Note: This website is no longer being updated and is being maintained for archive purposes by the Department of Botany, National Museum of Natural History, Smithsonian Institution. Please see About the Project for further details.

Centres of Plant Diversity: the Americas

Link to North America map of regional study sites
North America map

Link to Middle America map
Middle America map

Link to South America
South America map

Link to Centres of Plant Diversity home page


Botany

SOUTH AMERICA
REGIONAL OVERVIEW

Introduction: Total land area: c.17,519,900-17,529,250 km².
Population (1995): 320,000,000.
Geology: Maximum altitude: 6959 m.
Vegetation: Essentially all life zones and physiognomic vegetational formations. Principal vegetation types are tropical evergreen and semi-evergreen moist forest, dry forest to woodland, open grassy savanna, desert and arid steppe, Mediterranean-climate communities, temperate evergreen forest, and several montane formations (e,g. páramo, puna).
Flora: Number of vascular plant species: Over 83,000.
Number of endemic vascular plant species: Over 50,000; regions with 5%-76% endemism.
Number of genera of flowering plants: 4200.
Social and environmental values
Threats

Centres of plant diversity and endemism
References

Introduction

The continent of South America has about one-eighth of the Earth's land surface, situated between latitudes 12°N-55°S and longitudes 80°-35°W; no other continent has a greater latitudinal span. Eighty percent of its land mass is within the tropical zone, yet it extends into the subantarctic. The extensive zones of temperate and cold climates in the vicinity of the Equator, in the Andes, are unique. The land area of about 17,519,900-17,529,250 km² is under the jurisdiction of 13 countries (Table 49); French Guiana is governed as an overseas department of France. The region's 1995 population of c. 320 million people is estimated to reach 452 million people in 2025. Three of the world's 21 megacities are in South America: São Paulo, Buenos Aires and Rio de Janeiro (WRI, UNEP and UNDP 1994).

Geological setting

Although the neotropics may be conveniently considered as a single phytogeographic unit, the region is geologically complex. The neotropics include not only the South American continental plate but the southern portion of the North American plate, as well as the independent Caribbean plate (Clapperton 1993). The complicated geological history of the region, for example as these plates intermittently separated and collided through the Cretaceous and the Tertiary, provides the milieu within which plant evolution has been superimposed.

South America has been an island continent during most of the period of angiosperm evolution, whereas Central America constitutes one of the two tropical parts of the Laurasian "world continent". Both South America and North America have been moving westward, roughly in tandem, since the breakup of Pangaea in the Mesozoic. In contrast, the Antillean plate with its flotsam of Antillean islands formed only during the Cenozoic and has moved in a retrograde eastern direction, at least with respect to its larger neighbours. Whereas South America and North America have been widely separated through most of their geological histories, there has been generally increasing contact between them through most of the Cenozoic, culminating in their coalescence with formation of the Isthmus of Panama c. 3.1 million years ago (Keigwin 1978). The date of this epochal event in neotropical geological history has been gradually estimated to be younger, with estimates of 5.7 million years ago giving way to as recently as 1.8 million years ago (Keller, Zenker and Stone 1989). In addition to their Pleistocene connection via the Isthmus of Panama, South America and North America apparently were more or less directly interconnected via the protoAntilles for a short time near the end of the Cretaceous, prior to formation of the Caribbean plate (Buskirk 1992).

The outstanding geological feature of South America is the Andes, the longest mountain range in the world, which extends in a nearly straight line of over 7000 km from the north to the southern tip of the continent. The Andes have the highest mountain in the Western Hemisphere, the highest mountain in the world's tropics, and as measured from the centre of the Earth (rather than metres above sea-level), the highest mountain in the world. The most important break in the north-south sweep of the "cordillera" is the Huancabamba Depression in northern Peru, where the eastern chain of the cordillera is entirely ruptured (by the Marañón River) and even the western chain dips to 2145 m (at the Abra de Porculla). The existence of this massive mountain range has had profound effects on plant and animal evolution in South America, and consequently has profound effects on essential conservation priorities.

In essence, the Andes represent a classical plate tectonic upthrust of continental rock, as the leading edge of the westward-moving South American plate collides with the oceanic Pacific plates. The Southern Andes are the oldest, with significant uplift already present in early Cenozoic times, prior to the Oligocene. Most of the uplift of the Central Andes was in the Miocene or later, whereas most of the uplift of the northern portion of the cordillera has been Plio-Pleistocene (van der Hammen 1974). To the north the Andes become more geologically complex, breaking into three separate cordilleras on the Ecuador/Colombia border. Much of the north-western margin of South America, including Colombia's western and central cordilleras, appears to be amassed "suspect terrane" rather than an integral part of the South American continental plate (Juteau et al. 1977; McCourt, Aspden and Brook 1984).

