NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Wilson EO, Peter FM, editors. Biodiversity. Washington (DC): National Academies Press (US); 1988.

Cover of Biodiversity


Show details

Chapter 14Tropical Dry Forests The Most Endangered Major Tropical Ecosystem


professor of Biology, University of Pennsylvania, Philadelphia, Pennsylvania

The rain forest is not the most threatened of the major tropical forest types. The tropical dry forests hold this honor. When the Spaniards arrived in the Western Hemisphere, there were 550,000 square kilometers of dry forest (approximately five times the size of Guatemala, or the size of France) on the Pacific coast of Mesoamerica (an area extending north from Panama to western Mexico). Today, only 0.09% of that region (approximately 480 square kilometers) has official conservation status, and less than 2% is sufficiently intact to attract the attention of the traditional conservationist. If there is to be a conserved neotropical (i.e., Western Hemisphere) dry forest wildland large enough to maintain the organisms and the habitats that were present when the Spaniards arrived, and if it is to be large enough to be easily maintained and thus a project willingly undertaken and managed into the indefinite future by the society in which it is imbedded, then we will have to grow it (Janzen, 1986a).

The story is the same for the dry tropical regions of Australia, Southeast Asia, Africa, and major parts of South America. The cause of the severe habitat loss is straightforward. Dry forest is easily cleared with fire, and woody regeneration in fields or pastures is easily suppressed with fire. Furthermore, fire does not stay where you put it; many areas are unintentionally cleared. The farmer is also aided by the severe dry season; it suppresses pest and weed populations, facilitates the use of fire as a tool to clean up pastures and fields, and slows soil degradation caused by continuous rain and farming. Many tropical dry forest regions even have good soils; they are downwind of volcanic mountain ranges or are situated on alluvia.

What are the conditions in a tropical dry forest (cf. Hartshorn, 1983; Janzen, 1986a; Murphy and Lugo, 1986)? Its 4- to 7-month rain-free dry season is sufficiently harsh that many species of trees, vines, and herbs are deciduous for 2 to 6 months. Its rainy season, during which 1 to 3 meters of rain can fall, is as wet, if not wetter, than that of a rain forest. In the dry season, the sun penetrates to the forest floor, the leaf litter becomes very dry (and virtually ceases to decompose), watercourses dry up or greatly diminish in flow, and daytime relative humidity ranges from 20 to 60%. The dry forest may appear uniformly green during the rainy season, but during the dry season this homogeneity changes into a complex mosaic of tens of habitat types distinguished by the differential drying rates of different soils and exposures, different ages of succession, and different vegetation types. Many animals migrate to moist refugia (hollow logs and caves, moist riparian sites, north-facing slopes protected from the wind, and sites close to rain forests). During the dry season, most plants cease their vegetative activities, but many species of woody plants flower, mature their fruits, and disperse their seeds. Some species of animals feed on dry season fruits, seeds, and flowers; for them, the dry season is the bountiful time of year and the rainy season, inimical.

Diversity in the Dry Forest

What is the level of diversity in tropical dry forests? A lowland dry forest adjacent to a lowland rain forest (such as a portion of the Pacific dry and Atlantic wet sides of northern Costa Rica) sustains a fauna and flora about 50 to 100% as species-rich as does the neighboring rain forest (Janzen, 1986a). Floras are the least similar in richness of species, largely because the dry forest epiphytes and trees are substantially less rich in species. The greatest similarity in species richness is represented by mammals and major insect groups such as butterflies and moths (Lepidoptera) and Hymenoptera such as bees, wasps, and ants. Species overlap between the two areas, ranging from less than 5% (e.g., epiphytes, amphibians) to as high as 80% (e.g., sphingid moths, mammals). In the 11,000-hectare dry forest of Santa Rosa National Park in Costa Rica, I estimate that there are 13,000 species of insects, and fairly accurate counts indicate that there are 175 breeding species of birds (Stiles, 1983). There are also 115 species of nonmarine mammals (Wilson, 1983) and about 75 species of reptiles and amphibians (Savage and Villa, 1986). All the species of small herbs and grasses have not yet been collected, but the final list of angiosperms (which include vascular plants such as orchids and trees) will probably not exceed 700 species (Janzen and Liesner, 1980). When such a dry forest habitat is replaced by fencerows, ditchsides, unkempt pastures, and woodlots, the species richness of the breeding fauna and flora is reduced by 90 to 95%.

