Note: This website is no longer being updated and is being maintained for archive purposes by the Department of Botany, National Museum of Natural History, Smithsonian Institution.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

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



Link to South America Regional Overview



VI. (Tropical) Andes


The Andes are by far the most significant mountain range in the world's tropics. They span 64° of latitude: from 11°N, where the highest point in Colombia overlooks the Caribbean Sea near the northernmost tip of South America, to the continent's southern tip near 53°S in Tierra del Fuego. The highest mountain in the Western Hemisphere is Aconcagua (6959 m), on the border between Chile and Argentina in the Altoandina region. The highest mountain in the world's tropics is Huascarán (6768 m) in Peru's Cordillera Blanca. Even the highest mountain in the world by some calculations is Andean, since the summit of Ecuador's Chimborazo Volcano (6267 m) is farther from the Earth's centre than is the summit of Mount Everest, due to Chimborazo's more equatorial location and the Earth's oblate form.

The Andes essentially represent the upthrust leading edge of the South American continent, formed as the South American continental plate collides with the adjacent Nazca plate. Parts of the western Andes of Colombia and Ecuador have a more complicated origin as suspect terrane. In general the Southern Andes are oldest, the Northern Andes youngest. The range generally broadens to the north, at the Nudo de Pasto splitting into three separate cordilleras in Colombia. The Sierra Nevada de Santa Marta is somewhat isolated from the rest of the Eastern Cordillera, but is here considered part of the Andean region. Although the Eastern Cordillera is rather continuous with Venezuela's Coastal Cordillera, that area has a different geological history and is here considered part of the Caribbean region, despite obvious Andean affinities.

Since vegetation is continuous from the Andean slopes into the adjacent lowlands, delimitation of the montane Andean region is necessarily somewhat subjective. The Andean region is defined here by the (900-) 1000 m contour line to the west (following Gentry 1993a; Dodson and Gentry 1991; Forero and Gentry 1989) and by the 500 m contour to the east (following Gentry 1982a, 1993a). The valleys of the Magdalena and Cauca rivers that separate the three Colombian cordilleras are included in the Caribbean region, approximately following the 1000 m contour. Even though the upper Marañón River and its tributaries nearly sever the Andes in northern Peru, the entire Huancabamba Depression is above 500 m in elevation and is included here in the Andes.

Return to Top


Andean vegetational zonation is largely defined by altitude, thus tending to form long narrow bands along the slopes of the cordillera. The most striking differentiation is at treeline, with the lower slopes of the Andes covered by forest and the upper slopes by herbaceous vegetation usually dominated by grasses. Treeline in the tropical Andes generally occurs around 3200-3500 m and is associated with an average annual temperature of c. 6°C (Holdridge 1967; van der Hammen 1974; Terborgh 1971; RivasMartínez and Tovar 1982). Tree-line tends to be slightly higher in the more subtropical Bolivian Andes, and gradually descends in the Altoandina region to near sea-level at the southern tip of the continent. On most of the dry Pacific-facing slopes of the central and southern Andes, trees are limited to tiny remnants in protected microsites and the differentiation between highAndean and mid-elevation vegetation is much less distinct.

Precipitation in the high Andes generally decreases from north to south, with the high-Andean grass-dominated vegetations falling into three major types correlated with the precipitation differences. The wettest northern high-Andean formation, called "páramo", typically has the distinctive thick-stemmed (pachycaul) Compositae genus Espeletia as a co-dominant with grasses. The grass-dominated vegetation of the drier southern high Andes is called "puna". Páramo with Espeletia reaches south only to northernmost Ecuador; puna occurs from near the Huancabamba Depression (and Cajamarca) in north-eastern Peru southward to northern Argentina. The intermediate precipitation area, called "jalca" in north-eastern Peru and páramo in Ecuador, is floristically similar to the northern Andean páramos except for lack of the characteristically dominant Espeletia. There are also limited areas of a similar jalca-like wet puna along the upper forest margin of the eastern escarpment of the Peruvian Andes.

Peculiar growth forms characterize many puna and páramo plants. In addition to the pachycaul-rosette growth form of Espeletia, cushion plants are common, as are sessile rosette growth forms. The woody plants are mostly dense subshrubs with very small sclerophyllous leaves.