Much of the rest of the South American continent consists of two great crystalline shields that represent the western portion of what was once Gondwanaland. The north-eastern portion of the continent constitutes the Guayana Shield, whereas much of Brazil south of Amazonia is underlain by the Brazilian Shield. These two major shields were formerly interconnected across what is today the Lower Amazon. They consist of a Precambrian igneous basement overlain by ancient mucheroded Precambrian sediments. The Guayana region has been the most heavily eroded, with basement elevations mostly below 500 m interrupted by massive flattopped table mountains, the fabled "tepuis", typically rising to 2000 m or 2500 m. The peak of the highest of these, Cerro Neblina or Pico da Neblina on the Venezuela/Brazil border, reaches an altitude of 3015 m and is the highest point in South America outside the Andes. The tepuis and similar formations are highest and most extensive in southern Venezuela, becoming smaller and more isolated to the west and east where La Macarena near the base of the Andes in Colombia and the Inini-Camopi Range in French Guiana respectively represent their ultimate vestiges.

The quartzite and sandstone of the Guayana Shield erode into nutrient-poor sands, and much of the Guayana region is characterized by extreme impoverishment of soils. The rivers draining this region are largely very acidic blackwater rivers, of which the Rio Negro is the most famous.

The Brazilian Shield is generally higher and less dissected, with much of central Brazil having an elevation of 800-1000 m. The Brazilian Shield is mostly drained by clearwater rivers such as the Tapajós and Xingú.

In contrast to these ancient shields, the Amazonian heartland of South America is low and geologically young. Prior to the Miocene most of Amazonia constituted a large inland sea opening to the Pacific. With uplift of the Central Andes, this sea became a giant lake that gradually filled with Andean sediments. When the Amazon River broke through the narrow connection between the Guayanan and Brazilian shields near Santarém, Brazil, Amazonia began to drain eastward into the Atlantic. Nevertheless, the region remains so flat that ocean-going ships can reach Iquitos, Peru, which is only 110 m above sea-level, yet 3000 km from the mouth of the Amazon and less than 800 km from the Pacific Ocean. Most of Amazonian Ecuador, Peru and Bolivia is below 200 m in elevation. The process of Amazonian sedimentation is continuing, as the sediment-laden white-water rivers course down from the Andes, continually changing their channels and depositing and redepositing their sediments along the way. About 26% of Peruvian Amazonia shows direct evidence of recent riverine reworking (Salo et al. 1986). With the lack of relief, it is not surprising that rather fine nuances of drainage, topography and depositional history are often major determinants of vegetation.

Like Amazonia, some other distinctive geological features of the South American continent are relatively low, flat and geologically young, such as the chaco/pantanal/pampa region to the south, the Venezuelan/Colombian Llanos to the north and the trans-Andean Chocó region of Colombia and Ecuador to the west. Large portions of these areas have been inundated during periods of high sea-level in the past, and large portions of all of these regions are seasonally inundated presently.

One aspect of the geological history of Latin America that has received much biogeographic attention is the series of Pleistocene climatic fluctuations and their effects on distribution and evolution of the present neotropical biota. It is clear from the palynological record that major changes in vegetation were associated with the cycles of Pleistocene glaciation (e.g. van der Hammen 1974), although to what extent lowland Amazonia was predominantly drier (e.g. Haffer 1969; van der Hammen 1974), colder (Colinvaux 1987; Liu and Colinvaux 1988) or both, and how this affected the Pleistocene distribution of tropical forest, remain hotly contested (Colinvaux 1987; Räsänen, Salo and Kalliola 1991). Although most of the corroborative geomorphological evidence for dry periods in the tropical lowlands during the Pleistocene is now otherwise interpreted (Irion 1989; Colinvaux 1987), some new data look promising. There are also several other theories that attempt to explain aspects of present biogeography on the basis of past geological events, including river-channel formation and migration (Capparella 1988; Salo et al. 1986; Salo and Räsänen 1989), hypothesized massive flooding in south-western Amazonia (Campbell and Frailey 1984), and the formation of a putative giant Pleistocene lake in Amazonia (Frailey et al. 1988).

Mesoamerica

For its size, Middle America is even more complex geologically than South America (see Central America regional overview). Nuclear Central America, an integral part of the North American continent, reaches south to central Nicaragua. The region from southern Nicaragua to the isthmus of Darién in Panama is geologically younger and presents recent volcanism, uplift and associated sedimentation.

Like South America, the northern neotropics have a mountainous spine that breaks into separate cordilleras in the north. In general the Middle American cordilleras are highest to the north in Mexico, and lowest in Panama to the south-east. In Mexico, the geological picture is complicated by a band of volcanoes that bisects the continent from east to west at the latitude of Mexico City. This "eje volcanico transversal" is associated with the Mexican megashear, along which the southern half of the country has gradually moved eastward with respect to the northern half.

In southern Central America, volcanism has been most intensive in Costa Rica, which has two sections of its Central Cordillera reaching above treeline. In northern Costa Rica and adjacent Nicaragua the volcanoes become gradually reduced in size and more isolated from each other to the north. Similarly in Panama the Central Cordillera is over 2000 m high to the west near the Costa Rican border but only about 500 m high in most of the eastern part of the country. In central Panama, the Panama Canal cuts through a continental divide of only 100 m elevation, and in the San Juan River/Lake Nicaragua area of Nicaragua the maximum elevation is even less. For montane organisms, these interruptions in the cordillera represent major biological discontinuities.