It is true that long lists of species have been used as criteria for identifying tropical habitats worthy of conservation. However, an approach that merely considers the number of species present is incomplete. This can be seen by examining the tropical dry forest, which is less rich than the rain forest in total species but is much richer in its variety of species' activities. It contains many species that remain dormant in inclement (wet or dry) weather, and species that magically find enough water to develop flowers, fruits, and leaves at the height of the dry season (e.g., Janzen, 1967, 1982a,b). Its parasitoids and grazers range from absolutely monophagous to extremely polyphagous; that is, some subsist only on one type of food, whereas others feed on a variety of organisms. Where else can you find white-tailed deer eating fruits dropped by spider monkeys and coyotes foraging side by side with jaguars? Moreover, the Santa Rosa dry forest has the only wind-pollinated legume (Janzen, in press a), the fiercest ant-plant mutualism (Janzen, 1966), and an enormous seasonal pulse of caterpillars that changes to a nearly total absence of caterpillars while the host plants are still in full leaf (Janzen, in press b). If you want a plantation timber tree that will grow throughout the year in a tropical rain forest, yet withstand the droughts produced by agricultural clearing, look to dry forest trees rather than to rain forest trees. The dry forest is also the parental climate for many major tropical crops and food animals such as cebu cattle, chickens, cotton, rice, corn, beans, sweet potatoes, sorghum, and pasture grasses (Janzen, 1986b).

Rather than focusing just on lists of species, tropical conservationists should also concentrate on saving interactions among species. What good are such interactions? Interactions make wildlands interesting, and they provide the raw materials used by the natural historian to construct the stories and the visions that are the real value of the natural world to humanity (Janzen, 1986b).

The conservation world has by and large failed to exploit the real enticements of the areas it conserves. Tropical forests can be likened to libraries and books. The value of a book is not measured by the number of words it contains or even by the number of kinds of words it contains. Put most simply, how long will the public continue to support a library whose goal is to have enormous holdings but no card catalog, no librarians, and thus no readers. Such a library is doomed to fail during the next paper shortage or governmental budget cut. Books are also comparable to the species in the dry forest in that they have little meaning except in context. The context of the dry forest is unique because of its species; it was once widespread and certainly has as much potential for biocultural development as does any type of tropical vegetation. It is this context that we must save.

The likelihood of long-term survival of a conserved wildland area is directly and strongly proportional to the economic health and stability of the society in which that wildland is imbedded. The farm- and ranchland once occupied by dry forest often sustains economically strong regional subcultures, which can thrive without the need to exploit the conserved area. Land blessed with conservation status in such a subculture has a much higher chance of survival than do the more abundant wildlands in frontier like subcultures, many of which are also based on marginal farmland.

Overemphasis on the length of species lists is also potentially misleading, because the extraordinary peaks in species richness encountered at certain sites can be highly atypical. They are very interesting ecologically but are not representative of the tropics as a whole. They may not even be representative of the site itself, since many of the species-rich sites bear accumulations of strays at the overlap between several less species-rich habitats. Furthermore, the remaining sites of very great endemism and extraordinarily extensive species richness are often sites with peculiar physical characteristics that render them less likely to be occupied by humans at present and therefore easier to conserve. The apparently successful conservation of such areas gives a feeling of accomplishment that makes it easier to accept conservation failures or inactivity with less species-rich, and therefore more ordinary, tropical sites. In the headlong rush to conserve diversity, we risk leaving the next generation with a handful of pretty baubles rather than the substance of the tropics. Saving a habitat with 300 species of endemic orchids on an Andean mountaintop may not have the same long-term ecological, intellectual, or economic value as does saving remnants of once widespread and less species-rich lowland forest.

A Many-Faced Threat

The threats to the tropical dry forest are multiple and complex. The concerned observer is correct to be no longer stirred to action by the simplistic chant, "The beef cow is responsible for the demise of the tropics"; the music is more daunting, more complicated, and more site-specific.