The forested tropical Andean slopes ("selva nublada" in Colombia, "ceja de la montaña" in Peru, "yungas" in Bolivia) form a continuous band along the eastern Andean slopes roughly between 500-3500 m. This forest band is usually subdivided into about three altitudinal zones: premontane or submontane forest extending up to (1000-) 1500 m, lower montane or tropical montane forest between c. 1500 m and c. (2300-) 2500 m, and upper montane forest between c. 2500 m and (3200-) 3500 m. Sometimes a fourth altitudinal zone is recognized, subpáramo or prepuna, consisting of the usually rather narrow shrub-dominated transition zone from forest to non-forest that generally occurs between 3000-3500 m (Cleef et al. 1984; Frahm and Gradstein 1991; van der Hammen 1974; Cuatrecasas 1958; Terborgh 1971). The correlations between these vegetational types and the altitude are not constant, and the zones generally occur at lower altitudes on narrower cordilleras and outlying ridges - the well-known Massenerhebung effect (Flenley 1995). The transition between premontane and montane forest around 1500 m elevation is usually associated with the dew-line or cloud-line and is characterized especially by an abruptly greater density of both vascular and non-vascular epiphytes (Terborgh 1971; Cleef et al. 1984; Frahm and Gradstein 1991). The transition from lower montane to upper montane forest around 2500 m tends to be associated with the 12°C isohyet of average annual temperature (Holdridge 1967; RivasMartínez and Tovar 1982). Above this altitude, soil becomes strongly peaty, vascular epiphytes are reduced, and bryophytes become much more prevalent, forming thick moss layers on the trees.

On the western slopes of the Andes similar cloud-forest formations occur south to central Ecuador, although their zonation is much less obvious in the super-saturated conditions of the Chocó region, where some cloud-forest features and elements extend down to sea-level (Gentry 1986b). In southern Ecuador and northernmost Peru, inland from the Humboldt Current and thus drier, the cloud-forest band becomes more restricted altitudinally. South of 5°S latitude in Peru, forest is restricted to small isolated patches in the most protected valleys. Farther south, the lower limit of these forest patches recedes upward. At the latitude of Lima (12°S), relict forest patches occur only at very high altitudes over 3000 m; farther south they are altogether lacking.

In addition to the prevalent cloud-forest vegetation, the Andes include important areas of dry shrub dominated vegetation in various inter-Andean valleys and on the Pacific-facing western slopes of Peru. The largest and most important of the inter-Andean dry areas is the Huancabamba Depression of northern Peru and southern Ecuador, where the main Andean cordilleras are extensively dissected by the Marañón River and its major tributaries. Similar but less extensive shrubby areas occur along the upper Apurimac and Huallaga rivers in Peru, in the Cochabamba area of Bolivia and on a smaller scale in Ecuador and Colombia.

Return to Top


In the context of South America, the Andes are unusual in their prevalence of Laurasian plant families and genera, many of which do not extend into the adjacent tropical lowlands (Gentry 1982a). These northern elements, mostly newly arrived from the north in Plio-Pleistocene times as the Isthmus of Panama closed, mixed complexly with tropical Gondwananderived and southern subtropical taxa like Podocarpus, Drimys, Panopsis and Weinmannia to give rise to today's Andean flora. Perhaps a quarter of the genera of the puna and páramo belong to the old australantarctic flora (van der Hammen and Cleef 1983). In the Southern Cone, the puna together with the Altoandina and Patagonia constitute the Andino-Patagonian Dominion (Cabrera and Willink 1973). The southern taxa are less prevalent in the Andean forests, which are largely a mixture of the Laurasian and tropical Gondwanan elements.

The floristic composition of the high-Andean páramo and puna vegetation is remarkably consistent at the familial level, with Compositae always the largest family, nearly always followed by Gramineae. The other most diverse families are Scrophulariaceae, Orchidaceae, Leguminosae, Melastomataceae, Solanaceae, Caryophyllaceae, Rosaceae, Cyperaceae, Cruciferae (and sometimes Ericaceae - e.g. in the Puracé, Colombia area), with over half of the species of a given high-Andean site belonging to these dozen most speciose families, nearly all of which tend to be of Laurasian origin and poorly represented in the tropical South American lowlands (Vareschi 1970; Cleef 1981; Smith 1988; Galeano 1990; Ruthsatz 1977). Senecio (Compositae) is consistently the most speciose high-Andean genus; Solanum, Calceolaria and Valeriana are also notably speciesrich in the Andes (Gentry 1993a). Several of these genera are famous for their local endemism, presumably associated with the archipelago-like nature of the high-Andean vegetation.

A similar combination of predictable familial and generic composition combined with strong local endemism in certain groups characterizes the Andean cloud forests. The premontane vegetation zone below 1500 m is largely tropical in its familial and generic composition, contrasting with the middle and upper elevation forests which are floristically distinctive (Gentry 1992b).

