,10-hexachloro-1,4,4a,5,8,8a-hexahydro-1,4,5,8-dimethanonaphthalene), were detected in mammals, birds, reptiles, and fish in tropical Australia . Most samples from undeveloped areas had no detectable OC residues, whereas OC residues were found in 69% to 94% of the samples from developed areas. Similarly, in Kenya, 18 species of raptors from agricultural areas  contained residues of DDT metabolites and dieldrin, although no detectable levels of OCs were found in raptors from nonagriculrural areas. Both of these studies are more than 20 years old, however, and lack of detection could be attributed to the inability of older techniques to detect lower levels of contaminants.
Persistent OCs, such as HCH (BHC, or 1,2,3,4,5,6-hexachlorocyclohexane) isomers, DDT, PCBs (polychlorinated biphenyls), and HCB, were detected in wildlife occupying an agricultural watershed in south India . In birds, OC residue levels varied according to feeding habits. The highest levels were found in inland piscivores and scavengers, then coastal piscivores, insectivores, omnivores, and granivores. Vermeer et al.  documented large fish kills in Surinam after application of PCP (pentachlorophenol). The chemical had been sprayed to eliminate Pomacea snails from rice fields. Fish mortality exceeded tens of thousands. Upon further analysis, pesticide effects were found to extend beyond mortality of fish. Kites, herons, egrets, and jacanas were found sick or dead in surrounding roosts. Snail kite (R. sociabilis) mortality was attributed specifically to PCP. Sickness and mortality in the egrets, herons, and jacanas coincided with intense applications of endrin.
Allsopp  looked at the long-term photoisomerization of dieldrin in Kenyan savannas. Photoisomerization of dieldrin did not increase short- or long-term hazards to wildlife species, although residues in vegetative and wildlife tissues did take approx. 2 years to return to prespray levels. From 1987 to 1991, Douthwaite and Tingle  investigated and compared the burdens and effects of the use of OCs to control tsetse flies on animal populations in Zimbabwe. Fauna monitored included bats, birds, lizards, fish, and insects. Adverse effects at the population level were found in four bird and one lizard species. It was believed that DDT spraying was responsible for the comparative scarcity of several bird and terrestrial invertebrate species in sprayed areas. Similarly, survivability indices indicated increased risk to bat species exposed to other environmental stressors, such as drought. No significant effects were detected in fish or on soil processes. Mora  also looked at OC levels in 51 endemic and migratory avian species in Texas. Although DDE was found in 97% and PCBs in 66% of the avian tissues, a trend showing decreasing residue concentrations was found through comparisons of data from 1965 to 1988.
Persistent OCs and heavy metals have been detected in marine mammals since 1980. The major source of contamination is believed to be tropical countries, but large volumes of volatized OCs may also be dispersed through the atmosphere, and residues have been found in the open ocean from the tropics to the Arctic . Large natural metal deposits may also account for some of the mercury accumulation that was detected in several organs of 35 specimens of Stenella coeruleoalba stranded on the French Atlantic and Mediterranean coasts and 45 S. attenuata captured in the eastern tropical Pacific . Bioaccumulation via consumption of contaminated prey is considered to be the major route of mercury contamination in dolphins . In the Bay of Bengal, India, large quantities of OCs were detected in cetaceans . Marine mammals may accumulate large quantities of these chemicals because of the high quantity of lipid-rich blubber they possess . Large quantities of lipophilic pollutants can also be transferred to young during lactation . One reason for such dramatic accumulation detected in marine mammals and cetaceans is the lower capacity of metabolic enzymes to break down and excrete OCs [36,39].
Chemical use in agroecosystems
The most prominent issue in the tropics concerning toxicants is the use of agrochemicals to meet food requirements of growing populations. Agricultural chemical use has enhanced tropical agricultural productivity and provided an array of relatively inexpensive fruits, vegetables, and grains worldwide. The wide use of chemicals has improved nutritional standards for inhabitants of tropical nations. Chemicals are also increasingly used to accommodate human encroachment into non-developed areas to allow for additional agricultural development through vector eradication and weed control. Edwards  completed an earlier review of chemical use in developing countries. Hazards of pesticide use identified in 1977 included the development of new pest problems, including development of resistance to pesticides, contamination of the human food supply, and wildlife mortality. All remain concerns 20 years later.
