This text is a modified and slightly expanded version of a section commissioned for the encyclopedia 'Biosphere' (Biosfera) by Encyclopedia Catalana, published in Barcelona, Spain in 1994.

THE DISTRIBUTION AND VARIETY OF EQUATORIAL RAIN FOREST

by Jonathan Adams, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA

The natural distribution of equatorial rainforest. The global distribution of equatorial rainforest is closely tied to the warm, moist climates that occur near to the Equator. Ecologists recognize this class of forests from a characteristic assemblage of features; the trees tend to have wide buttress roots that splay out near the ground, and the leaves tend to be large, evergreen and laurel-like, with an elongated tip (a 'drip tip') on the end of each leaf. In many areas the equatorial forest grades almost imperceptably over hundreds of kilometres into the vegetation of cooler and drier climates - with drip tips and buttress roots becoming progressively less common - so that there is no single point where one can objectively say that equatorial forest ends and another vegetation type begins. A gradation to colder climate vegetation occurs much more rapidly on the tropical mountains; above a few hundred metres altitude the character of the forest starts to change, until beyond about 2000m above sea level it is seen to be different enough to warrant the separate name of 'montane forest'.

The equatorial rainforest climate. Near to the Equator, the intense energy input from the sun produces the intertropical convergence zone (the ITCZ), a convection zone of rising air that loses its moisture as frequent, intense rainstorms. In equatorial areas that are relatively isolated from the sea winds that carry water vapour inland, there are breaks in the rainforest belt. Likewise, travelling north and south away from the Equator one generally finds a decrease in rainfall, as the influence of the ITCZ becomes weaker. In these drier places, wherever the annual rainfall falls below about 1600mm with an intense dry season during part of the year, the rainforest gives way to either monsoon (seasonal) forest, open woodland or grassland. However, the actual limits of rainforest do also vary greatly with soil type, and the amount of disturbance from humans and fires.

Where the climate remains moist enough to support equatorial rainforest outside the main tropical belt, what finally puts paid to it are low temperatures and frosts. In a few places in the world, such as southern China, a belt of moist evergreen forest continues north or southwards well away from the Equator (as far as 26 degrees North in south-western China), nourished by moisture-bearing winds from the oceans. The ultimate limits to equatorial rainforest in these areas seem to be related to the mean temperature of the coldest winter month, with the final (and rather subjective) boundary between equatorial rainforest and temperate rainforest being drawn on maps at around the point where there is a significant probability of occasional frosts occurring on cold winter nights.

A common evolutionary heritage. Despite their wide geographical separation, there are strong floristic similarities between all the world's rainforest regions that reflect a common origin. Many of the same families and even genera of plants occur in rainforests on different continents, separated by thousands of miles of ocean water.

Some of the closest floristic similarities between rainforest regions are probably due to the exceptional dispersal abilities of certain plants that may have had their seeds carried across great distances by birds or ocean currents within the last few million years. This mechanism may explain how, for example, a single species of the highly specialised carnivorous pitcher plant Nepenthes occurs in the rainforests of Madagascar, whilst around 70 species of this same genus occur in the Far East.

However, to a large extent the causes of these common floristic links must lie in the more distant geological past, when several of the world's rainforest regions were in much closer contact with one another. There is fossil evidence indicating that that many of the world's widely separated regions of rainforest were once linked together into continuous forest belts during times of moister and warmer climate. For example, during the early Tertiary Period around 55 million years ago, equatorial rainforests seem to have formed an almost continous belt from Africa across southern Europe and southern Asia to the Far East. Though this rainforest belt was broken up intermittantly by shifting narrow seaways, one can imagine the African and far Eastern rainforest regions would have exchanged newly evolving types of plants rapidly and efficiently. This linkage seems to have been broken by about 30 million years ago, as global climates became cooler and drier, but the closer common heritage between the African and far Eastern regions still leads to them being classified together as the palaeotropical (= old world tropics) realm.

A more fundamental set of resemblances between the rainforest regions exists because of the geological history of plate tectonics, with different land masses that were once linked having been carried away from one another. During the Cretaceous (around 140-65 myr ago), the time when the flowering plants (angiosperms) were undergoing their main diversification and rise to dominance, the land masses of the Southern Hemisphere were still more-or-less in the form of the ancient southern supercontinent, known as Gondwanaland.

During the key phase of diversification of tropical angiosperms from about 140 and 65 myr ago, Africa and South America (previously parts of Gondwanaland) were in the process of rifting apart but were still quite near to one another as shown in this animation by Rowley & Ziegler from the University of Chicago. Many of the ancestors of what are now tropical rainforest plant families would have been able to cross between the continents. Sometimes these newly evolving forms would have spread directly across the last remaining landbridges or even drifting fragments of continent present during the rifting process. No doubt over the millions of years there were also many occasional long-distance dispersal events across tens or hundreds of miles of sea water, aided by shorter distance 'island hopping' across the scattered groups of volcanic islands that certainly would have existed in such a geologically active region. As a result of the combination of overland and overwater spread between Africa and South America, we find many of the same plant families (such as the Euphorbiaceae, Meliaceae, Moraceae and Sapotaceae) occurring between widely these separated continents, and indeed all across the world's rainforest flora. Often, the overlay of subsequent evolution has given an east-west split between separate subgroups shared between Africa and Asia on one hand, and South and Central America on the other. The sub-family level differences that have evolved are sufficient for the rainforests of the Americas to be classified separately as the neotropical ( = new world tropics) realm.

The Australasian region (including Australia and New Guinea along with other smaller fragments) separated off from the southern part of Gondwanaland at about the same time as South America, and evolved its own distinctive flora from the groups of plants that had managed to reach there. Moist tropical environments were not present in the Australasian region at the time that it split off from the other continents, but arose afterwards during the Tertiary as it moved north towards the equator. Consequently the region evolved its own very distinctive rainforest flora from the plant groups that happened to be present, including an important role for certain ancient conifer groups that had dominated the Gondwanaland forests before flowering plants existed. However, many of the same families and even genera of angiosperms and conifers that evolved to fill the Australian rainforest realm do also occur elsewhere in the world because of the ancient Gondwanaland link.