The middle elevation forests are usually dominated by Lauraceae, Melastomataceae and Rubiaceae. They are also characterized by a small group of tree genera each belonging to a different family, including Saurauia, Ilex, Alnus, Brunellia, Styloceras, Viburnum, Hedyosmum, Clethra, Weinmannia, Billia, Juglans, Prunus, Meliosma, Styrax, Symplocos and Gordonia. Most of these genera belong to Laurasian families and most are absent or very poorly represented in the lowlands (Gentry 1992b). If complete species lists are compared, predominantly epiphytic taxa like Araceae (especially Anthurium), Ericaceae, Peperomia (Piperaceae), Orchidaceae (especially Epidendrum, Maxillaria, Pleurothallis) and ferns (especially Polypodium) are among the most diverse components of middle elevation Andean forests. Considering the shrubs and trees of upland Ecuador, Compositae, Melastomataceae, Ericaceae, Solanaceae and Rubiaceae are the largest families and Miconia, Piper and Solanum the largest genera (Ulloa-U. and Jørgensen 1993). These typically Andean taxa tend to be very poorly known taxonomically; some are very prone to local speciation (Gentry and Dodson 1987; Gentry 1992b).

The highest altitude Andean forests show similar patterns but with different taxa. Forests near treeline are dominated by Compositae, Ericaceae, Myrsinaceae, Melastomataceae and Rosaceae, and include genera like Drimys, Cervantesia, Vallea, Hesperomeles, Polylepis, Escallonia and Myrica that are poorly represented or absent in lower altitude forests.

It is conservationally relevant that although there is a general tendency for diversity to decrease with increasing altitude (Gentry 1988), individual páramo and puna florulas can have 600-800 species and be almost as species-rich as many lowland tropical florulas (Smith 1988; Galeano 1990; Ruthsatz 1977). Presumably this richness is due to the unusually close juxtaposition of different habitats in montane environments. Similarly the species lists for individual altitudinal transects in Colombia (by the EcoAndes project) include c. 1200 vascular plant species, again reflecting the telescoping of habitats. Young and León (see Data Sheet SA37) estimate that the 1500-3500 m strip of eastern Peruvian Andean forest includes 14% of Peru's flora in 5% of the country's area. Expanding the slope area to the 500 m contour, they estimate that about half of Peru's species might be included in the 20% of the country represented by the eastern Andean forests. Similarly, Balslev (1988) estimates that half of Ecuador's species occur in the 10% of the country represented by middle-elevation (900-3000 m) Andean forests.

Sparse data suggest that endemism generally increases in montane forests. For example Balslev (1988) found in his sample of recently monographed Ecuadorian plants that 40% of the high-altitude (over 3000 m) species and 39% of the middle-elevation (900-3000 m) species are endemic to Ecuador, compared to 16% of lowland species. Half of the lowland species are widespread, as compared to one-quarter of the mid-elevation taxa and only one-sixth of the high altitude taxa.

There are regional peculiarities between different areas of Andean vegetation. Thus it is imperative to apportion conservation efforts to include representation of many different, floristically distinctive vegetations in the Andean region. As examples, Quercus is dominant in many high-Andean forests in Colombia, but absent farther south. The western slope of the Colombian Andes is especially rich in Araceae, mostly still undescribed, and Ericaceae, of which Luteyn (1989) reported finding 41 species, 12 new to science, in a few days' collecting along a single 5-km stretch of road. Isolated massifs like Colombia's Sierra Nevada de Santa Marta tend to have unusually high rates of endemism (e.g. 11 endemic Diplostephium spp., Compositae) but relatively depauperate floras; Santa Marta even has endemic genera like Cabreriella, Castanedia and Raouliopsis (Compositae), but only half as many páramo genera as the Eastern Cordillera of the Colombian Andes (Cleef and Rangel 1984). The relatively young Eastern Cordillera has a much richer high-Andean flora and greater endemism than the older, but more actively volcanic Central Cordillera (Rangel, pers. comm.). Solanaceae are more dominant in relatively dry cloud forests in north-western Peru (Gentry 1992b). The dry western slopes of the Peruvian Andes and the dry interAndean valleys are dominated by shrubby and herbaceous Compositae.