Agribusiness uses fungicides, herbicides, and pesticides applied through a variety of methods. Some chemicals have been thoroughly tested and evaluated by the U.S. Environmental Protection Agency (EPA), whereas others have been banned or, because of extensive wildlife study requirements, not registered. Some of the latter chemicals are heavily used throughout the tropics. For example, monocrotophos (dimethyl-(E)-1-methyl-2-(methyl carbamoyl)-vinyl phosphate), an organophosphate (OP) insecticide, was registered in the United States until 1989, when it was withdrawn because exhaustive avian studies were required to reregister the chemical . Although it is no longer in the United States, it remains the third most widely used insecticide in the world .
Trees are used to shade cocoa plantations  in equatorial Guinea. Rodents living in the trees will collect cocoa pods or scratch and bite the pods. The damage to pods allows the fungal contamination to spread. The removal of rodent habitat reduces the quality of cocoa seed produced and eliminates wildlife habitat. In the past the most cost- and time-effective method of removing rodents was to use a wax block formulation of the second-generation anticoagulant brodifacoum (3–(3,4′-bromo(1–1′-biphenyl)-4-yl)-1,2,3,4-tetrahydro-1-napthalenyl)-4-hydroxy-2H-1-benzopyran-2-one). In large-scale field trials in the Philippines, Hoque and Olvida  pulse baited the rodenticide flocoumafen (4-hydroxy-3-(1,2,3,4-tetrahydro-3-(4-(4 trifluoromethyl-benzyl-oxy)phenyl)-1 napthyl) coumarin) at a rate of 1.175 kg/ha per season. Residues in wild species of rats were highest 2 days after baiting and fell below 0.03 mg/kg 10 to 25 days after baiting. The authors concluded that when used as directed, flocoumafen provides cost-effective rodent control with minimal environmental impact. In a related article, integrated pest managers and crop protection agencies in southeast Asia used brodifacoum as a cost-effective method of increasing yield, with production increases of 1 to 2%, usually enough to repay the cost of treatment . No domestic or wildlife causalities were reported in either of these projects, but the effects of bait blocks on the full complement of endemic wildlife species were not studied. In Malaysian cocoa and coconut plantations, although field poisoning of wood rats (Rattus tiomanicus) reduced the population to 13% of its original size, the population recovered to 88% of its original size after only 6 months . Recruitment into the depopulated areas was also found to be via immigration.
Macadamia orchards in Hawaii were studied  to determine the toxicity of rat bait containing 1.88% zinc phosphide (Zn3P2) to avian species. Although consumption of this bait by granivorous birds in sugarcane fields has been previously shown , no avian mortality was documented. Widespread acceptance of bait by birds indicated that the bait should not be applied when indigenous seed-eating birds are present.
As intensely managed agroecosystems with frequent pesticide applications, banana plantations are ideally suited to assess wildlife exposures [20,49] Mortensen et al.  developed and tested methods of estimating pesticide distribution and fate on and around banana plantations in Costa Rica while simultaneously assessing wildlife. Drift cards proved to be a useful tool in monitoring aerial applications of fungicides, whereas water samples provided detectable levels of pesticides only after rain. In individuals from 30 avian species captured, reactivation of blood cholinesterase was sufficient to determine exposure to cholinesterase-inhibiting compounds. Carbamate reactivation was seen in 39% of the plasma samples obtained from the plantations, whereas only 13% of samples obtained from birds captured outside the plantations showed reactivation. This suggested higher probabilities of exposures proximal to chemical applications.
Banana production on Ecuador's Pacific slope increased dramatically during 1991–1992. Taura syndrome was first recognized as mortality in commercial penaeid shrimp farms in the Gulf of Guayaquil in 1992 . A decrease in shrimp growth and survivability reduced income for shrimp exporters in the Gulf of Guayaquil. The synergistic action of fungicides and herbicides, transported in high concentrations downstream into the gulf, appeared to cause the mortality. Losses to the shrimp industry were estimated at more than $45 million (U.S. dollars) . Taura syndrome was later proven experimentally to be caused by a virus [51,52]. The syndrome has since been found on shrimp farms in Peru, Colombia, Honduras, and Hawaii . Viral comparisons of isolated, yet naturally infected, penaeid samples from Hawaii and Ecuador confirmed that the same virus was responsible for outbreaks in both shrimp-growing regions . Results of wildlife studies testing chemical effects on wildlife on Ecuadorian banana plantations may be suspiciously met by industry in the future; this exemplifies the need for better development of cause-and-effect linkages and the danger in presuming cause without adequate evidence.
