See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/226173322

The changing ecology of tropical forests

Article in Biodiversity and Conservation · February 1997

DOI: 10.1023/A:1018352405482

CITATIONS READS

136 558

1 author:

Oliver Lawrence Phillips

University of Leeds
689 PUBLICATIONS 80,365 CITATIONS

SEE PROFILE

All content following this page was uploaded by Oliver Lawrence Phillips on 20 September 2019.

The user has requested enhancement of the downloaded file.

Biodiversity and Conservation 6, 291±311 (1997)

The changing ecology of tropical forests

O. L. PHILLIPS
School of Geography, University of Leeds, Leeds LS2 9JT, UK

Received 12 August 1995; revised and accepted 8 May 1996

The threat to tropical forests is often gauged in terms of deforestation rates and the total area
remaining. Recently, however, there has been a growing realization that forest can appear intact on a
satellite image yet be biologically degraded or vulnerable to degradation. The array of direct threats
to humid tropical forest biodiversity, in addition to deforestation, includes: selective extraction of
plants; selective extraction of animals; biological invasion; fragmentation; climate change; changing
atmospheric composition; and increasing tree turnover rates. The threats are linked to one another
by a poorly understood network of causality and feedback e€ects. Moreover, their potential impacts
on forest biodiversity are hard to assess because each threat is as likely to precipitate indirect e€ects
as direct e€ects, and because several threats are likely to interact synergistically with one another. In
spite of the uncertainties, it is clear that the biological health of tropical forests can become seriously
degraded as a result of these threats, and it is unlikely that any tropical forest will escape signi®cant
ecological changes. Some groups of plants and animals are likely to bene®t at the expense of others.
Species diversity is expected to decline as a consequence of the changes in forest ecology. In the 21st
century scientists and conservationists will be increasingly challenged to monitor, understand, pre-
vent and head o€ these threats.
Keywords: biodiversity; degradation; defaunation; fragmentation; climate change; carbon fertiliza-
tion; increasing turnover; interactions; monitoring.

Introduction
Most land in the tropics has some history of human use (Flenley, 1979; Sanford et al.,
1985; BaleÂe, 1989, 1994; Horn and Sanford, 1992). These human activities, such as hunting
and deforestation, were once very restricted in time and space. Therefore, defaunated areas
were usually recolonized and deforested clearings usually regenerated. However, with
technological and population expansion, the same activities are now becoming increas-
ingly pervasive, giving tropical forests less opportunity to recuperate.
Satellite images present one, very graphic, view of the spreading threat. They show that
large-scale deforestation is proceeding in many parts of the tropics (Myers, 1989, 1993;
FAO, 1993; Grainger, 1993; Skole and Tucker, 1993). Less than half the original extent of
tropical forests remains in primary forest, so deforestation must have already set in train a
massive wave of extinction (Myers, 1988; Sayer and Whitmore, 1991). Yet outright
clearing of the forest is only one, albeit the most extreme, among a series of changes that
are a€ecting the ecology of the earth's most diverse ecosystems. The other changes will not
necessarily remove forest cover. Instead, they are more likely to cause progressive de-
gradation of forest structure and biodiversity. Unlike the obvious devastation caused by
clear-felling or large ®res, these other threats share the quality of being hard to perceive
without careful monitoring. (Only one, fragmentation, is potentially visible with satellite

0960-3115 Ó 1997 Chapman & Hall

292 Phillips
imagery, and the cascade of knock-on e€ects following fragmentation cannot be perceived
by remote sensing.) These threats to forest biodiversity are additionally disquieting for
conservationists because they will not operate in isolation, but rather may coincide with
one another and with the dominant threat posed by deforestation. Therefore, in some
cases we can expect that the threats will interact synergistically to magnify their impact on
forests.
Recent geographical work has developed a framework for classifying the complex
nature of tropical forest degradation from a cultural biogeographic perspective (Grainger,
1996). But forest degradation is also a biological process, so we also need to understand
the ecology of changing forests. So far, biologists have mostly only examined the impacts
of individual degradatory processes ± particularly fragmentation (e.g. Wilcove, 1987;
Saunders et al., 1991), climate change (e.g. Hartshorn, 1992), and defaunation (e.g.
Redford, 1992). In this paper I attempt to develop a broader biological approach to the
degradation of humid tropical forests by considering the full spectrum of the forest de-
gradation threats from a biocentric perspective and exploring some of the links between
di€erent agents of degradation.
Non-deforestation threats to tropical forest biodiversity can be considered in seven
distinct categories: (a) selective extraction of plants; (b) selective extraction of animals; (c)
biological invasion; (d) fragmentation; (e) climate change; (f) changing atmospheric
composition; and (g) changing tree turnover rates. Each phenomenon merits serious
consideration as an agent of change in the ecology of tropical forests. Nevertheless, the
seven threats are extremely diverse and defy simple classi®cation. The ®rst two represent
an obvious immediate bodily removal of plants and animals, followed by a much less
obvious and potentially more signi®cant series of indirect ecological e€ects on the re-
maining biodiversity. Biological invasions (c) may initially increase local diversity, but
eventually lead to extinctions caused by direct and indirect ecological e€ects. Threats (d),
(e), (f) and (g) might also have immediate direct impacts on organisms, but more im-
portantly they also involve changes to the wider physical environment which will alter the
long-term ability of the forest to maintain natural levels of biodiversity. Finally, threats (c)
through (g) are themselves partly or wholly higher products of other forest threats. Indeed,
most of the seven threats are a€ected by other threats, including deforestation, in a
complex and poorly understood web of feedback e€ects. Figure 1 represents a simpli®ed
model of the known and hypothetical links among the threats.
The central questions addressed here are: `What are the biological implications of each
of these degradatory processes?', and `How might interactions and feedback e€ects modify
the biological responses to the processes?'. To answer these questions, I ®rst present a brief
overview of the threats of tropical forest biodiversity. Two particular threats (defaunation
and increasing tree turnover) are then discussed in more detail, as they provide instructive