The same general patterns are reflected in the faunas of the rainforest regions, but not always for the same reasons. Thus for example the monkeys of Africa and Asia together form a linked group (known as the catarrhines) distinct from those of the Americas (the platyrhines), but it is known that monkeys evolved relatively recently in geological time - after the major breakup of Gondwanaland - so it is unlikely they managed to get directly across the widening Atlantic. Instead, they most likely spread across from the old world tropics by forested land links that existed between northern Eurasia and North America in the mid-Tertiary period. They would then have spread southwards across the land bridge of central America, eventually reaching South America. After the Northern Hemisphere link was broken, the new and old world monkeys evolved into distinct groups. A similar east-west pattern is found for the bats of the new and old worlds, with the fruit eating niche of the African and Asian 'flying foxes' of the Macrochiroptera being filled in South America by a distinct and less specialised group descended from the insect-eating Microchiroptera.

As the belt of warm, moist climate that once linked the African and Asian rainforests is now long gone, evolution has also taken different directions between these two regions. Thus, the Asian forests are today dominated by trees of the family Dipterocarpaceae, whilst the handful of African dipterocarps are savanna and woodland trees, belonging to a quite distinct subfamily.

During the last 30 million years since the mid-Tertiary, the history of the world's equatorial forests has been one of further decline. A trend towards aridity in Africa and Australia has left only small remnants of the much more extensive blocks of rainforest that once existed there. Often these isolated pockets of rainforest survive hundreds of kilometres away from one another, in coastal strips or on mountain slopes where the climate remains moist enough. During each of the cool, dry glacial phases which have occurred repeatedly over the last two-and-a-half million years, these forest remnants seem to have been even further reduced below what exists at present, perhaps almost to vanishing point. Likewise, in the Americas, there is strong evidence for major arid periods having occurred in the rainforest regions during the past few million years.

As an additional complication to the story, two of the world's main rainforest realms are currently in the process of merging into one another. As the isolated Australasian plate drew northwards during the early Tertiary (about 40-30 million years ago), distinctively Asian species of plants started to appear in its flora as it it approached the edge of the Asian plate. Following the eventual collision of the Asian and Australasian continental plates beginning about 20-30 million years ago, dispersal of plants and animals across the relatively narrow straits that now separate these two continental masses has been a gradual but ongoing process. A legacy of their previously distinct histories remains in the form of Wallace's Line (named after A.R. Wallace, the 19th century naturalist who first gave a detailed account of it), the zone of faunal and floral discontinuity that closely follows the meeting point of the plate boundaries. Thus, after diverging since the dawn of the angiosperm era, these two rainforest regions are now engaged in another vast ecological and evolutionary experiment.

Similarities and differences. As a result of this complex history of climatic change and continental drift, the world's rainforests show many regional floristic and structural differences that can only be put down to chance. Yet the differences are superimposed upon a background of striking similarity, produced by the same evolutionary pressures producing similar growth forms amongst many independent lineages of plants, giving equatorial rainforests in any part of the world the same characteristic general appearance.

Thus in Asia, the role of gap colonizer in disturbed rainforest sites is filled by the large-leaved fast growing trees of the genus Macaranga (family Euphorbiaceae), whilst in the Americas it is filled by the superficially rather similar Cecropia (family Urticaceae). In contrast, whilst epiphytes growing in the branches of canopy trees occur in all rainforest regions, they are far more abundant in the American tropics where bromeliads (Bromeliaceae) have successfully diversified to take up the role.

Large 'emergent' trees which project high above the main canopy occur in all the world's main rainforest regions, but different tree families predominate in different places. In south-east Asia, large trees of the family Dipterocarpaceae are particularly important as emergents, whereas in other regions the commonest emergents belong to a range of other families.

Palms occur as part of the forest canopy in all rainforest regions, but for some reason they are far more abundant in a range of ecological roles in South American rainforests than elsewhere.

The list of similarities and differences could go on and on. Looking across the range of the world's rainforests, it is obvious that there must be certain common evolutionary pressures which are responsible for the striking similarities between unrelated groups of plants growing in different parts of the world. However, there is also a certain amount of leeway which allows particular types of plants with distinctive attributes and aptitudes to come to dominate the structure of the rainforest in a particular region.

The global distribution of species richness in equatorial forests. Equally as striking as the convergences in plant form within and between rainforest regions are the results of recent studies on numbers of species present. It has long been known that equatorial rainforests around the world tend to be exceptionally rich in species of animals and plants, with some areas being richer than others. It is also evident that within each region, certain groups of plants have undergone explosive diversification due to the vagaries of history and opportunity. Thus for example there are 700 species of screwpalms (Pandanus) in Eastern Asia, but only a handful of species in Africa. However, at the broadest scale there may also be a suprising amount of order in terms of the patterns of species richness in the rainforest ecosystem

Recent studies have shown that tree species richness within each tropical forest region tends to follow the trend of mean annual rainfall. In the drier parts of rainforest regions (around 1000-2000 mm), the number of tree species per 0.1 hectare starts at around 50-100 species, and climbs to over 250 species in areas of very high rainfall (about 6000mm annually) and very weak seasonality. Above this level (in areas where as much as 9000 mm of rain can fall in a year), the curve of rising tree species richness tends to flatten off.

It is striking that despite tens of millions of years of isolation and independent evolution, the richness of tree species between different tropical regions of similar climate remains fairly constant. Thus, on the data gathered so far, a hectare of primary rainforest in Africa can be expected to have roughly the same number of species as one in Asia or South America existing under similar mean annual levels of rainfall (although the pattern of seasonality in rainfall does affect this as well). This pattern contradicts the previously accepted notion that the African rainforests are relatively poor in numbers of species because of their geological history of arid phases, with these having occurred more severely than in either Asia or South America. Having said this, the larger area of relatively dry species-poor rainforest in the African region may account for the lower overall number of species in that region.

It seems that not only is there some ecological driving force which gives rise to more species of rainforest trees in the wettest tropical climates, but the intensity of this driving force is modulated to give similar levels of richness between regions that have been separated long enough for them to have very different floras. There has been a great deal of speculation on the nature of this mysterious driving force, but as yet no really convincing explanation has been put forward.