The Yucatán Peninsula area of Mexico, Guatemala and Belize represents a geologically anomalous portion of Middle America. It is a flat limestone formation more like the Greater Antilles or Peninsular Florida than the mountainous terrain and volcanic soil of most of Middle America. Limestone is otherwise relatively rare in the continental neotropics, in contrast to many other parts of the world, with small outcrops like those in the Madden Lake region of central Panama or the Coloso area of northern Colombia being associated with peculiar floras. These areas, like the Yucatán Peninsula, tend to show distinctly Antillean floristic affinities, paralleling the geological ones.

Caribbean

The Antillean islands constitute the third geologic unit of the neotropics (see Caribbean Islands regional overview). The Antilles make up in geological complexity what they lack in size. The most striking geological anomaly is Hispaniola, which is a composite of what were three separate islands during much of the Cenozoic. In addition to being completely submerged during part of the midCenozoic, the southern peninsula of Hispaniola was probably attached to Cuba instead of Hispaniola until the end of the Cenozoic. Jamaica too was completely submerged during much of the mid-Cenozoic, and has a different geological history from the rest of the Greater Antilles, with closer connections to Central America via the nowsubmerged Nicaraguan Rise. Possibly a collision of the western end of the Greater Antilles island arc with Mexico-Guatemala fragmented its western end to form Jamaica.

Also phytogeographically and conservationally important, some of the Antilles have extensive areas of distinctive substrates. In addition to large areas of limestone, most of the Greater Antilles (Cuba, Hispaniola, Puerto Rico) have significant areas of serpentine and other ultrabasic rocks formed from uplift of patches of oceanic crust during the north-eastward movement of the Caribbean plate. The Lesser Antilles are small and actively volcanic. Most of the other smaller islands are low limestone keys with little or no geological relief. These patterns are clearly reflected in the Antillean flora. The most striking concentrations of local endemism occur in areas of ultrabasic rocks or on unusual types of limestone on the larger islands. The Lesser Antilles, Bahamas and other smaller islands have only a depauperate subset of the generally most widespread Antillean taxa.

Return to Top

Vegetation

The neotropics include a broad array of vegetation types commensurate with their ecological diversity. Along the west coast of South America are both one of the wettest places in the world - Tutunendó in the Chocó region of Colombia, with 11,770 mm of annual precipitation, and the driest - no rain has been recorded in parts of the Atacama Desert of Chile. The largest tract of rain forest in the world is in the Amazon Basin, and Amazonia has received a perhaps disproportionate share of the world's conservation attention. While the forests of Upper Amazonia are the most diverse in the world for many kinds of organisms, including trees as well as butterflies, amphibians, reptiles, birds and mammals, other vegetation types have equal or greater concentrations of local endemism and are more acutely threatened. In particular, the plight of dry forests and of Andean montane forests are beginning to receive increased attention. Some isolated areas of lowland moist forest outside of Amazonia also have highly endemic floras and are currently much more threatened than Amazonia. In the following paragraphs are sketched the major neotropical vegetation types, followed by a conservation assessment of each.

At the very broadest level, the lowland vegetation types of South America and the rest of the neotropics may be summarized as:

1. Tropical moist forest (evergreen or semi-evergreen rain forest) in Amazonia, the coastal region of Brazil, the Chocó and the lower Magdalena Valley, and along the Atlantic coast of Central America to Mexico.

2. Dry forest (intergrading into woodland) along the Pacific side of Mexico and Central America, in northern Colombia and Venezuela, coastal Ecuador and adjacent Peru, the Velasco area (Chiquitania) of eastern Bolivia, a broad swath from north-west Argentina to north-east Brazil encompassing chaco, cerrado and caatinga, and with scattered smaller patches elsewhere.

3. Open grassy savanna in the pampas region of north-eastern Argentina and adjacent Uruguay and southernmost Brazil, the Llanos de Mojos and adjacent pantanal of Bolivia and Brazil, the Llanos of Colombia and Venezuela, and the Gran Sabana and Sipaliwini savanna in the Guayana region.

4. Desert and arid steppe in northern Mexico, the dry Sechura and Atacama regions along the west coast of South America between 5°S and 30°S, and in the monte and Patagonian steppes of the south-eastern part of the Southern Cone of South America.

5. The Mediterranean-climate region of central Chile.

6. The temperate evergreen forests of southern Chile with an adjacent fringe of Argentina.

More complex montane formations occur along the Andean Cordillera which stretches the length of the western periphery of South America, in the more interrupted Central American/Mexican cordilleran system, in the tepuis of the Guayana region and in the coastal cordillera of southern Brazil.

Moist and wet forests

In general, forests receiving more than 1600 mm (Gentry 1995) or 2000 mm (Holdridge 1967) of annual rainfall are evergreen or semi-evergreen and may be referred to as tropical moist forest. In the neotropics, lowland tropical moist forest is often further subdivided, following the Holdridge life-zone system, into moist forest (2000-4000 mm of precipitation annually), wet forest (4000-8000 mm) and pluvial forest (over 8000 mm). Nearly all of the Amazon Basin receives 2000 mm or more of annual rainfall and constitutes variants of the moist forest. There are also several major regions of lowland moist forest variously disjunct from the Amazonian core area. These include the region along the Atlantic coast of Central America (extending into Mexico), the lower Magdalena Valley of northern Colombia, the Chocó region along the Pacific coast of Colombia and northern Ecuador, and the coastal forests of Brazil.