Speculation continues over the effects of spraying tropical drug plantations with the herbicide paraquat (1,1′-dimethyl-4,4′-bipyridinium ion). Aerial spraying for drug eradication, funded by the United States, began with the approval of the U.S. Congress in 1981. Most applications were targeted at marijuana and coca plantations, although paraquat has been aerially applied over opium plantations in Colombia . Indiscriminate poisoning of people, agriculture, wildlife, soil, and water may occur. Furthermore, drug users in countries into which contaminated drugs are illegally imported may also be affected by lung necrosis due to paraquat inhalation or ingestion.
In Australia, the abundance of mammalian fauna has declined during the past 20 years. Short and Smith  found that in Australia extinctions occurred at a disproportionately higher rate among medium sized ground-dwelling mammals (0.035–5.500 kg); they concluded that this was due to exotic predators. Pesticides were considered to be a minor factor, but intensive applications of 1080 (sodium monofluoroacetate) were used to reduce and eliminate predator populations. No studies were conducted on the effects of a nonselective mammalian citric acid cycle inhibitor on wildlife populations.
Neotropical migratory birds: Overview and case studies
Ornithologists have noted declines in the populations of migratory birds during the past several decades, particularly those that migrate to tropical regions [55–57]. The dramatic and mysterious nature of these declines sparked a series of workshops and symposia to address the issue [58–61]. Numerous causes have been hypothesized, including habitat fragmentation in North America, increased levels of nest predation, parasitism in brown-headed cowbird nests, and tropical deforestation [55,56]. Recently, both the popular  and activist  press have highlighted the expansion of export agriculture, in particular bananas, as being responsible for a number of environmental ills. Pesticide use is frequently highlighted by the media.
Gard and Hooper  and Gard et al.  reviewed the potential hazards of pesticides and other environmental contaminants to Neotropical migrants. The effects of many of these chemicals are well documented on some species of birds. Effects can be lethal or sublethal, and action can be direct or indirect . Direct effects include mortality, decreased reproductive success, compromised immunological function, and carcinogenesis. Indirect effects can result as a consequence of reduced prey abundance, altered habitat structure, and changes in community interactions. Agroecosystems offer the best medium for the study of these alterations, and the insights derived have applications beyond ecotoxicology .
Tropical ornithologists have shown that Neotropical migrants use altered, disturbed, and agricultural habitats while overwintering in the tropics , although the effect of agricultural chemicals typically has not been evaluated. Up to 30% of the avifauna populations in disturbed areas of the tropics is made up of migrants , and almost all rely on small forest patches in a mosaic of agricultural habitats [68–70]. This means that they are at high risk for exposure to agricultural chemicals used in these areas. Indeed, many species of Neotropical migrants have elevated body burdens of some persistent contaminants as compared with Nearctic resident species .
Ehrlich  commented that the decline in populations of songbirds in the eastern United States has followed of the loss of tropical forests in Central and South America and that diminished coastal wetlands have resulted in reduced availability of food resources for shorebirds. The destruction and alteration of habitat could have caused carrying capacities to be lower for wintering birds in tropical countries, although similar problems within North American breeding habitats share responsibility for declines. Use of chemical pesticides combined with habitat losses may severely impact sensitive species. For example, deforestation and the pollution of water with pesticides and by-products of fuel alcohol production are the major causes of the deterioration of the Brazilian Pantanal . Studying patterns of avian migration will provide data on the relative quality of breeding grounds and wintering environments for selected species.
Seasonal and geographic variations in OC residues have been observed in migratory bird populations. The Peregrine falcon (Falco peregrinus) has been studied most intensively. Henny et al.  determined that OC levels were significantly higher in falcons returning from Latin America in the spring than in birds leaving North America in the fall. In addition, migrant and resident avian prey of Peregrine falcons were collected in Surinam, Ecuador, Peru, and Costa Rica , Resident species had higher DDE levels than migrants, and carnivorous and insectivorous species had higher DDE levels than omnivorous and granivorous species. The highest levels of contamination were found in Peru and Ecuador. Similarly, Mora and Anderson  showed that birds continued to be exposed to OCs in wetland habitats and agricultural systems in Mexico.