There are almost no large blocks of dry forest still standing that can be destroyed and thus cause concern among the public and academic world. Equally important, there are few opportunities to recognize the biocultural deprivation of the ranching and farming cultures that have been sustained for hundreds to thousands of years by the soils that once supported dry forest. As tropical conservation has swung into high gear during the past three decades, it has become comfortable to focus largely on the remaining rain forest and not to worry about scraps of other scattered vegetation types such as the dry forest. A traditional conservation battle for tropical dry forests would have to have been fought in 1900. Today, restoration ecology and habitat management (e.g., Janzen, in press c) are the only answers.

The acquisition and restoration of dry forest wildlands conflict with traditional conservationist protocol in numerous ways:

  • Land apparently used for agricultural production, or land that has produced something, is being set aside. This is expensive land, and its acquisition is often accompanied by a last-minute harvest of the few remaining trees and other resources by sellers who can do very serious damage to the anticipated restoration of the site.
  • The sellers—poor to very wealthy—are likely to be involved in neighborhood functions and politics far into the future; they cannot be bought out and left as resentful recipients of a bad deal.
  • The frontier is gone. The audience is local. The power is local. Within a few decades, if it hasn't happened already, almost all members of tropical societies situated on dry forest soils will be settled on firmly titled land and will be leading a real or vicarious urban life with amenities such as good roads and schools. Survival of a wildland will depend on regional policy decisions by government institutions and planning commissions, and those decisions will be made by or in conjunction with the local community.
  • Many dry forest species are relatively robust, largely due to their evolutionary history of exposure to seasonal changes. Thus even tiny population fragments and severely altered populations can be ecologically reworked into viable, interacting populations and complex habitats that are replicas or facsimiles of what once was on the site. Habitats that appear well beyond recovery can be restored if the seed sources are present.
  • Intensive cleaning up of the Mesoamerican dry forest agroscape during the last two to three decades is leading to the final extirpation of species and habitats that survived the first wave of megafaunal extinction by hunters 9,000 years ago, the extensive agriculture that began 5,000 years ago, and the ever more intensive agriculture that began 500 years ago. Either we act very soon or we will witness the elimination of many species that have persisted through many seemingly more severe perturbations than the contemporary, innocuous-appearing clearing of the last fencerows.
  • Conserved areas of dry forest wildlands will be rich in plant and animal opportunistic species and may well be the only places where most of today's weeds survive. Weeds may be the most information-rich carriers of the genetic information for environmental toughness—information of obvious value in genetic engineering for crop species in harsh habitats.
  • Fires and invasion by grasses are the most serious contemporary ecological threats to the restoration and maintenance of dry forest wildlands. Properly manipulated, domestic animals may be the best tools for managing these threats, and they may even pay for their own maintenance: they mow the competing grass, they eat the fuel for the next dry season's grass fires, and they disperse tree seeds far into pastures.
  • Dry forest conservation requires not only restoration but also explicit efforts to eliminate the various species initially introduced for agricultural or ranching purposes.
  • The biggest and a perpetual problem in dry forest management lies in deciding which areas of the wildland will be managed in what manner and to what end. Yes, it can be returned to a natural state, given the availability of seed sources. But which natural state do you want? On a time scale of at least thousands of years, the state to which it returns or in which it remains depends on many factors, such as the initial condition of the site, the species that arrived, and the order in which they arrived.

What means can be used to restore a tropical dry forest habitat?