Lowland moist forest is the most diverse neotropical vegetation type, structurally as well as taxonomically. In most lowland moist-forest and wet-forest regions around a quarter of the species are vines and lianas, a quarter to a half terrestrial herbs (including weeds), up to a quarter vascular epiphytes and only about a quarter trees (Gentry and Dodson 1987; Gentry 1990b). To the extent that smaller organisms such as herbs and epiphytes may demand different conservation strategies than large organisms like trees (or top predators), this habitat diversity assumes conservation importance.

Diversity patterns are also important for conservation planning. There is a strong correlation of plant community diversity with precipitation - wetter forests generally are more botanically diverse. For plants the most speciesrich forests in the world are the aseasonal lowland moist and wet forests of Upper Amazonia and the Chocó region. For plants over 2.5 cm dbh in 0.1-ha samples, world record sites are in the pluvial-forest area of the Colombian Chocó (258-265 species); for plants over 10 cm dbh in 1-ha plots, the world record is near Iquitos, Peru (300 species out of 606 individual trees and lianas).

Concentrations of endemism do not necessarily follow those of diversity. Local endemism appears to be concentrated in cloud-forest regions along the base of the northern Andes and in adjacent southern Central America (cf. Vázquez-García 1995), and in the north-western sector of Amazonia where the substrate mosaic associated with sediments from the Guayana Shield is most complex (Gentry 1986a). Overall regional endemism in predominantly moist-forest areas is greatest in Amazonia, with an estimated 13,700 endemic species constituting 76% of the flora (Gentry 1992d). However many of these species are relatively widespread within Amazonia.

The much more restricted (and devastated, see below) Mata Atlântica forests of coastal Brazil have almost three-quarters as many endemic species (c. 9500) as Amazonia and similarly high endemism (73% of the flora) (Gentry 1992d). Moreover a larger proportion of the Mata Atlântica species probably are locally endemic.

On the other side of South America, the trans-Andean very wet to wet and moist forests of the Chocó and coastal Ecuador are also geographically isolated and highly endemic (cf. Terborgh and Winter 1982). Estimates of endemism in the Chocó phytogeographic region are c. 20% (Gentry 1982b). Probably about 1260 or 20% of western Ecuador's 6300 naturally occurring species also are endemic (Dodson and Gentry 1991). For the northern Andean region as a whole, including both the coastal lowlands of western Colombia and Ecuador and the adjacent uplands, Gentry (1992d) estimated over 8000 endemic species, constituting 56% of the flora. Moreover this is probably the floristically most poorly known part of the neotropics, perhaps of the world, surely with several thousand mostly endemic species awaiting discovery and description.

Dry forests

There are seven main areas of dry forest in the neotropics, and by some estimations this may be the most acutely threatened of all neotropical vegetations. The interior dry areas of South America are outstanding in their regional endemism, estimated at 73%. Two of the most extensive neotropical dry-forest areas represent manifestations of the standard interface between the subtropical high pressure desert areas and the moist equatorial tropics.

In Middle America, this area of strongly seasonal climate occurs mostly along the Pacific coast in a narrow but formerly continuous band from Mexico to the Guanacaste region of north-western Costa Rica. There are also outliers farther south in the Terraba Valley of Costa Rica, Azuero Peninsula of Panama, and even around Garachiné in the Darién (Panama), partially connecting the main Middle American dry forest with that of northern South America. These western Middle American dry forests are made up almost entirely of broadleaved deciduous species. In addition, the northern part of the Yucatán and large areas of the Antilles are covered by dry-forest variants. Most of the Caribbean dry forests are on limestone, and their woody species tend to be distinctively more sclerophyllous and smaller leaved than are the Pacific coast dry-forest plants. In the driest areas, both these types of dry forest tend to smaller stature and merge into various kinds of thorn-scrub matorral.

In South America, only the extreme northern parts of Colombia and Venezuela reach far enough from the Equator to enter the strongly seasonal subtropical zone. Floristically and physiognomically this northern dry area is very much like similarly dry areas of western Middle America. The strongly seasonal region of northern South America also includes the open savannas of the Llanos extending from the Orinoco River west and north to the base of the Eastern Cordillera of the Colombian Andes and the north slope of the Coast Range of Venezuela. Large areas of the lowlying, often poorly drained Llanos are seasonally inundated, especially in the Apure region.

The main area of tropical dry forest in South America is the chaco region, encompassing the western half of Paraguay and adjacent areas of Bolivia and Argentina, south of 17°S latitude. The "chaco" is physiognomically distinctive in being a dense scrubby vegetation of mostly smallleaved, spiny branched small trees interspersed with scattered large individuals of a few characteristic species of large trees. To the south, the chaco gives way to the desert scrub of the Argentine monte.