Since the reduction or ban of OCs and the general move from use of OC to OP and carbamate pesticides, studies have indicated declines in wildlife OC burdens. Mora et al.  found no significant difference in concentrations of DDE in birds along the Pacific flyway in Mexico versus California. Similarly, Elliot and Shutt  showed that the residue levels of the more persistent OC chemicals in Sharp-shinned hawks (Accipiter striatus) were not significantly different before and after migration. Dieldrin and heptachlor epoxide (epoxide of 1,4,5,6,7,8,8-heptachloro-3a,4,7,7a-tetrahydro-4,7-methanoin-dene), however, have shorter half-lives in the body and were significantly increased in the blood of all age classes after each winter and significantly decreased after each breeding season. Mora  also demonstrated a reduction in OC residues in a variety of avian species. What was once a long-term reproductive and immunological concern encountered with OC use in pelican and raptor species has become a serious international problem of acute toxicity in NTMBs due to widespread use of OP and carbamate pesticides.
Basili and Temple  studied the migration and wintering behavior of the dickcissel (Spiza americana) between the United States and the Llanos of Venezuela. Farmers in Venezuela aerially sprayed roosts of dickcissels with OP pesticides to eliminate this grassland bird from foraging on their rice and sorghum crops. The majority of the breeding population of dickcissels migrates to a very restricted area within the Venezuelan Llanos, and roosts may contain hundreds of thousands to 3 million birds (G. Basili, personal communication). With a total population of about 10 million birds, a significant portion could be wiped out by killing one large roost. Through educational programs, financial compensation for crop losses, and alteration of crop rotation schedules, farmers have been invited to join in the international conservation of this species.
The migratory patterns of Swainson's hawk (Buteo swainsoni) recently have been studied as the birds travel between North America and the grassland provinces of Argentina  to evaluate stressors encountered during migration and on wintering grounds. It has been postulated that this species feeds while migrating through the tropics ; the low probability of finding sufficient suitable foraging areas in Central America and northern South America could provide the impetus for the long flights between available habitats. A roost with more than 700 dead birds was discovered in 1995 ; the cause of death at the time was not known, but pesticide exposure was suspected. Researchers returned to the site in 1996 equipped to assess pesticide contamination. More than 4,000 mortalities were initially observed, and samples were collected for subsequent analyses . Tests for cholinesterase activity, pesticide residues, and degradates confirmed that the deaths were due to monocrotophos intoxication (M. Hooper, personal communication). This is the largest mortality incident in a bird of prey ever recorded [16,17].
This review assessed the use and impacts of toxic substances on wildlife species in the tropics. Within the boundaries of this region, few detailed, long-term toxicological studies have been completed. Results of existing studies demonstrate potential impacts both on and off sites where wildlife exposures have occurred, although the highest risks of exposure and accumulation are in agricultural areas . The growth of wildlife toxicology in the tropics will need to draw upon developments in the assessment of exposure through the use of biomarkers . The joining of biochemical and behavioral toxicology allows increasingly sensitive levels of acute and chronic exposures to be measured and used for management purposes . Long-term assessment of chemical impacts on wildlife will provide data for more accurate evaluation of effects on endemic and migrating tropical wildlife. Although the movement and fate of chemicals is poorly understood in tropical systems, the call for in-depth and long-term research must be answered. Recent studies documenting mortality in migratory passerine and raptor species have set off wide-ranging philosophical and moral discussions about land management and tenure and the regulation of agricultural chemicals as they relate to economic viability, species survivability, integrated land use management, and international responsibility for the conservation of wildlife.
RESEARCH PRIORITIES FOR TROPICAL ECOTOXICOLOGY
Priority areas for research
Given the diversity of tropical ecosystems, it is likely that environmental contaminant issues will be far more complex than those encountered in temperate countries. It is not possible, with our current state of knowledge, to predict where the most serious problems will occur. Several areas with well-documented problems are presently of concern; these deserve immediate attention.
Gold mining has expanded in many tropical countries. In most cases it is unregulated, and the amount of mercury released into the environment during these operations is a major health hazard for wildlife and humans in several parts of Brazil [25,26]. Larger mining operations tend to be less problematic than very small-scale, local operations, which are almost impossible to regulate. Almost all are located near rivers high in organic material that favors the formation of organic methyl mercury . Even small mines can be located fairly easily with aerial photographs, and areas of potential contamination can be identified.