c
Figure 1. Schematic chart of interactions among factors threatening tropical forest biodiversity and
ecology. Thick lines represent known e€ects, dashed lines represent potential e€ects. The ®gure
distinguishes direct e€ects on plants from direct e€ects on animals where possible, but does not
attempt to model the plant ± plant, plant ± animal, and animal±animal interactions that will de-
termine the ultimate e€ects on biodiversity of each threat. `Deforestation' includes clear-felling, as
well as intensive logging, and human-caused ®res that removes the forest cover. `Biodiversity' re-
presents the species-level and ecosystem-level diversity.
The changing ecology of tropical forests 293
294 Phillips
examples of the potential for indirect e€ects to alter tropical forest ecology in unexpected
ways. Finally, I explore the potential for synergisms among threats, and the conservation
implications of biotic degradation.

Overview of threats
The individual threats are brie¯y summarized here, with an outline of each threat and
tropical examples of likely direct and indirect impacts on plants and animals. Where
appropriate, the summaries and examples also consider likely synergistic interactions with
other threats.

SELECTIVE EXTRACTION OF PLANTS

General
Almost any plant species may be the target of extraction, and primary forest is often
utilized for a wide variety of both subsistence and commercial uses (Phillips et al., 1994a).
Exploitation can involve rapid, non-sustainable harvesting of particular species (Gentry
and VaÂsquez, 1988), leaving behind a standing but biologically and economically depleted
forest. Except for heavy selective logging (Myers, 1989), this form of resource use rarely
results in obvious structural changes, and therefore extractive forest is typically indis-
tinguishable from pristine forest on satellite images. National forest assessments that do
not include extensive ground-truthing by local experts may wrongly assume that forests
are intact when in reality they have already been mined of economically valuable plants.

Direct e€ect examples

Extraction of individual timber trees has left species like Swietenia macrophylla (maho-
gany) and Cedrela odorata (South American cedar) extinct across most of their former
range, even in areas still maintaining 100% forest cover. Several fruit trees and medicinal
plants are also extinct in most forests within 100 km of the small Peruvian city of Iquitos,
since it is often easier to cut trees for their fruit or latex than to climb or tap them (Padoch,
1988; VaÂsquez and Gentry, 1989).

Indirect e€ect examples

Where only scattered individuals or populations survive harvesting, species may be vul-
nerable to inbreeding, leading to further declines in genetic diversity. Extinctions can be
anticipated among other forest plants and animals (e.g. pollinators, fruit-dispersers) if they
depend on the over-harvested economic species for their own survival. If extractive logging
substantially alters the forest structure, humid tropical forest may become susceptible to
®re.

REMOVAL OF ANIMALS (`DEFAUNATION')

Very few tropical forest areas now have pristine assemblages of mammals and birds.
Surprisingly low population densities of humans can devastate a fauna, both directly (for
example, by subsistence hunting) or indirectly (for example, by removing many species'
critical food supplies by cutting fruit trees). Later I will explore the consequences of such
direct e€ects on the fauna, and indirect e€ects on the ¯ora and fauna.
The changing ecology of tropical forests 295
BIOLOGICAL INVASIONS
General
Exotic plant and animal species are important factors in causing extinctions, particularly
among endemic species on remote islands (e.g. Vitousek, 1988; Simberlo€, 1992; Cronk
and Fuller, 1996). Invading plant species typically have superior competitive abilities to
the native species, and invading predators may wipe out native herbivores that have little
ability to defend themselves. In more continental settings most communities are less
susceptible to invasion, although disturbance to the natural vegetation can still facilitate
invasion by competitive alien species.