There is a distinctive but generally neglected area of dry forest at the interface between the chaco and Amazonia in Bolivia. The names Chiquitania and Velasco forest have been used locally in Bolivia to refer to this vegetation, which extends from the Tucuvaca Valley and Serranía de Chiquitos in easternmost Santa Cruz Department interruptedly westward to the base of the Andes and along much of the lower Andean slopes of the southern half of Bolivia. This region of closed-canopy dry forest is physiognomically similar to that of western Central America, with tall broadleaved completely deciduous (caducifolious) trees. Although it has been locally regarded as merely representing the transition between the chaco and Amazonia, it is a floristically and physiognomically distinctive unit that should be accorded equivalent conservation importance to the other major dry-forest vegetation types (Gentry 1994).

The chaco is adjoined to the north by two large and phytogeographically distinctive areas of dry forest, the cerrado and caatinga, which cover a small portion of easternmost Bolivia and most of the Brazilian Shield area of central and north-eastern Brazil. The typical vegetation of the "cerrado" region consists of wooded savanna with characteristically gnarled sclerophyllous-leaved trees with thick twisted branches and thick bark, widely enough separated to allow a ground cover of grass intermixed with a rich assortment of woody-rooted (xylopodial) subshrubs. The cerrado also includes areas where the trees form a nearly closed canopy ("cerradão"), and large open areas of grasses and subshrubs with no trees at all ("campo limpio" and "campo rupestre"). Although the cerrado is appropriately considered a kind of dry forest, some cerrado regions actually receive more rainfall than do adjacent forest regions; excess aluminium in the soil may be as important as the climate in determining its distribution.

The even drier forest of the caatinga of north-eastern Brazil extends from an appropriately subtropical 17°S latitude farther north to a surprisingly equatorial 3°S. Why this region should have such low rainfall remains poorly understood. Another climatic peculiarity is the irregularity of its rainfall, not only with low annual precipitation, but also with frequent years when the rains fail almost completely. The typical vegetation of the "caatinga" - relatively low, dense, small-leaved and completely deciduous in the dry season - is physiognomically similar to that of the chaco.

The final major South American dry-forest area is the coastal forest of north-western Peru and south-western Ecuador. Even more anomalous in its geographical setting than the caatinga, this dry-forest region is positioned almost on the Equator. The occurrence of dry forest so near the Equator is due to the offshore Humboldt Current. While similar cold-water currents occur along mid-latitude western coasts of other continents, the Humboldt Current is perhaps the strongest of these and is the only cold current reaching so near the Equator. The dry forest of coastal Peru and adjacent Ecuador is (or at least was, see below) physiognomically similar to that of western Central America, tall with a closed canopy of broadleaved completely deciduous trees.

There also are a number of scattered smaller patches of tropical dry forest and/or savanna in various interAndean valleys, around Tarapoto, Peru, the Trinidad region of Bolivia, Brazil's Roraima area, the Surinam/Brazil border region, on Marajó Island, and in the pantanal region of the upper Paraguay River.

Grasslands and deserts

Grasslands and deserts occupy smaller areas of the neotropics than they do in Africa or most higher latitude continents. The main grassland region of the neotropics is the pampas region between about 39°S and 28°S and encompassing most of Uruguay as well as adjacent eastern Argentina and southernmost Brazil. The other major grassland area is the llanos region of Colombia and Venezuela. Smaller predominantly grassland regions occur in north-eastern Bolivia (Llanos de Mojos) and the south-eastern Guayana region (Gran Sabana and Sipaliwini savanna). There are also areas with few or no trees and dominated by grasses in the cerrado and pantanal regions of Brazil, and scattered outliers associated with local edaphic peculiarities elsewhere.

None of the major grassland regions has many endemic species, in contrast to the campos rupestres of the Brazilian Shield and the Guayana area whitesand savannas, which have many endemics. This contrast is especially marked in southern Venezuela where some savanna patches have clay soils and a llanos-type flora of widespread species, whereas others have sandy soils and a flora of Amazonian affinities with many endemic species (Huber 1982).

The desert regions of Latin America are confined to northern Mexico, the monte (Morello 1958; Orians and Solbrig 1977) and Patagonian steppes of Argentina, and the narrow Pacific coastal strip of northern Chile and Peru. The 3500-km long South American coastal desert is one of the most arid in the world - most of it is largely devoid of vegetation. This region is saved from conservational obscurity, however, by the occurrence of islandlike patches of mostly herbaceous vegetation in places where steep coastal slopes are regularly bathed in winter fog. Although these "lomas" formations are individually not very rich in species (mostly fewer than 100 spp.), they have a very high degree of endemism due to their insular nature. The overall lomas flora includes nearly 1000 species, mostly annuals or geophytes. Diversity and endemism in the lomas formations generally increase southward, where cacti and other succulents are also increasingly represented (Müller 1985; Rundel et al. 1991).