The rainforest regions

African and Madagascan Rain Forests

The forest of the Congo Basin. By far the largest block of rainforest in Africa lies in and around the Congo Basin, a horseshoe-shaped catchment area formed from sedimentary rocks which overlie an ancient Precambrian basement layer. Most of the Basin is below 1000m in elevation, having a flat or gently rolling topography with large swampy areas at it centre. The Congo River drains the Basin westward into the Gulf of Guinea, across the coastal plain of Gabon.

West African rainforests To the northwest, a fringe of rainforest continues to the Cameroon Highlands, and along the northern coast of the Gulf of Guinea. After a gap of several hundred kilometres in the drier region between Togo and Benin, rainforest appears once again and extends as far as the Atlantic coast. Climates also become drier beyond the east of the Congo Basin; in the rift valley region the rainforest is mainly confined to localised mountain slopes. To the south, the rainforest of the Congo breaks up into a mosaic of forest and savanna, with the forest component becoming less and less significant as the rainfall becomes more seasonal.

Compared to the other main rainforest regions of the Americas and the Far East, a striking feature of the whole African rainforest region is its dryness. Only the very wettest parts of the Cameroon highlands match the annual rainfall of large areas of Amazonia and the Far East, and even they tend to have quite a strong dry season during part of the year. Indeed, most of the forest in Africa seems to be existing close to the climatic limits of what we would define as rainforest vegetation. During the dry seasons that occur once or twice each year, the canopy trees across much of the Congo Basin show a marked tendency towards deciduousness, although understory trees remain evergreen.

In the wettest climates of eastern Zaire and Cameroon, large trees of the leguminous (pea) family Caesalpiniaceae are abundant and sometimes form pure stands of a single species. Palms are relatively rare in the African rainforest, and they tend to occur only in particular situations, such as where there has been disturbance of the forest. In contrast to the Americas and the Far East, there are far fewer epiphytic plants growing on the branches of the trees in African forests. This might partly be due to the drier climates in Africa, but is perhaps also a legacy of the fact that the important epiphyte family Bromeliaceae has failed to reach there, in contrast to the hundreds of species which occur in the Americas.

There has has been much discussion of the extent to which the present-day flora and fauna of the African rainforest reflect a history of arid episodes that wiped out the forest in some areas and allowed it to persist in others. These refugia might then have retained species which have failed to expand their populations fast enough to spread out and recover their full potential range.

There are indeed strong signs that forest cover was reduced during the last Ice Age, between about 25,000 and 12,000 years ago. Whilst forest is known to have persisted in some areas, such as Cameroon, it was certainly greatly reduced further west. There are indications that rainforest may have been replaced by savanna and grassland over much of the Congo Basin, although considerable areas of riverine and swamp forest may have persisted. In general, some of the fossil and geological evidence suggests that the areas which are richer in species today (e.g. Cameroon, and the central Congo Basin) were glacial refugia, but then again these are among the areas of wettest climate at present. It is extremely difficult to disentangle the effects of glacial history and present-day climate, especially when fossil data on the distribution of forest during the last Ice Age remain sparse and ambiguous.

Humans and the African rainforests. The distribution of rainforest that one sees today in Africa is to some extent natural and to some extent the legacy of a long history of human influence. Modern-day humans seem to have existed longer in Africa than anywhere else, but it is difficult to gauge what effects they were having on the forests before the spread of agriculture to the region about 3,000 years ago. The savanna-forest mosaic which exists to the south of the Congo rainforest, and extends north as a finger almost to the Equator in west Congo, has been widely thought of as the recent result of denser agricultural populations of humans burning the savanna and keeping the rainforest at bay. However, fossil evidence indicates that at least parts of this mosaic have existed continuously since the last Ice Age, around 20,000 years ago. It seems that the relative proportions of forest and savanna have fluctuated in line with climatic changes and perhaps in response to changes in the density of human populations. It may be that preagricultual humans were also burning the savanna often enough (in order to make hunting easier) to have a significant effect on the extent of forest.

Since the spread of agriculture, the influence of shifting cultivation has been so pervasive that it has led many authors to wonder if there is any truly untouched forest left in Africa. Ecological studies of what might appear to be pristine rainforest have often found ancient traces of cultivation and habitation, such as charcoal, pottery fragments and hut foundations.

During the 20th century, parts of Africa have undergone the same rapid loss of forest cover that is occurring in other parts of the tropics. West Africa has now lost most of its natural forest cover to an expanding agricultural population, and Cameroon is starting to suffer the same problem.

In contrast, most of the central region of Africa remains densely forested, and sparsely inhabited. There is relatively little sign of the burning activities of cultivators showing up in satellite imagery, in contrast to the situation in many areas elsewhere in the moist tropics. However there are now indications from remote sensing of a reticulate network of plantations and croplands spreading out through the rainforest in Zaire, corresponding to the lines of the main road system. There are certainly some large areas of cultivated land in coastal Gabon, and intensive forestry activity in parts of eastern Zaire, although its extent has never been thoroughly mapped. With the pressure of growing populations, and increasing investment by foreign timber companies, all the expectations are that the demands on the central African rainforests will increase over the coming decades.

Madagascar forests. Geologically, Madagascar is a micro-continent which split off from the eastern side of Africa at some time during the Jurassic Period (more than 140 million years ago). Its flora and fauna reflect its continuing geographical proximity to Africa, with the imprint of partial isolation and independent evolution, and also with the addition of a suprising number of Asian groups of plants and animals. Many of these links to the Asian and African mainland must be the product of chance dispersal within the last few million years, carried by ocean currents and winds. However, some of the characteristic features of the Malagasy biota may in fact be a much older legacy of the former distribution of rainforest animal and plant groups that were present in the long-vanished Tertiary rainforests of eastern Africa (or on the northwards moving sub-continent of India) that have managed to reach Madagascar but have since died out where they came from. Madagascar is perhaps most famous for its lemurs, descendants of a group of primitive primates that was once much more widespread in the world. It seems likely that it was their isolation on this island continent which enabled them to escape competition and extinction at the hands of other more recently evolved primate groups.

In general, Madagascar is a rather arid place, but along its eastern edge the climate is more than moist enough to support a long band of rainforest. The high rate of endemism of plant and animal species (around 85% of the plant species on Madagascar are unique to the island) reflects the strength of its isolation from the African mainland, so it is perhaps suprising that the rainforest nevertheless has levels of species richness that are comparable with continental rainforest areas with similar climate. It is as if evolution has worked overtime on Madagascar to ensure that it has its full quota of diversity!