Direct e€ect examples

In primary wet forest on Mauritius, a few exotic plant taxa (Psidium, Ligustrum, Ardisia)
constitute most of the seedling individuals and threaten to completely take over the forest
(Lorence and Sussman, 1986). In the Hawaiian archipelago, predation by introduced rats
may have precipitated the extinction of several endemic birds (Atkinson, 1977).

Indirect e€ect examples

When island birds are extirpated by rats, cats or humans, endemic plant species may lose
their natural fruit dispersal agents. Even in intact continental ecosystems, alien species can
have substantial impacts: rainbow trout (Oncorhynchus mykiss) have become top pre-
dators in some east Andean montane forest streams, indirectly modifying both the di-
versity and the trophic structure of forest streams and rivers (Young et al., 1994).

FRAGMENTATION
General
Spatial aspects of fragmentation are typically visible and quanti®able from satellite ima-
gery (Skole and Tucker, 1993; Grainger et al., 1996). However, subsequent ecosystem
decay in isolated fragments (Wilcove, 1987; Terborgh, 1992) is largely invisible from above
as fragments can acquire a simpler biological state without obvious changes in vegetation
structure and canopy re¯ectance. Other things being equal, larger fragments support more
species than smaller fragments (MacArthur and Wilson, 1967; Dale et al., 1994), but even
the largest existing forest reserves may be inadequate to protect all their species (Thiollay,
1989).

Direct e€ect examples

Species occuring at very low densities over large areas may be especially vulnerable to local
or global extinction as a result of reductions in their total available habitat. Also, since
tropical forest regions can contain numerous, distinct biological communities (Tuomisto
et al., 1995), those species that specialize in restricted environments will be prone to
extinction as a result of random sampling error when larger areas become subdivided.

Indirect e€ect examples

After isolation, a number of interacting processes, including loss of genetic diversity and
greater vulnerability to natural catastrophes, may threaten species' long-term survival.
Small fragments, or fragments with high fractal dimension, will be especially vulnerable to
edge e€ects, although the nature and penetration of forest edge habitat into the forest
296 Phillips
interior will vary greatly from species to species (Williams-Linera, 1990). The potential
variety of edge e€ects that may impact the forest is almost in®nite, including vegetation
desiccation (Kapos, 1989), higher rates of tree mortality and therefore of gap formation
(Lovejoy et al., 1984), greater exposure to hunting pressure and other threats such as heavy
seed rain from weedy species ± alien or native ± growing in adjacent secondary vegetation
(Janzen, 1983, 1986; Simberlo€, 1992; Turner and Tan, 1992; Turner et al., 1994).
CLIMATE CHANGE
General
Recent work suggests that tropical climates may be very susceptible to global climatic
changes. In the last ice-age, temperatures in the tropics may have been ³ 5°C cooler than
today (e.g. Bush et al., 1993; Guilderson et al., 1994; Stute et al., 1995). In the warmer
world of the near future, outer tropical forests are likely to experience more frequent and
more powerful hurricanes (O'Brien et al., 1992) and most tropical regions will experience
greater seasonal extremes between wet and dry seasons. Seasonality is known to be a
principal determinant of vegetation type, species richness, and biological activity in the
tropics (Foster, 1982; Hartshorn, 1992; Clinebell et al., 1995).
Regional environmental changes may also have a climatic impact. Tropical rainfall
depends partly on evapotranspiration from standing forest (Meher-Homji, 1991; Salati
and Nobre, 1991), so that deforestation might lead to regional climate changes that de-
siccate forests, accelerating forest decline. Climate changes have been documented in
several tropical regions that have been substantially deforested (e.g. Zhang, 1986; Wind-
sor, 1990), but a link between the two processes is far from proven.
Direct e€ect examples
Increasing seasonality of climate could bene®t species physiologically pre-adapted to live
in more seasonally arid forests (e.g. some reptiles, Bromeliad epiphytes, and small-leaved
deciduous vines with tubers) at the expense of those requiring year-round humid en-
vironments (e.g. some amphibians, some fern epiphytes, and evergreen vines with large
leaves). A change to drier climatic conditions in central Panama has coincided with sharp
population declines in small tree and shrub species that specialize in moist microhabitats,
and at least 16 species are predicted to become locally extinct in a 50 hectare inventory plot
(Condit et al., 1996).