Montane vegetation

The main montane-forest area of the neotropics is associated with the Andes. A major but more interrupted montane-forest strip is associated with the mountainous backbone of Central America. Venezuela's Cordillera de la Costa phytogeographically is essentially an Andean extension, although geologically distinct from the Eastern Cordillera of the Colombian Andes. The tepui summits of the Guayana Highlands, though small in area, constitute a highly distinctive and phytogeographically fascinating montane environment. The Serra do Mar along Brazil's south-eastern coast is mostly low elevation but has a few peaks reaching above treeline with a depauperate páramo-like vegetation.

The Andes may be conveniently recognized in three segments: northern - Venezuela, Colombia and Ecuador; central-Peru and Bolivia; and southern-Chile and Argentina. In general the northern Andes are wetter, the central and southern regions drier. The main biogeographic discontinuity in the Andean forests is associated with the Huancabamba Depression in northern Peru, where the extensive system of dry interAndean valleys of the Marañón River and its tributaries entirely bisects the Eastern Cordillera and is associated with a topographically complex region having unusually high local endemism.

Treeline in the tropical Andes occurs around 3500 m, depending on latitude and local factors. Above treeline, the wet grass-dominated vegetation of the Venezuelan, Colombian and northern Ecuadorian Andes is termed "páramo"; this drier vegetation, occurring from Peru to Argentina and Chile, is the "puna". Colombian and Venezuelan páramos are characterized by Espeletia (Compositae) with its typical pachycaul-rosette growth form. The vegetation above treeline of most of Ecuador and northernmost Peru, locally called "jalca" in Peru, is ecologically as well as geographically intermediate; although generally called páramo in Ecuador, this region lacks the definitive Espeletia aspect of the typical northern páramos.

While individual high-Andean plant communities are not very rich in species, many different communities can occur in close proximity in broken montane terrain. Thus the several high-Andean sites for which Florulas are available (Cleef 1981; Smith 1988; Galeano 1990; Ruthsatz 1977) have between 500-800 species, approaching the size of some lowland tropical Florulas.

The moist Andean slopes generally show a distinctive floristic zonation, with woody plant diversity decreasing linearly with altitude from c. 1500 m to treeline. Below 1500 m Andean forests are generally similar both in floristic composition and diversity to equivalent samples of lowland forest. There are also structural changes at different elevations. For example hemi-epiphytic climbers show a strong peak in abundance between 1500-2400 m, epiphytes are usually more numerous in middleelevation cloud forests, and the stem density of woody plants is usually greater at higher elevations (Gentry 1992a). While the northern Andes have cloud forest on both western and eastern slopes, increasing aridity south from the Equator limits cloud forest to an ever narrower band on the Pacific slope. South of 7°S latitude, forest on the western slopes of the Andes is restricted to isolated protected pockets, and the predominant slope vegetation becomes chaparral, thorn scrub and desert.

One of the most striking features of the Andes phytogeographically is the high level of floristic endemism. In part this is associated with the discontinuity of high-altitude vegetation types, which are strongly fragmented into habitat islands. In addition to microgeographic allopatric speciation related to habitat fragmentation, it seems likely that unusually dynamic speciation, perhaps associated with genetic drift in small founder populations, may be a prevalent evolutionary theme in Andean cloud forests (Gentry and Dodson 1987; Gentry 1989).

The combination of high local endemism (Gentry 1986a, 1993a; Luteyn 1989; Henderson, Churchill and Luteyn 1991) with major deforestation makes the Andes one of South America's conservationally most critical regions. As with the dry forests, the Andean forests have recently begun to receive greater conservation attention (Henderson, Churchill and Luteyn 1991; Young and Valencia 1992). Estimates of deforestation for the northern Andes as a whole are generally over 90%. Some areas are even more critical - perhaps less than 5% of Colombia's high-altitude montane forests remain (Hernández-C. 1990) and only c. 4% of the original forest persists on the western Andean slopes of Ecuador (Dodson and Gentry 1991). Most of the northern Peruvian Andes are similarly deforested (cf. Dillon 1994). Although relatively extensive forests still remain on the Amazonfacing slopes of Peru and Bolivia, much of this area is being actively deforested, in large part to grow "coca" (Erythroxylum coca) and opium poppy (Papaver somniferum).

Return to Top

Flora

From a conservation perspective, the neotropical region merits very special attention. Just as South America is sometimes called the "bird continent", the neotropics might well be termed the "plant continent" in deference to their uniquely rich botanical diversity (Table 50). If current estimates are accurate, the neotropical region contains 90,000-100,000 plant species, twice to nearly three times as many as in either tropical Africa or tropical Australasia (cf. Prance 1994).