However, the distinctiveness of Malagasy plants does not run very deep. Perhaps because Madagascar was not so isolated during the critical early stages of evolution of the angiosperms, the process of evolutionary divergence has not been carried very far. In contrast to the high level of species endemism, there is only one family of flowering plants that is unique to the Madagascar rainforest, the Humbertiaceae. Many Malagasy plants also belong to genera that occur elsewhere in the old world tropics, perhaps having reached the island by dispersing across from the African mainland, although some of these have radiated out into quite a large number of species on the island. Thus for example, one can find five species of wild malagasy coffees (Coffea) in only 2 km2 of rainforest. Since these wild coffees are resistant to the diseases which afflict the African-derived cultivated coffees, one can see that there may be great potential here for producing commercially important hybrids.

One group of plants which has diversified strongly on Madacascar is the palms (Palmaceae). There are some 12 endemic genera of palms on the island (compared to only three endemic palm genera in Africa), but all have strong Asiatic affinities. The remaining 7 non-endemic palm genera also all occur in eastern Asia, but have close relatives in Africa.

Orchids are also very diverse on the island, with nearly 1000 species having been found (more than in the whole of continental Africa), many of these being confined to the rainforest. Perhaps the most remarkable is the white-flowered Agraecum sesquipedale, which has a spur some 35cm long. The great 19th century naturalist A.R. Wallace - already mentioned above in the context of Wallace's Line - predicted that there must exist a moth with a tongue long enough to reach down the spur and drink the nectar at the end, for otherwise the orchid would not be visited by insects and would not be pollinated. He was laughed at by entomologists for suggesting this, but sure enough 40 years later a moth was discovered that has a 35cm tongue capable of reaching down the full length of the spur. This moth, Xanthopan morganii forma praedicta('predicted form'), only ever fully unfurls its tongue when it nears a flower of the long-spurred Argraecum, and drinks hovering over the flower while the orchid glues pollen masses to its proboscis, ready for them to be carried on to pollinate the next flower of the species.

Sadly, the flora and fauna of Madagascar is suffering like no other from the pressures of human population. Already, a fantastical array of giant flightless birds and large bear-like lemurs has been wiped out by hunting and habitat loss since the first humans arrived 1500 years ago. The human population of Madagascar is now growing at an explosive rate, and fragmentation of the forest is occurring so rapidly that it cannot be long before many more extinctions follow. Many unique and potentially valuable species are holding on in only a few scraps of forest, encroached upon year by year as settlers grab more land to feed their families.

As the original vegetation is lost, artificially introduced species of plants are taking its place in the disturbed secondary forest and scrub that remains. It seems to be a characteristic of islands in general that they are susceptible to invasion by garden escapes and weeds brought over accidentally from other parts of the world, and Madagascar is no exception to this rule.

Rainforests of the Amazon and Orinoco.

Within the linked catchment zones of the great Amazon and Orinoco rivers lies the world's largest contiguous area of equatorial rainforest. Moist Atlantic air travelling westwards is thought to supply most of the region's rainfall, but the wettest climates are found on the opposite side of the continent close to the Andes, where the air is forced to rise and shed its heavy burden of water vapour. As it travels west from the coast, the prevailing air stream picks up much of the water vapour evaporated back up from the forest and redeposits it as rain further on. Despite the recycling, there is also a constant loss of water as it enters the massive Amazonian river system and flows straight down to the Atlantic Ocean. The role of moist Pacific airmasses in contributing to the high rainfall in the western Amazon remains uncertain, although some studies suggest that they are in fact important. Whatever the causes, the climate becomes drier from West to East across the Amazon Basin, but moister again as one approaches the Atlantic coast in the north-east.

In general, the Amazonian forests exist under a rather moister climate than their African counterparts. Although there are huge areas of semi-evergreen rainforest towards the southern and eastern margins of the Amazon region, there is also a very extensive core area of evergreen forest. In these wetter parts of the region, with annual rainfall above about 2000mm and with no strong dry season, the forest is taller and richer in species. Without the regimentation of strong dry seasons, there is less reason or opportunity for biological processes to be synchronized, and growth and reproduction of plants and animals occurs more evenly around the year.

The annual rainfall along the eastern foothills of the Andes is particularly high (reaching 9,000mm in places), and distributed fairly evenly throughout the year. In this ever-wet climate, species richness of trees, lianas, shrubs and herbaceous plants reaches almost unbelievable levels. One study found 283 species of trees and lianas with stem diameters over 10cm in a tiny 0.1 ha sample of forest in the Napo region of Peru. Furthermore, many of the plant species within these forests of the Andean foothills have very localised distributions, sometimes being found on just one or two hilltops, with their place being taken by other similarly localised species elsewhere.This adds even further to the species richness on a broader scale, making these forests a truly remarkable reservoir of biological diversity.

Forest types of the Amazon region.The great extent of the Amazonian forests, together with the regional peculiarities of the soil types and river systems within the Basin, allows a number of very distinctive forest types to exist, each with its own specialized flora. By far the largest area is taken up by tierra firme forest, which is the 'standard' forest type that one usually has in mind when referring to 'the' Amazon rainforest. Tierra firme forest occupies fairly well-drained soils that are relatively rich in available nutrients. It is tall (40m or more) and usually very rich in tree species, with a dense upper canopy and a relatively dark, open interior.

Other forest types occur under conditions on soils that are particulary poor in nutrients, or in areas subject to fire or flooding. White sand caatinga is a low, scrubby forest with small hard leaves, that occurs in places where the soil is mainly bleached sand. It is scattered throughout the Amazon region, but particularly in the Rio Negro catchment area where the soils are formed on ancient sea beaches and other sandy deposits.

Liana forest (known as cipoal in Brazil) is a relatively open forest in terms of its tree cover, but is almost smothered by huge numbers of vines growing on, over and between the trees. It often seems to occur on the intensively weathered latosols which are associated with important ore deposits of iron (as in the Serra dos Carajas region) and aluminium, but it can also be found on the more fertile terra roxa soils. The liana forest tends to form a close patchwork with other types of forest, and is particularly abundant in the area between Maraba and Itaituba through which the TransAmazon Highway passes.