Indirect e€ect examples

As a result of deforestation and fragmentation, in the next century there will be few large
unbroken areas of trans-altitudinal and trans-latitudinal forest. This will be a critical
factor in distinguishing the biological impact of future rapid global change from that of
past natural changes. Many plant and animal populations will be unable to track their
optimal climate by migrating across future landscapes. Montane forests may be most
severely impacted of all, since their lower limits may be forced upward by warmer and
drier climates and their upper limits will be unable to advance in the face of human
activities. By the standards of tropical forests, montane forests are not unusually diverse
but they do contain a very high proportion of narrow-range endemic plants and animals
(Young, 1992; Gentry, 1995; Jorgensen et al., 1995). Climate change, in concert with
deforestation and fragmentation, therefore presents a very serious threat to the survival of
many species.
The changing ecology of tropical forests 297
OTHER ATMOSPHERIC CHANGES
General
Carbon dioxide is not only a greenhouse gas, but also a basic plant nutrient used by plants
during photosynthesis. The increasing concentrations of atmospheric carbon dioxide will
directly a€ect plant metabolism everywhere (e.g. Condon et al., 1992; Johnson et al.,
1993). Other atmospheric changes are less globally pervasive, but trophospheric ozone
pollution, acid rain, smoky atmospheres from ®res, and nutrient deposition from smoke,
will all have local impacts on forest metabolism and may precipitate knock-on e€ects on
forest composition.

Direct e€ect examples

The increase in atmospheric CO2 levels has already had a biological impact. It is known to
have caused: reductions in the number of leaf stomata in some temperate species
(Woodward, 1987, 1992; Beerling and Woodward, 1993); increases in water-use eciency
(Wayne Polley et al., 1993); and possibly an increased cambial growth rates in temperate
trees (Graybill and Idso, 1993). Increasing CO2 is likely to directly bene®t intrinsically fast-
growing species (e.g. Condon et al., 1992) because they are pre-adapted to exploit high-
resource environments.

Indirect e€ect examples

It is dicult to predict how these metabolic changes will a€ect tropical forests, but we can
expect impacts on ecosystem-level processes and therefore ultimately on the species
composition and diversity of ecosystems (Bazzaz, 1990; Woodward, 1992; Johnson et al.,
1993). The ratio of carbon to nitrogen in litterfall may increase, which would favour some
detritivorous animal and fungal species over others and perhaps slow overall decom-
position rates.

CHANGING TREE TURNOVER RATES

For unknown reasons, possibly related to one or more of the above anthropogenic im-
pacts, inventory plot data suggest that the rate at which trees are dying and being replaced
in mature tropical forests has increased (Phillips and Gentry, 1994). Increasing turnover
will almost certainly have impacts on biodiversity, as will be discussed later.

Two invisible threats

Defaunation and changing turnover are discussed here in more detail. They are both
`invisible' processes, i.e. not visible on remotely-sensed images and therefore perceived
only by careful ground monitoring. Direct, indirect, and feedback e€ects of both phe-
nomena could create surprising and profound changes in the ecology of tropical forests.

DEFAUNATION

A few human hunters can defaunate a large area (Yost and Kelley, 1983; Thiollay, 1989;
Peres, 1990). The largest primates like Ateles (spider monkeys) and Lagothrix (woolly
monkeys), and some large cats, caimans, raptors, macaws and game-birds, are e€ectively
extinct in large parts of Amazonia (Thiollay, 1989; Peres, 1990). Many other animal
298 Phillips
species are also depleted by hunting. The `empty forest' (Redford, 1992) is an apt char-
acterization for the vast tracts of ostensibly pristine forest which have had many mammal
and bird populations hunted out. Moreover, most tropical rivers are ®shed and many are
over-exploited; this too may a€ect forest ecosystems since some ®sh species are important
seed dispersers in seasonally ¯ooded forests (Goulding, 1980; Kubitzki and Ziburski,
1994).
Human predation is having an obvious direct impact on the hunted forest animals. But
since ecological communities are characterized by a complex web of interactions among
species, we can expect indirect e€ects to propagate through the ecosystem, almost certainly
having e€ects on plants and animals not obviously connected to the target species. Few
studies have attempted to trace how the indirect e€ects of defaunation propagate through
the community, but it is already clear that defaunation can cause progressive change in the
forest community toward a simpler state. However, most plants have longer life-cycles
than most animals, so the e€ects of defaunation on plants may not be immediately ob-
vious. Consideration of three classes of ecological interaction ± seed dispersal, herbivory,
and predation ± suggests how ¯oristic changes might occur in defaunated forests.