To the extent that high-Andean endemism reflects broken local topography, such patterns can be very useful for conservation planning. For example, Fuchsia (Onagraceae) has its greatest diversity in the Northern Andes' Central Cordillera of Colombia and into Ecuador, but much higher endemism in the more dissected terrain of the Peruvian Andes (Berry 1982). Telipogon (Orchidaceae) has its greatest concentrations of locally endemic species in the more complicated topography of the Nudo de Loja and Nudo de Pasto than in intervening regions (Gentry and Dodson 1987). In a 200-km² area of the Huancabamba region in Peru's Cajamarca and Amazonas departments are concentrated 63 of the 181 neotropical species of Calceolaria (Scrophulariaceae), 21 of them with distributional areas less than 50 km across (Molau 1988). Complexly dissected regions like the Huancabamba area and the Nudo de Pasto are clearly of exceptional conservational significance.

Return to Top

Species of economic importance

The Peruvian and Bolivian Andes constitute one of the major Vavilovian centres of domestication. This Andean centre of domestication rivals the Indo-Malayan and Mediterranean regions as the area that has produced the world's most important crops. Thus it is hardly surprising that the Andean region has a large number of economically important plant species. Many of the high-Andean food plants are tubers. In addition to the potato (Solanum tuberosum) and several wild relatives, "ollucos" (Ullucus tuberosus) and "oca" (Oxalis tuberosa) are local staples with broader commercial potential. Similarly, the native grain amaranths (Amaranthus) have been singled out as more protein-rich than grain cereals. Edible fruits from the region include several species of Rubus, Solanum (especially S. quitoense), Cyphomandra, Lycopersicon, Carica and Passiflora. There are many medicinal plants in the Andes, with the famous antimalarial Cinchona one of the best known; several CPD Data Sheets especially emphasize how many high-altitude páramo and puna plants have medicinal uses.

Important timber trees from the Andes include Cedrela, Juglans, Quercus and Podocarpus. With so many epiphytes and spectacular small herbs, the Andean flora also has significant, as yet largely unexploited potential for horticulture. An example is the red-flowered ladyslipper orchid Phragmipedium besseyi which was discovered on the eastern slopes of the Peruvian Andes not too many years ago, causing a major horticultural stir, with the first plants selling for several thousand US$ each (C. Dodson, pers. comm.); it continues to be popular now that it is produced in cultivation. There also are numerous Andean species used in local handicrafts. A particularly interesting example is "barniz de Pasto", obtained from the stipule exudate of a species of Elaeagia (Rubiaceae) and used in a kind of traditional lacquer-work in the Pasto region of southern Colombia.

In the Sierra Nevada de Santa Marta Data Sheet (SA25), Rangel and Garzón cite an unpublished 1987 thesis by E. Carbonó that makes abundantly clear the economic importance of Andean plants for one group of Andean natives. The Kogui Amerindians of the Sierra Nevada consume 32 species and use 81 for medicine, 44 in construction, 12 for dyes and 36 mythologically.

Perhaps the most important economic use of Andean plants is to provide ecological services such as erosion control and water retention. While difficult to quantify, such ecological services are clearly of paramount value in steep montane terrain. Unfortunately this value tends to be more appreciated once these services have been lost. The current crisis in electricity-generating capacity throughout the Andean region, due to lack of sufficient water to fill the hydroelectric reservoirs and run the turbines, is now widely appreciated to be a direct result of deforestation and other destruction of natural Andean ecosystems.

Return to Top


The threats to biodiversity in the Andean region that are most often cited in the CPD Data Sheets are especially: clearing (legally and illicitly) of too steep land for agricultural cultivation, overgrazing and the wide-scale burning associated with cattle husbandry in páramo and puna, habitat loss from fuelwood collecting at the highest altitudes where trees are rare, and pollution from mining.

At the highest altitudes, cultivation of potatoes and other tubers to feed a burgeoning population may be the biggest threat. While even larger (and lower) areas of Andean forest were destroyed for coffee cultivation, currently little expansion of the coffee growing area seems to be taking place. However, in the cloud-forest areas increasing cultivation of coca in Bolivia and Peru, mostly to produce cocaine for the European and North American markets, and of marijuana (Cannabis sativa) and opium poppy in Colombia, also for export markets, have led to destruction of huge areas of forest. Ironically, attempts to control the drug crops have had no effect on the drug supply, but have had a devastating effect on natural ecosystems, both directly by indiscriminate spraying of chemicals from airplanes, and indirectly, since drug growers merely clear more forest for their preferred crops on steeper slopes in more inaccessible sites.

In all the Andean countries, failure to appreciate the constraints on agricultural development imposed by steep mountainous terrain has led to major losses of biodiversity through inappropriate colonization schemes by shortsighted government agencies, funded typically by international lending agencies.