Another type of more open forest which is widespread in Amazonia is palm forest, dominated by the babacu palms (Orbignya spp). Palm forest is often partly the product of human influence, with frequent burning helping to encourage the fire-resistant babacu. In terms of the indigenous economy of the forested regions, this forest type is important for the Brazil-nut trees (Bertholletia excelsia) which are also present.

Although there are some areas of permanent swamp forest, this is a much more minor component than the seasonally inundated forests that line the riverbanks and floodplains of the Amazon and its tributaries. The highly seasonal nature of the inundation is a result of the vast extent of the Amazon catchment. The Basin stretches north and south across the equator and westwards as far as the Andes, so when the rainy season reaches its peak in either the north or the south, the runoff is sent rushing downstream in a great burst. Hence, the water levels of rivers even in the relatively non-seasonal areas of the Amazon Basin closest to the equator can reflect this seasonality occurring in the catchment areas far upstream.

There are various types of seasonally or intermittently inundated forests along the river systems of the Amazon Basin. Where a river is fed from clay-soil areas, the suspended sediment in the water tends to be be deposited over the bank tops as it floods. This builds up broad low ridges, known as levees, along either side of the river, channelling its flow and limiting the extent of flooding. This levee-top and floodplain forest, known as varzea, is very similar in structure to the terra firme forest. The trees tend to be heavily buttressed (perhaps an adaptation for anchorage in the moist clayey soils), and often have seeds with special water flotation mechanisms that enable them to be dispersed when the river is in flood. The rubber tree (Hevea brazilensis) is an example of a species which is clearly adapted to the varzea habitat, having seeds which can float for up to two months and which form an important part of the diet of certain large fish. Perhaps because of the regular input of nutrient-rich sediment, the herbaceous understory layer is particularly lush and rich in species of such plants as gingers (Zingiberaceae) and heliconias (Scitaminae).

Diversity centres - real or imaginary? Within the Amazonian rainforest block, there is undoubtedly a general gradient in tree species richness from west to east, following the trend in annual rainfall. However, many groups of animals and plants have also been purported to show scattered centres of high diversity, containing relatively large numbers of endemic species. It has been suggested that these diversity centres might have been refugia for rainforest species during dry glacial periods in the past, with many of the species having been too slow to disperse out of their refugia as the forest expanded again during the wetter interglacials. Some recent work indicates that many of these apparent diversity centres could be no more than an illusion, resulting from differences between the amount of fieldwork done in different areas. If more collecting work has been done in any particular area, as compared to its neighbouring areas, it is to be expected that more species will have been discovered there! It seems that more research will be needed to sort out whether these diversity centres actually do exist.

The Atlantic Rainforest Along the Atlantic coastal strip of south-eastern Brazil, the moist maritime air gives enough rainfall to sustain a long belt of equatorial rainforest. Towards the south, this merges imperceptably into cooler moist evergreen temperate forest, losing much of its species richness as it does so.

Separated from the the main Amazon rainforest block by many hundreds of kilometres of dry scrub and savanna, the Atlantic forest has a high proportion of endemic species. Nevertheless, around half of the tree species found in this forest also occur in the Amazon rainforests or elsewhere in the Americas (e.g. the tree Guarea guldonea), reflecting some sort of common history in the recent past. Many of these non-endemic trees also occur scattered through the broad intervening scrub/savanna regions as small populations within strips of 'gallery forest' along river floodplains. Thus, it is not difficult to imagine how their seeds might have been dispersed stepwise over much shorter distances between the separate areas of forest. It is also of course possible that past phases of moister climate could have brought the gallery forests extending in further towards one another, allowing the gradual exchange of plants and animals to occur more readily.

The Caribbean rainforests.To the north of the main Amazonian rainforest block, many of the scattered islands of the Caribbean Basin are partly or wholly covered by rainforest. Although the climate over much of this region is seasonally quite dry, the steep topography of many of the islands forces the maritime air to rise and cool, losing its moisture as rainfall as it does so. Where this happens, rainfall may exceed 2000mm and moist evergreen or semi-evergreen forest occurs on the lower mountain slopes and hills.

On Trinidad, mora (Morea excelsa) forms extensive stands within the moist forests, but elsewhere in the Caribbean the forests are generally not dominated by any one particular species. In the Windward Islands, at the southeastern edge of the Caribbean, typical trees of the forest canopies include Slonea and Canarium. On the Greater Antilles, comprising such islands as Jamaica and Cuba, species of Ficus and Psidium are particularly important trees. These island forests are generally less species-rich than those of the South American mainland, perhaps partly because of their drier climates at present and their history of greater aridity during ice ages, and also due to their relative isolation from larger sources of evolutionary novelty on the mainland. It is a general characteristic of islands everywhere that, when other things such as climate are equalled out, they tend to be poorer in species than mainland areas of the same size (a notable exception to this rule is Madagascar, but this has a long geological history and is in any case quite a large island). This must be partly because newly evolved species on large landmasses are less likely to reach them, but is also surely because the smaller populations of plants and animals that exist on these islands are more likely to be wiped out through chance catastrophes.

Amongst those tree species which do occur on the Caribbean Islands, there is a close affinity to the South and Central American rainforests. Most of the genera are the same, and quite a large proportion of the species are also shared with the mainland (e.g. the same Gaurea guldonia as also occurs in the Amazon, Central American and Atlantic rainforest).

The Pacific rainforests. On the west side of the Andes, through western Ecuador, Peru and Colombia, there is a thin band of rainforest stretching parallel to the coast. Both floristically and geographically, it is very close to the Amazonian forests that lie to the east, but it has probably been completely separated from them for millions of years after the Andes grew up from beneath. Thus, although one finds many of the same plant genera that are present to the east, they often belong to distinct species.

Central America. Further north, the west Andean forest connects with the central American rainforest at Panama, and stretches north to Mexico. Whilst tending to have a rather more seasonal climate, these central American forests are nevertheless very rich in species. In Costa Rica, 233 species of plants (of all types taken together) have been found within a sample area of only 100m2 (0.01 ha) of forest! Generally, the rainforests of central America are very similar in composition to those of the Amazon and west Andean regions, with local endemic species mixed together with more widespread ones that occur throughout the region.