Seed dispersal
Most trees are typically dispersed by frugivorous guilds of several animal species, rather
than by one co-evolved species, and most animal species consume fruit from several plant
species (but see Roth, 1984, and von Roosmalen, 1985, for some important exceptions).
Therefore, there is a degree of redundancy built in to most fruit/disperser systems.
However, the same large animal species favoured by hunters are often the most important
seed dispersers in tropical forests. So when populations of several large animal species are
depleted for an extended period of time, some plant species may simply fail to regenerate.
Evidence for `dispersal breakdowns' is largely anecdotal, but nonetheless persuasive:
ecologists who work in defaunated tropical forests are often struck by the obvious failure
of some proli®cally fruiting trees to be dispersed, as piles of uneaten fruit accumulate and
rot under the parent tree, in contrast to the situation at ®eld sites that are not a€ected by
hunting.
Dispersal may be a critical process for maintainance of tree diversity, by removing seeds
from unsafe sites crowded with conspeci®cs, that may, for example, attract predators or
pathogens (Janzen, 1970; Connell, 1971). Numerous studies have shown the advantages of
dispersal for a species' regeneration. For example, Hubbell and Foster (1990) found that
saplings under conspeci®c adults have reduced survivorship and growth, compared with
saplings dispersed a distance from conspeci®c adults. Chapman and Chapman (1995)
estimated that up to 60% of tree species at Kibale Forest, Uganda, could fail to recruit
adequately if frugivorous species were removed, based on the proportion of species that do
not to regenerate freely under conspeci®cs. Eventually, therefore, large imbalances in
populations of several frugivorous animals could eliminate populations of some trees.
Since mammal-dispersed trees and lianas tend to be localized in distribution (Gentry,
1983), a localized loss of large mammalian dispersers may result in disproportionately
many plant species extinctions.

Herbivory and seed predation

Research in Panama and Mexico has revealed the critical role that some herbivorous
animals may have in determining the composition and diversity of tree species in the
The changing ecology of tropical forests 299
forest. Both examples consist of comparisons of partly defaunated forest fragments with
less defaunated larger fragments. They also show, incidentally, that defaunation is an
inevitable consequence of fragmentation, and therefore that the lines dividing the threat
categories can become very blurred.
Putz et al. (1990) and Leigh et al. (1993) found that a depleted fauna can lead to a
decline in tree species richness. They surveyed islands in Panama, isolated by arti®cial
lake-waters in 1913, that were too small to support rodent seed predators. The island
forests have changed greatly, losing ¯oristic diversity and becoming dominated by a few
large-seeded species which have been released from normal levels of predator pressure. By
contrast, on the nearby much larger Barro Colorado Island (BCI), population densities of
important seed predators such as squirrels (Scireus) and agoutis (Dasyprocta) have per-
sisted at very high densities, apparently because the island is too small to support their
own predators (Glanz 1982, 1990). Terborgh (1988, 1992) has suggested that the super-
abundance of large seed predators could be depressing the regeneration of large-seeded
plants on BCI, which would be a completely contrary result to that pertaining on the
smaller islands. This highlights the diculties in predicting which animal and plant species
will `win' and which will `lose' under di€erent defaunation scenarios.
In a further demonstration of the importance of forest herbivores, Dirzo and Miranda
(1990, 1991) made an elegant comparison of the fauna and understorey of two Mexican
rainforests. One (Los Tuxtlas) is a 700 ha forest fragment which has su€ered extinctions
and sharp declines among its larger mammals; the other (the Lacandon rainforest in
Chiapas) is far more extensive and still has near-normal mammalian populations, in-
cluding the full complement of large herbivores (peccaries, tapir and deer). In the Los
Tuxtlas understorey there was no sign of vertebrate browsing, the seedling density was
twice as high, and the diversity of plants was much lower than in the Chiapas forest which
was being browsed. Herbivory appears to be promoting the diversity of understorey plants
by preventing a few species from monopolizing the physical resources, in much the same
way that moderate levels of sheep grazing maintain plant diversity in British chalk
grasslands (Wells, 1971), or predatory star®shes can maintain mollusc diversity in inter-
tidal food webs (Paine, 1966). Eliminating herbivores, therefore, can not only shift the
competitive balance in favour of some plants over others, but it can also reduce overall
¯oristic diversity.

Predation
Several tropical forest studies indicate that carnivorous predators are `keystone species'
that help to structure animal communities, and thereby plant communities. In a wide-
ranging comparative analysis, Emmons (1987) suggested that prey species' densities be-
come very uneven in the absence of large felid predators, and the well-studied example of
BCI con®rms that predation pressure, or its absence, can have an enormous e€ect on prey
populations. Theoretical research further suggests that loss of a few top predators will
cause more knock-on extinctions in species-rich systems than would loss of a few species at
lower trophic levels (Pimm, 1986).
Few tropical ecosystems retain their full natural complement of top predators. In the
neotropics, for example, hunting has left populations of cat, otter, large eagle and caiman
species decimated. Predicting the consequences of this for plant species is especially hard,
because of the indirect e€ects of the top predators on plants and other animals (Terborgh,
1988), and the potential for synergisms with other ecosystem stresses. For example, if
300 Phillips
predator populations are already su€ering direct hunting pressure, human hunting of their
prey species could indirectly cause extinctions among the predators, which in turn could
increase the unevenness of herbivorous species' population densities, in turn altering
herbivore impacts on plants.