  • Initiate and maintain a heavy flow of biological information, both biocultural and economic, from the site into the neighboring social system. The process of restoration, and the biology and interactions of the organisms being restored, must become as familiar to the region as are its irrigation projects, school development, and health programs. This task is both more difficult and more sustainable if the conservation effort is focused on habitats, interactions, and caterpillars rather than on redwoods, lions, and condors.
  • Stop man-made fires, hunting, cattle ranching, and other free-ranging perturbations. That is to say, give the site back to the remaining or adjacent dry forest organisms to recolonize by their own means. However, while this multihectare regeneration appears to be natural, such megacolonization of pastures and fields does not occur anywhere in nature. It is also not risk-free; neighboring blocks of pristine vegetation are likely to be severely altered by invasions of secondary successional organisms from the large and ever-growing areas under restoration.
  • If some organisms are to be reintroduced from elsewhere, how far back in time shall we reach? Do we return the Pleistocene horse to Central American dry forest wildlands? How do we do that without adding its predators? If we put the tapir and white-lipped peccary back into El Salvador, do we also bring back their food plants? There are no correct choices per se, but it is clear that certain major dry forest areas must be set aside purely for agriculture and that no effort must be spared to maintain certain other (much smaller) areas as wildlands.
  • If the goal of restoration ecology is to conserve a maximum number of species, the management plan would be quite different than if the goal is to conserve habitats and interactions with whatever species they normally contain. Management directed at the conservation of a maximum number of species leads at least to the fragmentation of the wildland into a mosaic of successional types and ages and the introduction of species from other areas. If the goal is to conserve interactions as well as species, then the wildland manager must predict rather than simply react. The interactions that are saved will depend on the management steps taken, which depend on the interactions that are desired (and there will be errors and surprise outcomes). Conservation abruptly graduates from the art of patrolling a boundary against poachers to a variety of technical activities, such as the research-based studies of ecological succession, evolutionary biology, species packing, competitive exclusion, and epidemiology.
  • Rain forest wildlands must be conserved within migratory reach of the dry forest areas that are subject to restoration. A dry forest does not exist unto itself and neither does a rain forest. In Central America, the rain forest and the dry forest are the mutual recipients of each other's migrants—migrants that are important parts of the interactive structure that holds tropical habitats together. Birds migrating from Wisconsin to Costa Rica are not the only ecological link over large agroscapes.
  • The dry forest is not only a collection of many kinds of habitats, each rich in unique species, but the members of a given habitat can be important interactants in adjacent habitats. If only certain (usually species-rich) dry forest habitats are slated for conservation, one quickly discovers that a substantial portion of the species in those habitats spend critical parts of their lives in other nearby habitats. To put it another way, the conversion of highly deciduous forest on dry ridges to pasture may have a severely depressing effect on the species-richness of organisms in the very species-rich adjacent alluvial bottom lands.

The dry tropics contain adult remnants of a once thriving forest, juveniles from gradually dwindling seed reservoirs, and waifs from as yet intact wildlands. These organisms now stand on a trashed agroscape and will die without replacement. They are the living dead—all physiologically alive but can be regarded as dead if they were already lying in the litter (Janzen, 1986b). If they flower, they fail to set seed (lack of pollinators). If they set seed, the seeds do not disperse (lack of seed dispersers). If seeds disperse, they do not develop as new members of their population (lack of adequate conditions for growth and development). If they develop, they do not thrive in a sustained population (the caprice of agroecosystem development eradicates conditions needed for population maintenance within a generation or two). These organisms are usually included in the lists of species in a region and are often used to demonstrate that a species is not threatened with extinction—even though it is (Janzen, 1972). Agroscapes, seemingly still supporting long lists of widespread species, are primed for massive extinction as individuals of these species senesce or are killed through intensification of contemporary agriculture. If all these species were to be physically removed as soon as they have no future, the catastrophe would be much more noticeable and would therefore arouse the sentiments normally associated with massive extinction.

The dry forest is more prone to these less visible catastrophes than is the rain forest, in which scraps of vegetation left when the forest is cleared die more quickly than do those of the dry forest. Thus the threatened plant species in dry forest are available for a longer time and can be used as basic stock in restoration projects (though the price paid is that they dilute the visual impact of a largely demolished forest).

A Central American tropical dry forest wildland that is large enough to be visited and used by humans is substantially larger than a wildland that is to exist without human intrusion. Tropical habitats are very rich in behaviorally sensitive species and species that exist in low-density populations. Morever, dry forests contain many small habitats (e.g., springs, dry ridge tops, marshes, edaphic outcrops, temporary streams, and pickets of forest sheltered from the wind). Visitors (tourists researchers, seed collectors, and habitat managers) will perturb ecological interactions substantially more and effect them more permanently in dry forests than in most extratropical or rain forest habitats. This calls for strict zoning for habitat use and replicate habitats, both of which can be compatible with conservation management only if a large acreage is set aside.

Future Prospects

Ignore the voice that demands that a monetary value be placed on a wildland or a species before it can be conserved. Is that what you would do to determine the need for a public library, a public hospital, a public school? Can you tell me the dollar value of these institutions to your children? Are we to continue to be led by commercial interests to sanctify the production of material goods? The great majority of tropical humans live as draft animals; they are sold to the highest bidders along with the habitats that maintain them, and the purchasers are not generally benevolent. Through the swirl of changing market values, there will eventually come a day when the living organisms in a tropical wildland would be as doomed as would be libraries, if books were valued only for their paper pulp and the price of paper pulp were to rise.