Exploitation and loss of equatorial rainforests in the Americas. The central American rainforests once gave rise to the only ancient civilization ever to develop in a true lowland rainforest region. The Maya, whose empire was initially centered on the Yucatan Peninsula of Guatemala, began building stone temples around 400 B.C. and persisted as a powerful political and cultural force up until 1300 A.D.. They were initially able to support a large urban population by intensive agriculture, using terracing and manuring to help reduce soil erosion. Nevertheless, it seems to have been soil erosion and decline in soil fertility - combined with a sudden severe drought event - that eventually brought about a dramatic collapse of the population around 800 AD, from which the Maya Empire never truly recovered.

Other than the Maya, the only inhabitants of the rainforests of the Americas before European settlement were native hunter gathers and shifting cultivators. Other native civilizations, such the Aztecs and Incas, avoided the lowlands and concentrated on montane environments with their more fertile soils.

Over the last few centuries there has been an increasing drive to utilize the lowland forests commercially. The process of forest clearance for crop-growing and for ranching has accelerated greatly over the past few decades, partly as a result of the growing populations in tropical American countries. Thus for example, in Costa Rica the area of forest has been reduced from 67% primary forest in 1940 to only 17% in 1983. The Atlantic rainforest on the south-eastern coast of Brazil has now been almost totally destroyed, with primary forest occupying less than 2% of its original area.

The process of forest conversion in Amazonia has been increased by planned projects that have opened up previously inaccessable regions. In the southern Amazonian state of Rondonia, the Brazilian government built a road (BR-364) which allowed large scale migration of landless farmers into the region. The result is that much of the primary forest in Rondonia has now been destroyed. Other government-backed projects in Brazil have resulted in the large-scale clearance of rainforest for its use as iron smelter fuel in the Carajas region in the north-east, and substantial areas of forest have been flooded by the building of hydroelectric dams along the Amazon tributaries.

So far, it seems that somewhere between 5 and 12% of the primary rainforest of Amazonia has been cleared, although parts of this have now reverted back to relatively species-poor secondary forest. There are recent signs of a significant shift in Brazilian policy away from large-scale clearance of rainforest, following the failure of transmigration and ranching projects to benefit the country's economy.

On the western side of the Andes, the exceptionally species-rich Peruvian and Colombian forests are now being cleared rapidly, with coca (cocaine) and opium poppy cultivation being one of the major incentives. There are also plans afoot by foreign timber companies to exploit these forests on a large scale for chipboard manufacture. To the north, in Panama, the loss of forest from around the Canal Zone to illegal logging and shifting cultivation has resulted in a great increase in erosion and a significant decrease in rainfall. As a result, the canal is silting up and its water level has also fallen, bringing about difficulties for its use by shipping.

The Far Eastern Rain Forests

The far eastern rainforests seem to be the product of quite separate floras that had evolved on different crustal plates, which later collided and spilled their animals and plants across onto one another. Each of these collisions has left its distinctive mark on the biology of the forests, and on the patterns of distribution of plants and animals which we see today.

All across the far eastern region there exists a strong floristic resemblance to the African tropics, a reminder of its descent from the biota of the great northern continent of Laurasia which remained intermittently linked to Africa across a land bridge or separated only by narrow seaways. Still further to the east, there is the legacy of a long period of separate evolution on the Australasian plate, which has now collided with the edge of the old Laurasian plate and received much of its flora.

The west Malesian forests; Malaysia and Indonesia. The equatorial rainforests of west Malesia have been well described by naturalists and foresters over the past century, and more recently been they have been mapped in detail using satellite data. Towards the north of the region, the rainforests fade into deciduous monsoon forests, but bands of moist climate allow rainforest to continue to its most northerly point (26 Deg.N.) in Southern China.

The canopy of the west Malesian forests is generally between 30 and 40m tall, with large emergent trees projecting up to 60-70m, and with all the usual assemblage of rainforest features such as buttress roots and large evergreen leaves. For the most part, the forests of the region exist under very weakly seasonal climates, with an annual rainfall of over 2000mm distributed through the year. The wettest climates of all are found in the Kalimantan region of north-western Borneo, where tree species richness also reaches its highest levels.

Studies of the seasonal behaviour of rainforest trees in equatorial Malaysia have indicated that even the relatively slight dip in rainstorm frequency that occurs twice each year is enough to act as a cue for a burst of leaf renewal and flowering in many species. The actual cue that the trees are sensing may be an increase in drought stress, or perhaps a slight temperature change within the canopy. Not all trees reproduce every year; many species in the family Dipterocarpaceae show a pattern of 'mast fruiting' only once every few years, with all of the tree population across a large area coming into flower simultaneously.

Mast fruiting often seems to be synchronized by the more intense droughts that can occur in even the wettest regions, once in a while. These occasional droughts tend to be associated with 'El Nino' events; sudden temporary reversals in the tropical Pacific circulation. An exceptionally severe El Nino event in 1982-83 led to the drying out and destruction of large areas of rainforest in Borneo by fire, although these fires were made worse as a result of forestry activity that had left inflammable debris scattered through the forests. It is also thought that many of the fires were deliberately started by shifting cultivators to help clear small patches of forest, and that these quickly went out of control.

It may be suprising to learn that the scattered rainforest islands of west Malesia have a flora and fauna which is fairly similar throughout. After all, many of these islands are separated from one another by hundreds of kilometres of seawater, so one might have expected evolution to have taken different directions due to their relative isolation from one another. However, as is true at a more general level throughout the tropics, this similarity amongst the islands is a legacy of land links which once existed but have now vanished. In the case of west Malesia, these links were last present less than 10,000 years ago, in the period following the last Ice Age before sea level had risen fully to its present level. In fact, the 'normal' state of south-east Asia during the past two million years or so has been as a single linked landmass, with only brief, exceptional warm periods (such as the one we are living in at present) melting enough polar ice to separate off the islands from one another. As a result of this history, there has been plenty of opportunity for species to spread overland from one part of west Malesia to another. Nevertheless, a great many local endemics and races do remain confined to particular islands or parts of islands, although these differences may have more to do with present-day variations in climate and soil type than the recent history of isolation by sea level rise.

Variations of forest types. On particular soil conditions, caused by variations in geology or drainage, various specialized types of forest occur. Each forest type has its own assemblage of characteristic tree species, together with others that occur more widely.