CHANGING TREE TURNOVER RATES

Phillips and Gentry (1994) compiled data on tree turnover through time, by averaging tree
mortality and recruitment rates at 40 pantropical sites. The dataset showed an overall
trend toward an increase in turnover rates in the late 20th century, whether analysed as
change in the annualized average rate across all inventoried forests, or as changes within
19 sites that had three or more inventories. Since then, some additional sites have come to
light, and the overall pattern using updated datasets (Phillips, in press) remains largely
unchanged (Fig. 2).
Why tree turnover rates have increased is not clear ± indeed no large-scale observa-
tional study of ecological phenomena is easily amenable to a conclusive causal analysis ±
and so it is also not clear if and how turnover rates might change in the future. In our
original paper we hypothesized that the increase in turnover is related to enhanced pro-
ductivity, stimulated by increasing atmospheric CO2 and I propose more possible ex-
planations later on. In the context of this paper, the ®nding is most signi®cant in
demonstrating how poorly we understand the ecology of tropical forests and our impacts
on them. It shows that current ground-based research programmes are still insucient to
monitor or predict, let alone help prevent, the e€ects of long-term environmental change
on the most diverse ecosystems. In the continuing absence of a reliable science for pre-
dicting the consequences of environmental change, the ®nding suggests that we should
take a precautionary approach and tackle the causes of environmental change sooner
rather than later.
The consequences of increased tree turnover may include substantial changes in both
biological diversity and the carbon cycle, since tropical forests contain globally signi®cant
stores of both diversity and carbon (Pimm and Sugden, 1994). Consideration of the role of
forest dynamics in regulating diversity suggests how some biotic consequences of in-
creasing turnover might be manifested. Empirical evidence strongly indicates that gap
dynamics help to regulate diversity (Denslow, 1987), and that more dynamic forests tend
to be more diverse than less dynamic forests (Phillips et al., 1994b). Therefore, forests with
intrinsically low turnover rates might experience an increase in local species richness, as
more frequent canopy openings should permit a greater variety of species to co-exist. But
turnover rates are changing in ecological, not evolutionary time, and so they cannot lead
to a global increase in species diversity. A reasonable prediction is that in combination
with other factors likely to favour faster-growing plants (see later), increased turnover
could lead to signi®cant losses in biodiversity both regionally and globally. In mature
primary forests, the rarest, slowest-growing tree species (and some animal species, in-
cluding host-speci®c insects and probably including other animal and plant species in-
directly linked to the slowest-growing plants) may lose out to faster-growing opportunistic
tree and liana species which are better suited to exploiting small gaps and unprecendented
levels of atmospheric CO2.
The changing ecology of tropical forests 301

Figure 2. Annual mature tropical forest turnover rates measured since 1957. Tree turnover for each
site is computed as the average of annual mortality and recruitment rates. Turnover rates are
expressed as the annual mean (‹ standard error) of all simultaneously measured rates. The total
number of forest sites monitored is 59; the minimum number of simultaneous monitoring sites is 8
(in 1957±9, 1993), the maximum number of simultaneous monitoring sites is 30 (in 1980). The sites
used to calculate mean annual turnover rates switch through time, since few data are available for
more than 20 years from any one site. Therefore, the trend could conceivably result from a pro-
gressive and very pronounced change in the kinds of forest sampled. In order to control for such an
artefactual explanation, we also looked at change at forest sites with at least two consecutive
measurement intervals. A large majority of the sites with successive intervals had higher turnover in
later measurement intervals than in earlier measurement intervals (ratio of sites is 20:4; z = 3.63,
P <0.001, Wilcoxon matched-pair signed rank test). Further details are given in Phillips (in press).

Interactions between threats and forests

The individual threats do not operate in isolation, either from one another, or from the
pervasive threat posed by outright deforestation. The overview touched on some inter-
actions among the threats that could magnify their impact on biodiversity ± for example
the interactions of climate change with forest fragmentation, or of climate change with
deforestation. In many cases, if the ®rst order threats to plants and animals are themselves
poorly understood, some of the higher order interactive e€ects between them are purely
hypothetical. This section is therefore very speculative. But our poor understanding of the
nature of these synergistic e€ects does not diminish their potential impact on the ecology
of tropical forests. Interactions between one or more threats and forest ecology will be
built from one or more of three basic forms (Fig. 3). In practice, these basic forms will
probably be combined in various ways in complex networks of interactive cause and e€ect.
302 Phillips

Figure 3. Three basic kinds of interaction between threats and forest ecology. See text for ex-
planation and examples.