Given the high toxicity of mercury and the enormous quantities utilized in gold mining operations, research on the environmental consequences of mercury use in the tropics is a very high priority. Because of the tendency of mercury to bioaccumulate, special emphasis should be placed on studies of predatory fish (many of which are consumed by humans), piscivorous birds, raptors, and other top predators. Many of these top predators might serve as keystone species and be responsible for structuring community patterns and influencing community processes [81,82]. Mercury contamination has been documented in isolated regions of the Amazon . It will become increasingly difficult to describe the structure and function of undisturbed tropical aquatic communities if the use of mercury remains unregulated and unstudied.
Many tropical countries have greatly expanded the amount of land used for agriculture, either to feed growing populations or to export cash crops. Much of this land is intensively managed and receives large inputs of agricultural chemicals. Many chemicals used are not registered in the United States [e.g., 16]. In addition, the fate and transport of these chemicals under the environmental conditions of the tropics are poorly understood . Some research suggests that risk of exposure and intoxication is greater with high temperature and humidity . Several projects and pilot studies conducted in tropical agroecosystems have revealed either the potential for wildlife impacts [8,49,64] or extensive mortality [16,18,19]. These controlled agriculture environments are an ideal setting for experimental field studies on the impacts of agricultural chemicals on tropical wildlife species . Species of special concern include Neotropical migrants and organisms in drainage canals, streams, and rivers adjacent to agricultural areas, as well as species in outflow water bodies when more persistent chemicals are deposited.
Banana plantations occupy about 50,000 ha of the total lowland area of Costa Rica and support a variety of species of wildlife. Banana cultivation on these plantations is an intensely managed agroecosystem with high inputs of anthropogenic chemicals, generally in the form of pesticides and fertilizers. Much emphasis has been placed on estimating loss of wildlife due to deforestation, but there is little information on the exposure of migrant birds to agrochemicals used on banana plantations. Mortensen and colleagues  found that 42 species of birds were observed on several banana plantations in Costa Rica, which contradicts the notion that banana plantations are “avian deserts” . Eighteen species (43%) of these birds were classified as migratory by the United States Fish and Wildlife Service. Exposure data suggested that pesticides might be a complicating factor in the health of migrants that use banana plantations.
Little ecological and less toxicological work has been done in the agroecosystems of Costa Rica. Thus, only a small database is available on the biodiversity of and exposure of birds to agrochemicals used on banana plantations. Great opportunity exists to unite research on patterns of biodiversity with a pesticide exposure study. Such a study would serve as a catalyst for collaboration between scientists, environmental groups, and industry.
The discovery of heavy mortality of Swainson's hawks due to exposure to OP pesticides on their wintering grounds in Argentina [16,17] sparked an international conservation initiative. With close to 6,000 documented mortalities (likely a substantial underestimate of the total mortality) in two winter seasons, interactions with the Argentine government allowed for the development of an international cooperative approach, involving scientists and regulators from the United States, Canada, and Argentina, for the management of pesticide impacts on Swainson's hawks. Chemical labeling will be more precise, testing strategies will be re evaluated for avian species, and regulatory control of chemical use will be strengthened and delegated to provincial governments. It is hoped that these actions, combined with an intensive farmer extension information program, will prevent massive mortalities like those of 1996. The model being developed and implemented in Argentina is one that should be considered for other tropical regions and for other species of raptors.
A related concern is the possible impact of OPs like monocrotophos on species in other parts of the world. Virtually nothing is known of the impact of OPs on smaller tropical and migratory passerines in developing countries. The magnitude of the mortality observed in Swainson's hawks certainly suggests that programs of careful monitoring of the health of wildlife in and near agricultural areas in the developing world is of the highest priority.
Special issues in tropical ecotoxicology
Protecting tropical environments involves more than science. Tropical environments are found in a wide swath of equatorial countries that are also among the poorest on the planet. Tropical ecotoxicologists will, like tropical conservationists, need to deal with the complex social and economic problems that result from poverty if they are ever to develop effective research programs. In addition, tropical ecotoxicology is by nature international and multicultural; thus, political and cultural issues will also compete with science issues for attention. The challenge to tropical ecotoxicologists is, indeed, great, and for research programs to be effective, they will have to be multidisciplinary and international. Any “strategy for tropical ecotoxicology” will have several components that are paramount.