The last great places for plant collecting are in the northern half of South America (J. Wurdack 1995, pers. comm.), which is two to four times less documented by herbarium specimens than elsewhere in the tropics (cf. Campbell 1989). Some of the main relatively unexplored areas (according to Wurdack) are, in Brazil: Serra de Tumucumaque (Tumuc-Humac Mountains), along the border with Surinam and French Guiana; slopes, especially the eastern slopes, of Pico da Neblina; in north-western Mato Grosso State, along the Linea Telegráfica; in Venezuela: slopes and talus forests of the tepuis; páramos west of Piñango (north of Mérida); eastern slopes to Páramo de Tamá (State of Mérida, near border with Colombia); in Colombia: Páramo de Frontino (west of Medellín); Cuatrecasas' headwater localities of collection in western Colombia, particularly in the Department of Valle del Cauca (cf. Cuatrecasas 1958); upper elevations of the Serranía de La Macarena (Department of Meta); in Ecuador: Cordillera de Los Llanganates (which is east of Ambato) (cf. Kennerley and Bromley 1971); Cordillera de Cutucú (Province of Morona-Santiago); Cordillera del Cóndor, along the border with Peru; in Peru: elevations above 700 m of the Cerros Campanquiz, which are mostly in the Department of Amazonas; the eastern cordillera in the Department of Amazonas, Province of Chachapoyas (e.g. the Cerro de las Siete Lagunas east of Cerro Campanario); portions of the Cordillera de Vilcabamba (which is north-west of Cusco), including the northern Cutivireni region (Villa-Lobos 1995); and in Bolivia: the easternmost Andes and granitic outliers in the Department of Santa Cruz.

Floristic diversity is very asymmetrically distributed in South America (cf. Table 51). If the nine phytogeographic regions recognized by Gentry (1982a) for the neotropics are taken as a basis, Central America with Mexico (Mesoamerica) and Amazonia are the richest in species, with each of these two regions having about a quarter of the neotropical total. At the opposite extreme, the Antilles have an estimated 9% of the total neotropical flora and the Caribbean coastal region of Colombia and Venezuela has only 8%. The minuscule area of the Guayana Highlands (above 1500 m) accounts for only c. 2.5% of the neotropical flora, but has one of the highest rates of endemism (65%) in the region (Berry, Huber and Holst 1995). The three main tropical South American dry areas together include a relatively low 11% of the neotropical species total. Intermediate levels of regional plant species richness are found in the Northern Andean and Southern Andean regions and the Mata Atlântica area of Brazil, which each have between 16-18% of the tropical flora of the neotropical region.

Regional endemism is greatest in Amazonia including lowland Guayana (76%), but almost as great in coastal Brazil (73%) and the chacocerradocaatinga dry areas (73%). In contrast, those two Andean subregions, Central America, and the Antilles have endemism levels of 54-60%, and the northern Colombia/Venezuela region only 24%.

Farther south in the Southern Cone of South America, the monte of Argentina is estimated to include 700 species with 5% endemism, and Patagonia 1200 species with 30% endemism. Chile as a whole has 5215 species (Marticorena and Quezada 1985; Marticorena 1990), with 1800-2400 in the Mediterranean-climate area of central Chile where endemism is high, perhaps greater than for any of the equivalent tropical regions.

The reasons for the unique floristic diversity of the neotropics as compared to Africa or tropical Australasia continue to be hotly debated. A popular theory is allopatric multiplication of species in habitat-island forest refugia during Pleistocene glacial advances (Haffer 1969; Prance 1973, 1982). Africa, which is higher and drier, would have had fewer refugia and more extinction. Tropical Asia was less affected, being buffered by the nearby ocean due to the island status of its components and by its proximity to a rain source from the Pacific (the world's largest ocean). Other theories, not necessarily mutually exclusive (cf. Terborgh and Winter 1982), focus on explosive speciation in the more extensive cloud-forest area of the neotropics (Gentry 1982a, 1989; Gentry and Dodson 1987); "Endlerian" speciation associated with habitat specialization in the uniquely complicated habitat mosaic of north-western and north-central Amazonia (Gentry 1986a, 1989; Gentry and Ortiz-S. 1993); speciation associated with riverine barriers to gene flow in the largest river system of the world (Capparella 1988; Ducke and Black 1953); or biogeographical phenomena associated with the Great American Interchange and stemming from the direct juxtaposition of Laurasian and Gondwanan elements via the Isthmus of Panama (Gentry 1982a; Marshall et al. 1979).

Return to Top

Social and environmental values, and economic importance

The indigenous groups (nations) of South America (Gray 1987) are varyingly diverse peoples who often partly depend directly on the natural environment for their biological and cultural well or survival. Their approximate presence is shown in Table 52.

As the site of one of the Vavilovian centres of domestication, South America has played an important role in providing plants useful to people. The Andean centre of domestication rivals the Indo-Malayan and Mediterranean areas as the region that has produced the most important crop plants. Tobacco, potatoes, grain amaranths, quinoa, peanuts, lima beans, kidney beans, tomatoes and perhaps sweet potatoes and pineapples all derive from the Peruvian Andes and immediately adjacent regions (Anderson 1952). Based on land-race diversity, western Amazonia was the centre of domestication of a series of less well-known but increasingly important crops, including "pejibaye" or peach palm (Bactris gasipaes), "biriba" or "anona" (Rollinia mucosa), "abiu" or "caimito" (Pouteria caimito), "sapota" (Quararibea cordata), "araza" (Eugenia stipitata), "uvilla" (Pourouma cecropiifolia) and "cubiu" or "cocona" (Solanum sessiliflorum) (Clement 1989). Of the 86 major crops and their more than 100 species included in a summary of crop plant evolution (Simmonds 1976), 24 crops are neotropical in origin either wholly (19) or partly (5).