DIRECT CAUSATION

One of the seven direct threats to tropical forest biodiversity may also be the direct cause
of another threat. For example, on a local scale, once a forest is fragmented tree mortality
is enhanced close to the new forest edge, presumably due to altered exposure to physical
The changing ecology of tropical forests 303
stress factors such as wind and low humidity associated with fragmentation (Lovejoy et
al., 1984; Kapos, 1989; Bierregaard et al., 1992). Similarly, fragmentation alone typically
causes some local animal extinctions, especially in wide-ranging rare species such as top
predators. In some respects it therefore has a similar `defaunatory' impact as hunting (c.f.
`Herbivory and seed predation' under the earlier heading `Two invisible threats').
On larger scales, the phemonenon of recently increasing tree turnover may be explained
by one or more of the other processes. Rising atmospheric concentrations of CO2 could be
enhancing primary productivity, thereby increasing the rate with which wood cycles
through the community. Regional or global climate changes, in turn driven partly by
deforestation and/or rising atmospheric CO2 levels, could also a€ect tree turnover rates.
Alternatively, hunting of large frugivores is certain to a€ect large-fruited animal-dispersed
trees, so that trees with wind-dispersed and smaller animal-dispersed fruits could bene®t
from a competitive release e€ect. Since trees in the second group are more likely to be
earlier successional plants with naturally higher rates of growth and mortality (Foster and
Janson, 1985; Denslow, 1987; Swaine and Whitmore, 1988), increasing tree turnover rates
might partly be a result of shifts in tree species composition caused by the widespread
hunting of the largest tropical forest frugivores.

COMBINATIVE CAUSATION

Interaction between two or more processes may also generate an impact on biodiversity
di€erent from that caused by either alone. For example, low-level hunting pressure could
trigger extinctions among otherwise viable animal populations persisting in forest frag-
ments. Smaller fragments are also relatively more susceptible to biological invasion by
alien species or by species more typical of successional habitat (Janzen, 1983). An
otherwise sustainable level of selective cutting of fruit trees, combined with an otherwise
sustainable level of hunting of its major dispersers, could together eliminate several animal
and plant species, particularly if the fruit trees also happen to be keystone resources for
animals during seasons of scarcity (Howe, 1984; Terborgh, 1986). Logging (selective ex-
traction of plants) combined with drought (likely to become more frequent and more
intense as global warming proceeds) may make otherwise unburnable humid forests vul-
nerable to ®re.
It is also possible that some threats to biodiversity might help defuse, or reverse, the
impact of other threats. Brazil nut tree (Bertholletia excelsa) seeds are extracted from
forests in large quantities by humans and dispersed by rodents (agoutis); new data from
Peru (Enrique OrtõÂz, pers. comm.) suggest that moderate levels of some human activities
(including small-scale forest disturbance, and perhaps hunting of large predators) will lead
to increased agouti populations. Bertholletia regeneration may therefore be improved in
those areas where humans are collecting seeds relative to areas where human impact is
minimal. Bertholletia regeneration may also be boosted by unintentional seed scattering by
collectors, and by other human activities that create small light gaps.

FEEDBACK

In the third kind of interaction between threats and forest ecology, the individual pro-
cesses either feed on themselves in a `vicious circle' of positive feedback, thereby magni-
fying the impact on biodiversity, or eventually moderate themselves in a `virtuous circle' of
negative feedback. For example, increasing turnover may lead to greater opportunities for
304 Phillips
colonization by fast-growing tree species that regenerate preferentially in light gaps, which
in turn have intrinsically higher mortality rates than more shade-tolerant species (Swaine
and Whitmore, 1988). The result could be a further increase in tree turnover. The direct,
physiological e€ects of increasing atmospheric CO2 are more likely to bene®t intrinsically
fast-growing species, boosting stem turnover rates further. Lianas may bene®t from the
more dynamic and resource-rich environment, and their growth would be likely to ac-
celerate tree mortality further (Putz, 1982), thereby further generating conditions that suit
their own growth. This hypothetical positive feedback cycle could then directly cause other
threats to biodiversity. For example, any shift to faster-growing plants may result in net
loss of carbon from both living biomass (fast-growing species are less dense than slow-
growing species, and younger trees are less dense than their older conspeci®cs (Castro et
al., 1993)) and from detrital carbon (CebriaÂn and Duarte, 1995). Primary tropical forests
could therefore become net sources of atmospheric CO2. That, in turn, could have eco-
system and biome-level e€ects via climate change.
In other, more optimistic scenarios, negative feedback imposes an upper limit on the
biotic impact of individual processes. For example, if lianas do bene®t at some point they
will approach saturation of their trellis resource (suitable tree species), thereby halting
liana population growth. Another negative feedback e€ect might begin to modify hunting
pressure on animals once target populations fall below a certain level. Thus, human
hunters who are maximizing their time allocation to hunting e€ort may conclude that it is
no longer worth trying to hunt these now rare animals. This could result in sharply
decreasing pressure on animal population densities as they fall below critical thresholds,
and perhaps allow them to survive inde®nitely ± albeit as newly rare species newly vul-
nerable to stochastic events that could cause their extinction.