Many organisms we believe to be safe are really endangered, and those we call endangered are in reality extinct. Guards will not save tropical wildlands. The world's dry tropics are already way beyond their capacity for accommodating human activity. Thus a contract between managers of wildlands and society is mandatory. And the scientific community must aggressively participate in writing and executing the contract. Without this participation, tropical biology will be nothing but low-grade and gradually diminishing restoration ecology. The conservation community has valiantly propped up the fortress walls, but they are too few. The future lies in the children, but we cannot wait for a well-educated cohort to replace its parents. The tropical dry forest is a living classroom, and its students are its neighbors. The collective power to turn the game around resides with policy makers. We cannot force the world to conserve tropical nature; we must seduce it, and the bait is intellectual mutualism—not the dollar value of a caterpillar.


The people of Costa Rica have inspired me to believe there is still a chance. U.S. tax dollars through the National Science Foundation have financed the acquisition of the knowledge to see the chance, and the academic community has given the peer approval to know that this is the right direction. I thank P. Raven, W. Hallwachs, P. May, and A. Ugalde for help with the manuscript.


  • Hartshorn, G. S. 1983. Plants. Pp. 118–157 in D. H. Janzen, editor. , ed. Costa Rican Natural History. University of Chicago Press, Chicago.
  • Janzen, D. H. 1966. Coevolution of mutualism between ants and acacias in Central America. Evolution 20:249–275. [PubMed: 28562970]
  • Janzen, D. H. 1967. Synchronization of sexual reproduction of trees with the dry season in Central America. Evolution 21:620–637. [PubMed: 28563685]
  • Janzen, D. H. 1972. The uncertain future of the tropics. Nat. Hist. 81:80–89.
  • Janzen, D. H. 1982. a. Cenizero tree (Leguminosae: Pithecellobium saman) delayed fruit development in Costa Rican deciduous forests. Amer. J. Bot. 69:1269–1276.
  • Janzen, D. H. 1982. b. Variation in average seed size and fruit seediness in a fruit crop of a guanacaste tree (Leguminosae: Enterolobium cyclocarpum). Amer. J. Bot. 69:1169–1178.
  • Janzen, D. H. 1986. a. Guanacaste National Park: Tropical Ecological and Cultural Restoration. Editorial Universidad Estatal a Distancia, San Jose, Costa Rica. 103 pp.
  • Janzen, D. H. 1986. b. The future of tropical ecology. Ann. Rev. Ecol. Syst. 17:305–324.
  • Janzen, D. H. In press a. Natural history of a wind-pollinated Central American dry forest legume tree (Ateleia herbert-smithii Pittier). In C. H. Stirton, editor; and J. L. Zarucchi, editor. , eds. Advances in Legume Biology. Mo. Bot. Gard. Monogr. Syst. Bot.
  • Janzen, D. H. In press b. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica.
  • Janzen, D. H. In press c. Management of habitat fragments in a tropical dry forest: Growth. Ann. Mo. Bot. Gard.
  • Janzen, D. H., and R. Liesner. 1980. Annotated check-list of plants of lowland Guanacaste Province, Costa Rica, exclusive of grasses and non-vascular cryptogams. Brenesia 18:15–90.
  • Murphy, P. G., and A. E. Lugo. 1986. Ecology of tropical dry forest. Ann. Rev. Ecol. Syst. 17:67–88.
  • Savage, J. M., and J. Villa. 1986. Introduction to the Herpetofauna of Costa Rica. Society for the Study of Amphibians and Reptiles, Athens, Ohio. 207 pp.
  • Stiles, F. G. 1983. Checklist of birds. Pp. 530–544 in D. H. Janzen, editor. , ed. Costa Rican Natural History. University of Chicago Press, Chicago.
  • Wilson, D. E. 1983. Checklist of mammals. Pp. 443–447 in D. H. Janzen, editor. , ed. Costa Rican Natural History. University of Chicago Press, Chicago.
Copyright © 1988 by the National Academy of Sciences.
Bookshelf ID: NBK219281


Related information

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...