Andean forests are among the most threatened of all tropical forest vegetations. Some entire forest types, like the Podocarpus forests that used to cover significant parts of the Andes, have already all but disappeared. In Colombia estimates suggest that less than 10% of the Andean forests remain intact (Henderson, Churchill and Luteyn 1991) - perhaps even less than 5% of the high-altitude upper montane forest. In Ecuador almost nothing is left of the natural forests of the central valley and only 4% of the forests on the western Andean slopes (Dodson and Gentry 1991). North-western Peru north of the Huancabamba Depression probably retains even less intact forest, with the last Podocarpus forests of the Jaén/San Ignacio region currently being cut. Although there are still areas of relatively intact forest on the eastern slopes of the Andes of Ecuador, Peru and Bolivia, all three countries have active road-building programmes and rampant deforestation in this region.

Return to Top


The natural restriction or contraction of different vegetation types to relatively small areas is a general trend in the Andes, suggesting that appropriately planned conservation units on the Andean slopes would be likely to include more species than equivalent areas almost anywhere else.

A number of large and potentially significant conservation units exist in the Andean region, but only some of them are effective. Most of the National Parks, several of which are featured in CPD Data Sheets, include extensive areas of páramo or puna as well as Andean-slope cloud forest. A glaring omission from the park systems is the typically narrow strip of relatively wet fertile-soil forest that occurs in the 500-1000 m band. Most of the large upland parks either do not extend down to a low enough elevation to include this forest type or include it just on paper. Perhaps only the Sumaco region in Ecuador (CPD Site SA38) and Manu National Park in Peru (see CPD Site SA37) protect this critical zone, and even in Manu settlers are already making incursions. There are only a few places left in the Andes where the whole elevational sequence might be protected, perhaps including the Tambopata/Candamo region in southern Peru (see CPD Sites SA37 and SA10) and the Madidi region in Bolivia (CPD Site SA36). Clearly such areas need immediate evaluation and appropriate conservation, before colonization makes preservation unfeasible or impossible.

Perhaps a worse problem is that many Andean parks are "paper parks". The typical protection provided a National Park in the Andes is termed passive protection by Young and León (Data Sheet SA37) - inaccessibility rather than the park status provides a modicum of protection. They note that there are 20 park guards for the 20,000 km² conserved in the three relatively effective parks in the eastern Andes of Peru, an impossible total responsibility of 1000 km² per guard. They estimate that an additional 6000 km² of various other protected areas and reserves have no actual protection. To this total may be added Cutervo NP, which has been mostly clear-cut and had not a single guard when Gentry last visited it. Colombia's north-western Paramillo NNP has a single guard completely lacking in equipment or support and so without capability to even marginally effect the ongoing deforestation that has virtually eliminated natural vegetation from the great majority of the park. If National Parks are to be a useful mechanism for conservation of Andean biodiversity, major increases in their budgets and major changes in their administration will be necessary.

Protection of watershed forests in connection with hydroelectric generation is effective in some parts of the Andes, most notably of the Anchicaya watershed in Colombia by the Corporación Valle del Cauca (CVC); one of the advantages is that such protection is apparently politically more acceptable than is biodiversity protection per se. Private reserves like La Planada Nature Reserve (Stone 1985) in Colombia and Maquipucuna Reserve in Ecuador (Sarmiento 1995) seem to be more effective than most of the official parks, perhaps because the former tend to have relatively good financial support, but unfortunately the area of such reserves at present is minuscule. Several CPD Data Sheet authors emphasize the potential for ecotourism in the scenically spectacular Andes, although to date tourism has had little economic impact other than in the Cusco/Machu Picchu area of Peru. Clearly novel conservation strategies need to be implemented very rapidly in the Andean region; in many areas it is probably already too late.

Return to Top

Centres of Plant Diversity and Endemism

a. Paramo with Espeletiinae
(e.g. Espeletia)

SA25. Sierra Nevada de Santa Marta

SA26. Sierra Nevada del CocuyGuantiva

SA27. Páramo de Sumapaz region

SA28. Region of Los Nevados Natural National Park

SA29. Colombian Central Massif

Colombia, Ecuador

SA30. Volcanoes of Nariñense Plateau

b. Paramo without Espeletiinae to puna


SA31. Páramos and Andean forests of Sangay National Park

Peru, Ecuador

SA32. Huancabamba region


SA33. Peruvian puna

Argentina, Chile

SA34. Altoandina

c. TucumanBolivian region


SA35. Anconquija region

d. Eastern slope


SA36. Madidi-Apolo region


SA37. Eastern slopes of Peruvian Andes


SA38. Gran Sumaco and Upper Napo River region


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

Return to: South America Overview

North | Middle | South

CPD Home  

Botany Home Page | Smithsonian Home Page