Heath forest, or kerangas, is ecologically very similar to the white sand caatinga of Amazonia. It is shorter than other lowland rainforest types, with small leaves clustered on upturned twigs. As with the Amazonian caatinga, it occurs where the soil has been podsolized to form a nutrient-poor bleached sand. The abundant phenolics (such as tannins) which occur within the leaves are thought to be a defense mechanism to prevent insects from eating them and robbing the plant of precious nutrients. It is these phenolics which seem largely responsible for staining the water of the blackwater rivers in these areas, giving it the colour of tea.

Limestone forest occurs in the dramatic karst landscapes of Borneo and the Malaysian peninsula. These areas have a very rich flora with many endemics, including both trees and the herbaceous plants that cling to the limestone rock faces. Being so well drained, the forests growing on limestone areas suffer relatively severe drought stress during dry seasons, and often develop 'autumn' colours before shedding their leaves for part of the year.

Swamp forests of various sorts occupy considerable areas in western Malesia, particularly in Borneo. For example, peatswamp forest has a layer of between a few centimetres and several metres of peat beneath the shallow roots of the trees. In contrast to the forests on dry land, dipterocarps are seldom dominant, although species such as Shorea platycarpa do sometimes form dense stands. It was the peatswamp forest of eastern Borneo which fared particularly badly during the drought and fires of early 1983. At this time, the water table fell far enough to allow large areas of peat to dry out. The peat then burnt, killing the trees which were rooted into it.

The Indian rainforests. The equatorial rainforest in India occurs in two separate areas; along the strip of hills and mountains near to the west coast, the Western Ghats, and in the northeastern state of Assam close to the border with Myanmar (Burma). In both these areas, semi-evergreen rainforest is more widespread than evergreen forest, probably due to a long history of human influence which has degraded the structure of the forest and its soils, and led to a more precarious water balance.

Typically for the Far Eastern region, dipterocarps are important in both the western and northeastern rainforest areas of India, but there are none of these species in common between them. In the Assam valley, isolated individuals of Dipterocarpus macrocarpus and Shorea assamica tower above the main canopy, reaching a height of up to 50m and a girth of up to 7m. In the Western Ghats, giant buttressed specimens of other Dipterocarpus species commonly tower to a height of over 30m before reaching their first branch.

In general, the rainforests of the Western Ghats are more diverse than those of Assam, with over 4,000 plant species occurring in this relatively small area of hill land. Of these, 1,800 species are endemic to the Western Ghats, most of these being confined to the rainforest rather than the other drier vegetation types that occur mixed in with it. As one would expect from the general global patterns, the highest species richness is reached in the southernmost forests of the Western Ghats, where the dry season is shortest.

As with the African forests, there appears to have been a pervasive influence of humans on the Indian rainforests going back thousands of years. There are still a few of the ancient hunter gatherer groups which must have existed in this region before the arrival of agriculture. Other indigenous groups and more recent settlers still utilize the traditional system of shifting cultivation, growing crops in a clearing for a few years before abandoning the land to the forest. It is thought to be the history of disturbance through shifting agriculture that has led to clumps of bamboo being widespread within the forests along streams and in poorly drained hollows. As India's population expands, the intensity of shifting cultivation is becoming more intense, with shorter fallow periods giving less time for the forest to recolonize and recover.

Sri Lanka. In Sri Lanka, lowland rainforest would naturally occur over much of the south-western half of the island, where annual rainfall is between about 3500 and 5000mm, although over the past few decades it has been greatly depleted. There are many plant genera of Malesian affinities, such as Mesua and Vitex. Dipterocarpaceae are also locally important, with such species as Dipterocarpus hispidus and Shorea spp occurring on moist, fertile floodplain soils.

The Dipterocarps. There are some interesting features of the far Eastern rainforests that seem to be a legacy of chance evolutionary events in the region's history. Most striking of all, the west Malesian and Indian/Sri Lankan rainforests are dominated beyond all proportion by members of a single family of trees, the Dipterocarpaceae. Dipterocarp pollen first appears in the fossil record of west Malesia as an explosive burst about 30 million years ago. The timing of this event follows the early Tertiary collision of the northwards moving India-Ceylon plate with southern Asia, and it is possible that the group may originally have come from there. A few species of dipterocarps have managed to reach across the second collision zone to the east of the island of Sulawesi, so the eastward range of the family extends across Wallace's Line into New Guinea. Although several dipterocarps do occur in Africa, they are trees of savanna and woodland and not of the rainforest. A couple of species of what may be dipterocarps have recently been found in the Guiana Shield rainforests of South America, but their real evolutionary affinities are still in question. What is not in dispute is that the dipterocarps of the Far East belong to a quite distinct subfamily, the Dipterocarpoideae, and that they have diversified disproportionately into tens of genera and hundreds of species growing in a range of habitats throughout the rainforest.

Dipterocarps of the far east are not just abundant in terms of species; they also make their presence felt by their size and sheer weight of numbers. Flying low in an aircraft over the primary forest of Borneo or the Malayan peninsula, one would notice that the canopy is punctuated by clumps of huge dipterocarp trees projecting to a height of perhaps 60m. These emergent groups may consist of a single species (often of the genus Shorea), breaking the general rule that rainforest trees grow only as isolated individuals scattered in amongst other species. The surrounding forest is also likely to contain a large variety of different dipterocarp species, some of them scattered as rare individuals and others more abundant and gregarious.

The very abundance of dipterocarps, and the high quality of their timber, has been their downfall in the 20th century. They are a lucrative target for timber companies and as a result they have been virtually logged out from large areas of the Philippines, Sabah and elsewhere.

Loss of rainforest in the far East

The destruction of the far eastern rainforests has been exceptionally rapid during the last 50 years. As mentioned above, the most of the primary forest has been disturbed to some extent by timber extraction. In many areas, the forest cover has been lost altogether. For example Vietnam was about 45% forested in 1943, but this figure had fallen to about 20% by the mid-1980s. Population pressure for agriculural land seems to have been the main factor behind the loss of forest in Vietnam, but deliberate bombing and spraying for defoliation were also significant during the war of the 1960s and 70s. Many of the bombed and sprayed areas have so far shown little sign of natural recovery of forest vegetation, although replanting efforts are now underway.