1. International cooperation. Projects can easily be international and not be cooperative. Any resident of Brazil, Zaire, India, or any other developing tropical country can provide many examples. International projects are too often one-way streets, with researchers from developed countries collaborating with their counterparts from the tropical countries in only the most superficial manner. Frequently, research results are published without any of the collaborators from the developing country listed as an author; often, the simple courtesy of sending reprints is neglected.
International projects should be structured to be truly collaborative, with full participation of scientists from the developing country in the planning, research, analysis, and publication phases. Visiting scientists should always offer to present seminars at host institutions; if you don't speak the language, speak slowly and have a colleague translate. Interact with students and spend time to provide them with information. Always volunteer to copy articles or book chapters that are difficult to obtain and send them to the sponsoring institutions. Research in tropical ecotoxicology will have to be long-term to generate meaningful results. Striving to cooperate to the fullest extent and cultivating the confidence of colleagues in your host country is the most important aspect of international research.
Several examples of international collaboration provide a model for research collaborations in tropical ecotoxicology. The Organization for Tropical Studies, with its three field facilities in Costa Rica, has long been considered a model for international collaboration in tropical ecology . This organization has recently become more involved in applied ecology, but the major focus of the program is basic research. A facility for conducting research in applied ecotoxicology is greatly needed. Such a facility should provide laboratories for the analysis of environmental contaminants in a variety of matrices. Support for a centralized facility could be provided by the governments of collaborating countries as well as multilateral funding agencies. The facility could be staffed by an international team to facilitate the exchange of ideas and information. The creation of such a pilot facility, for a region such as Central America, the Pantanal, or the Brazilian Amazon, could be spearheaded by a coalition of scientific societies such as the Society of Environmental Toxicology and Chemistry and the Ecological Society of America.
2. Education, outreach, and technology transfer. Tropical countries seek international collaborators largely because the base of trained and qualified researchers in each country is too small to address all environmental problems. There is no lack of work to be done in the tropics and no justification for perpetuating the status quo. One of the greatest benefits institutions in the developed world can provide is education, especially at the graduate level , International projects should, whenever possible, provide assistantships or fellowships for students from the host country. Awarding a Ph.D. degree to a scientist from a developing country can have a major impact when the student returns to his or her country. Whenever possible, projects should be designed so that the field component is done in the tropics, not in a temperate country; this contributes to the information base for future research activities.
Another aspect of education is technology transfer. Projects should include funding for the purchase of much needed equipment to be left in the host country and should provide training on the use and maintenance of such equipment. Cooperative agreements should include reserve funding for the purchase of parts if possible. Education coupled with technology transfer will help to improve the infrastructure for science in the tropics and will allow scientists in developing countries to operate as peers in future international collaborations.
Technology transfer does not need to be for scientists only. The project previously discussed on the mitigation of the effects of agricultural pesticides on Swainson's hawks is heavily dependent on the transfer of technology to local farmers. Other grassroots projects, like the AMISCONDE initiative in Costa Rica and Panama , owe their success to involving all levels of the community in a sustainable development and conservation project and to providing access to appropriate information and technology. Any project that deals with the mitigation of the effects of small-scale gold mining or agricultural chemicals must include grassroots involvement or it will fail. As a consequence, social considerations will need to be a part of the initial planning of any ecotoxicology project in a developing country that deals with contaminant problems that arise at the local level.
3. Ecosystem and landscape-level analyses. Single-species toxicity testing may be of limited value in the tropics. Diversity is extremely high, and most species of vertebrates are found at much lower densities than is typical for temperate zone species. Unless a convincing argument can be made for a given species as a valuable indicator, the use of assemblage, ensemble, community, or ecosystem level indices should be pursued. An exception is a case in which a contaminant is affecting a threatened or endangered species. The issue of development and extrapolation of models and indices to the community and ecosystem level is contentious . The arguments for the protection of tropical ecosystems hinge on either structural (high biodiversity) or functional (complex interspecific interactions such as plant-pollinator relations, nutrient cycling of greenhouse gases) emergent properties of communities or ecosystems. Thus, there is interest in evaluating the impact of contaminants not so much on individual species but rather on systems. Tropical research can make a major contribution to the development of ecotoxicology as a whole by forcing the issue of the development of ecosystem toxicity testing. This will require extensive collaborations between field researchers, laboratory scientists, and modelers .