Also, a host of South American forest plants are used locally but have not reached world commerce. Amazonia is especially rich in wild fruits (e.g. Duke and Vásquez 1994). For example around Iquitos, Peru, 139 species of forest-harvested fruits are regularly consumed, 57 of them important enough to be sold in the local produce market (Vásquez and Gentry 1989). There are a multitude of other uses for neotropical plants. Gentry (1992b) notes that 38% of the Bignoniaceae species of north-western South America have specific ethnobotanical uses and suggests that this could be extrapolated to 10,000 species with uses in this part of the world alone.

Many studies have shown that the direct economic value of such products can be very high (e.g. Peters, Gentry and Mendelsohn 1989; Balick and Mendelsohn 1992). In a single hectare of speciesrich tropical forest near Iquitos, 454 of the 858 trees and lianas of dbh 10 cm or more have actual or potential uses (Gentry 1986c), with the hectare of forest potentially producing US$650 worth of fruit and US$50 worth of rubber per year. If the 93 m³ of sellable timber worth US$1000 is included, the net present value of the hectare of forest is US$9000, far more than the net present value of managed plantations or cattle-ranching.

Additionally, the major role of forested areas in controlling erosion, recycling rainfall and as a carbon sink are now well known. As the territory with the largest tropical forest remaining in the world, South America plays a major role in providing such regional and planetary environmental services.

Return to Top

Loss, threats and conservation

Although the neotropical region has the most forest, it is also losing more forest each year than any other area of tropical forest (Myers 1982; Reid 1992). In western Ecuador only 4% of the original forest cover remains (Dodson and Gentry 1991). Much attention has focused on Brazil, which includes 48% of the South American area. Perhaps the most definitive satellite analysis of deforestation in Amazonia to date (Skole and Tucker 1993) indicates that as of 1988 only c. 10% of Brazilian Amazonia had been deforested, but if allowance is made for a 1-km edge effect, fully 20% of Brazilian Amazonia had been impacted. Deforestation in Rondônia alone has been c. 4000 km² per year, reaching almost 40,000 km² or 15% of the state by 1989 (Malingreau and Tucker 1988; Fearnside 1991). In coastal Brazil estimates of surviving forest range from 2% (IUCN and WWF 1982) to 12% (Brown and Brown 1992).

Burgeoning populations are the biggest factor in the ongoing losses, although political and economic instability in some areas, and short-sighted "development" programmes in other areas, also play significant roles. In most of the neotropics, unlike much of the Old World, commercial lumbering operations have played a relatively small role so far.

Conservational awareness throughout the region has increased dramatically in the past few years. Not only are increasing numbers of National Parks and similar conservation units being set aside, but there is also rapidly growing interest in the possibility of sustainable use of tropical forests as a conservation strategy.

Unfortunately many destructive and unsustainable uses of forest can masquerade behind the banner of sustainable use. Making this promising new concept fulfil its potential remains a major challenge. Similarly the growing appreciation of the potential value of biodiversity has been accompanied by too much political preoccupation and posturing about sovereignty over potential genetic resources.

Despite such problems, it is clear that the diversity of rain-forest plant life is intrinsically valuable. South America, botanically the richest continent, is also the greatest repository of potentially useful plants. Conservation of South America's plant diversity is clearly a world conservational priority.

Return to Top

Centres of plant diversity and endemism

The overview of South American species diversity is subdivided into eight large regions: Caribbean, Guayana Highlands, Amazonia, Mata Atlântica, Interior dry and mesic forests, Andes, Pacific Coast, and Southern Cone. The delimitation of these regions has been made mainly on geographical and floristic criteria. They do not represent phytogeographic regions in the strict sense, since in some cases it was necessary to join several geographical regions or subregions together, in order to allow a broad evaluation of the floristic resources of the South American continent.

The 46 sites that have been recognized as centres of plant diversity (see Map 36 and Table 53) of course do not represent all of the important centres in South America. Although some additional important sites are mentioned below, comprehensive coverage will require a great deal more effort.

This overview was written by Dr. Alwyn H. Gentry (Missouri Botanical Garden, P. O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.); With contributions by Olga Herrera-MacBryde (Smithsonian Institution, Department of Botany, NHB-166, Washington, D.C. 20560, U.S.A.); Dr. Otto Huber (Research Associate, Apartado 80405 Caracas 1080-A, Venezuela); Dr. Bruce Nelson (Instituto Nacional de Pesquisas de Amazonia (INPA), Caixa Postal 478, 69000 Manaus, AM, Brazil); and Dr. Carlos B. Villamil (Universidad Nacional del Sur, Departamento de Biología, Perú 670, 8000 Bahía Blanca, Argentina).

While this overview was in preparation, Dr. Gentry died on 3 August 1993 in Ecuador in a plane crash, during a survey of an important coastal site of plant diversity.

S.A. Regional Overviews
I. Caribbean (of South America) V. Interior dry and mesic forests
II. Guayana Highlands VI. Andes
III. Amazonia VII. Pacific Coast
IV. Mata Atlântica VIII.Southern Cone

References

Return to Top


North | Middle | South

CPD Home

Botany Home Page