In Indonesia, a government-sponsored transmigration programme in the 1980's shifted some three million people from densely populated and heavily deforested islands such as Java, to less densely populated islands. This resulted in more rainforest being cleared for cropland and towns, but due to various problems such as the poor suitability of soils for agriculture, the programme has recently been scaled down.

In contrast, Borneo still retains a great deal of primary forest, although the extent of forestry and agriculture is on the increase, especially in the northern state of Sabah. Timber companies are now investing heavily in the region as the Malaysian forests become exhausted of useful timber. Population pressure is also leading to an increase in the area of forest cleared for slash-and-burn agriculture. As mentioned above, it seems that both agriculture and forestry practices combined to help produce the extensive fires that occurred in northern and eastern Borneo during the drought of 1982-83.

The Indian and Sri Lankan rainforests have suffered badly as a result of increasing human populations and poorly controlled logging during the past few decades. The longer-term result of unsustainable exploitation is a degradation in the quality of the forest resources for the people that benefit from them, and a loss of biological diversity. In Sri Lanka, uncontrolled deforestation occurred partly as a result of the breakdown of order resulting from civil war. Various endemic tree species with localised distributions on the island are thought to have gone extinct now, since the only areas they occurred in have been completely deforested.

In Myanmar (Burma), there is an increasing need for agricultural land from the growing population, with shifting cultivation occurring at intensities that cannot sustain allow the forest to recover fully. Timber poaching to feed markets in neighbouring countries, and excessive felling rates in production forests, is also depleting the forests on a large scale.

East Malesia; the New Guinean and Australian Rain Forests

The East Malesian region basically consists of the Australasian plate, a fragment of the ancient southern continent of Gondwanaland. Presently separated from one another by only a shallow continental sea, Australia and New Guinea were repeatedly linked by a land bridge (known as the Sahul shelf) in the past, particularly during the many low sea level phases that have occurred during the last two million years. The last land linkage between the two landmasses seems to have broken less than 10,000 years ago when sea levels rose after the last ice age. Although equatorial or sub-equatorial rainforest climates were apparently present over much of the northern half of Australia during the mid-Tertiary, the rainforest is now confined to a few fragments along the Queensland shore. In contrast, New Guinea remains largely covered in rainforest, with a few areas of monsoon forest at its southern and eastern edges.

For the most part, the Australian and New Guinea rainforests are now dominated by west Malesian groups of plants that have managed to disperse across the plate collision boundary. Many are isolated species belonging to genera which are much more diverse on the other side of Wallace's line in west Malesia. However, there are also many endemic or nearly-endemic genera of canopy trees present (e.g. Doryphora). These include a strong element of indigenous groups of plants that once formed an ancient rainforest flora in the east Malesian region, as part of the Australasian crustal plate. As a legacy of this past, the conifer genera Auracaria, Dacrydium, Podocarpus and Agathis still occur in the lowland rainforest of the region, sometimes as extensive stands. The only rainforest eucalypt, Eucalyptus deglupta, forms pure stands of trees in the New Guinea heathforests, where it is of considerable economic importance as a timber tree.

The east Malesian rainforests remain remarkably intact compared to those of most other regions. Very little of the forest of New Guinea has been touched by forestry or large-scale agriculture, although it has a long history of habitation by shifting cultivators. The Australian rainforest covers about 90% of the area it occupied when the first Europeans arrived, and much of what remains is now protected.

However, the native fauna and flora are of the Australian rainforest to some extent being menaced by introduced species, for example the American cane toad (Bufo marinus) which, if eaten, poisons the native marsupial cat (Dasyurus maculatus). Once they have been introduced, pest species such as the cane toad are often virtually impossible to eradicate from the forest; it seems that they will simply remain there forever. By bring such species with them, humans have managed to break geographical barriers between regions that had previously stood firm for many millions of years. It is the very fact that these barriers were once so effective that has allowed evolution to take so many different but parallel paths, giving rise to the fascinating range of rainforest biotas that exists across the world. It seems possible that from from this age forward, humans and the introduced animals and plants they leave behind will turn out to have a profound influence on the evolution of rainforest biotas.

Bibliography

Beazley M./IUCN. (1990). The Last Rainforests. M. Beazeley, London.

Briggs J.C. (1987). Biogeography and Plate Tectonics. Developments in palaeontology and stratigraphy v.10. Elsevier, Cambridge. 350 pp.

Jacobs M. (1983). The dipterocarps. Ch.3 in; Key Environments: Malaysia. Pergamon Press, Oxford.

Jacobs M. (1988) The Tropical Rainforests: a first encounter. Springer-Verlag, Berlin. 295 pp.

Guillaument J-L. (1984). Ch.2 The vegetation, an extraordinary diversity. p.27-51. in; Key Environments. Madagascar. ed. Jolly A., Oberle P., Albignac R. Pergamon Press, Oxford.

Hamilton (1983). African Forests. Ch.8 in Leith H. & Werger M.J.A. 'Tropical Rainforest Ecosystems'. v.14B in; Ecosystems of the World series. Elsevier, Amsterdam.

Heywood V. (1979). Flowering Plants of the World.

IUCN (1991). Conservation Atlas of Tropical Forests: Asia and the Pacific. ed.s Collins N.M., Sayer J.A. & Whitmore T.C.. Macmillan, London.

Mabberley D.J. 1983. Tropical Rain Forest Ecology. Blackwells, Oxford.

Walter H. (1971). Ecology of tropical and subtropical vegetation. Oliver & Boyd, Edinburgh.

WCMC (1992) Global Biodiversity: Status of the Earth's Living Resources. Chapman & Hall. London.

Whitmore (1983). Southeast Asian Tropical Forests. Ch 10. in Key Environments: Malaysia. Pergamon Press, Oxford.

Stocker & Unwin (1983). Northeast Australian Rainforests. Ch.12. Ibid.

Whitmore T.C. Forest types & forest zonation. Ch.2 in; Malaysia. Key Environments. Pergamon Press, Oxford.

Pires & Prance. Vegetation types of the Brazilian Amazon. Ch.7 in Key Environments; Amazonia. Pergamon Press, Oxford.

Walter H. (1971). Ecology of tropical and subtropical vegetation. Oliver & Boyd. London. 295 pp.

Wolfe J.A. (1985) Distribution of major vegetation types during the Tertiary. p. 233-256. in; The Carbon Cycle & Atmospheric CO2, Geophysical Monographs v.32.