If the focus of tropical ecotoxicology is at the community or ecosystem level, then all research should consider landscape-level analyses. Many factors will affect community structure and function, and most will be present in any study. The spatial representation of areas of deforestation, industrial contamination, and agriculture runoff will facilitate interpretation of their relative impacts on a target population or community. Geographic information systems (GISs) are becoming widely used in developing countries as the cost of hardware and software declines . The extra research effort involved is minimal. High-quality maps can be generated and are far more effective than white-paper reports when working with policy makers. In addition, once spatially related data on multiple impacts are entered into a GIS, the software allows spatially explicit modeling of scenarios of either continued impacts or mitigation. Landscape-level analyses are becoming increasingly more important in ecology, natural resources management, and conservation biology. They are also more widely used in ecotoxicology : this is another new research tool that can be developed more effectively via research in the tropics.
4. Risk-based approaches. The ultimate goal of tropical ecotoxicology must be the advancement of new paradigms in the management of land in the tropics. Environmental risk assessment models, like that recently developed in the United States by the EPA, , effectively combine research goals, economic constraints, and management objectives into a process that is open-ended and subject to future revision and modification. The poor database on wildlife populations and environmental contaminant issues in the tropics is exactly the situation that the environmental risk assessment paradigm is best suited to resolve. Evaluating projects under a risk assessment approach will also allow researchers and regulators to determine the weight to be placed on economic and social considerations, an especially important consideration in the developing world.
Applying the strategy for tropical ecotoxicology: An idealized example
Presenting a list of special issues in tropical ecotoxicology is only part of the task. Applying the strategy will require adapting the issues to the research question at hand. The case study on Neotropical migrants provides a good template. What is the best mechanism for addressing the question of the impact of contaminants on migratory birds? As a first step, we propose the creation of a “Center for the Study of the Ecotoxicology of Neotropical Migratory Birds” (“MIGRATOX”). The purpose of the center would be to provide a base for international collaboration on the study of contaminants in Neotropical migrants. The center would be staffed by an international team working on banding, radio and satellite tracking, behavioral ecology, and behavioral and biochemical toxicology.
Scientists from MIGRATOX would collaborate with researchers from leading academic institutions to provide several levels of education and technology transfer. The center would help to coordinate academic training at the master, doctoral, and postdoctoral level. The center would also conduct short courses for government officials on regulatory and development issues. Other educational activities would involve collaborations with teams of rural sociologists that would work with affected communities at the grassroots level.
The center would have an extensive geographic database on species distributions, migratory pathways, trends in population numbers, habitat, and agricultural or industrial activities. This would allow the broad, landscape-level analyses that will be required to track the myriad impacts present throughout the migratory pathway of any given species. The GIS approach would also allow the development of models that would examine scenarios of development and their impacts on migratory bird populations. These models would provide vivid, visual demonstrations to decision makers to assist them in the development of long-term plans for land-use management.
All long-term plans should be developed using an environmental risk-assessment model. This would allow researchers and planners to incorporate uncertainty into their plans; it will also allow new data to be introduced at later dates to upgrade and improve the predictive capabilities of the models. This is clearly an ambitious plan, but it also illustrates the power of an integrated strategy in resolving complex, multinational environmental problems.
Tropical ecotoxicology is more than traditional toxicology carried out in the rain forest. The vastly different biological, climatic, and physicochemical conditions that exist in the tropics are reason enough to argue for a new disciplinary approach. When the socioeconomic, political, cultural, and regulatory differences are brought into consideration, the contrast between approaches and methods that will work in temperate versus tropical regions becomes more sharp. The objective of this special issue of Environmental Toxicology and Chemistry is to bring these concerns to the broader scientific community. Anyone concerned about biodiversity must be aware of environmental problems in the tropics; the contaminant issue, especially regarding impacts on biodiversity, has largely been neglected. It is our goal that this series of reports will enlighten the readership to the problems, research, and possible solutions of ecotoxicology in tropical regions.
We thank M. Hooper, T. La Point, and P. Stangel for their critical comments and revisions.
Table . Complicating factors in developing countries that impede development and lead to economic decline and environmental degradation