Potential outcomes for plants and animals

How might the forests of the future compare with those of the recent past? Predicting the
biotic impacts of poorly understood changes on complex systems is hardly straightfor-
ward! However, some general e€ects on biodiversity resulting from these non-deforesta-
tion threats can be anticipated with some con®dence, from basic ecological principles,
unless preventative and remedial actions are taken. As we saw earlier, rare species and
habitat specialists are particularly vulnerable to extinction and should therefore be prio-
rities for conservation and monitoring (Terborgh and Winter, 1980). Other common-sense
predictions can also be made. Table 1 indicates hypothetical population changes among
some broadly de®ned groups of organisms in tropical forests. When these are compared
for each group across all the threat classes, some groups of organisms appear likely to do
consistently badly, others consistently well. For example, some large-seeded tree species
dispersed by large herbivores are predicted to be at a relative disadvantage from every one
of the threats. Small-seeded, wind-dispersed tropical vines, on the other hand, seem des-
tined to prosper as a result of most of the changes.
This crude tally is o€ered as a ®rst step toward indicating which groups of plants and
animals might most require intervention by conservationists. Obviously, these suggestions
should be treated with caution. For one, even among groups that are predicted to bene®t
as a whole, it is possible that non-deforestation extinctions will still occur. Most im-
portantly, this approach does not take into account all the potential higher-order inter-
actions among threats, simply because they are so hard to predict. In some cases these
The changing ecology of tropical forests 305
Table 1. Hypothetical population changes among native forest organisms in response to threats

Trees Lianas and vines Animals

Dispersed Dispersed Dispersed Dispersed Carnivores Herbivores
by large by wind by large by wind
animals or small animals or small
animals animals
Threat large small large small large small large small large small large small
seed seed seed seed seed seed seed seed
A )1 )1 ? ? )1 ? +1 +2 )1 ? )1 ?
B )2 )2 +1 +2 )2 )2 +1 +2 )2 +1 )2 +2
C )1 )1 )1 )1 ? ? ? ? ? ? )1 )2
D )2 )2 ? +2 )2 )2 ? +2 )2 ? )2 +1
E )1 )1 ? ? )1 )1 ? ? )1 ? )1 ?
F )1 +1 )1 +1 ? +1 +1 +2 )1 ? )1 +1
G )2 ? )1 +1 ? +1 ? +2 )1 ? )1 +1
Sum )10 )6 )2 +5 )6 )3 +3 +10 )8 +1 )9 +3
A = Selective extraction of plants (especially felling trees).
B = Selective extraction of animals (especially hunting animals).
C = Biological invasion.
D = Fragmentation (including isolation of populations, and enhanced edge e€ects).
E = Climate change (especially increased drought).
F = Atmospheric change (especially increased CO2 levels).
G = Increasing tree turnover.
? = E€ect, if any, as likely to be positive as negative.
+1 = Possibly bene®ting (i.e. if any e€ect, more likely to bene®t than su€er).
+2 = Probably bene®ting.
)1 = Possibly su€ering (i.e. if any e€ect, more likely to su€er than bene®t).
)2 = Probably su€ering.

e€ects could switch the balance of winners and losers. For example, large-seeded trees
could still bene®t, if they also have good intrinsic defence mechanism against climbers.
Palm trees, which usually remain free of climbers, and some myrmecophilous plants ±
which provide ant species with a home in return for their protection against various biotic
threats ± are some candidates for `surprise winners' as forest ecosystems change in reaction
to externally driven processes.
These considerations lead to some practical suggestions for future monitoring and
conservation in tropical forest reserves. There is an urgent need for a greater emphasis on
monitoring liana populations (Gentry, 1991; Hegarty and CaballeÂ, 1991) which are
overlooked in at least 90% of permanent inventory forest plots. Many lianas are likely to
bene®t in degraded forests by virtue of their opportunistic life-history characteristics. Since
some animals and plants are disproportionately likely to su€er as primary forests become
degraded, lists of `most-at-risk' species and ecological guilds would also be useful guides
for monitoring and conservation programmes. If major ecological imbalances are sus-
pected, there would be a case for conservationists to take an interventionist approach to
correct the imbalance. Examples might include culling seed predators that no longer have
306 Phillips
natural predators, and enrichment planting with plant species that are outcompeted by
other plant species under an altered herbivore regime. Such intervention will surely involve
intense scienti®c and ethical debate, but may become increasingly necessary.

Conclusions
Change and disturbance in tropical forests is not necessarily harmful. Small and large
disturbances (tree-falls, cyclones, droughts, and so on) are part of the natural regime, and
even those areas that today apparently support pristine forest communities have been the
sites of occasional human activity for thousands of years. So long-term biological de-
gradation is not an inevitable consequence of human impact, and all biologists would
probably agree that even a degraded forest is better than no forest at all. Moreover, the
time-scales on which these threats impact biodiversity may make them less immediately
dangerous than deforestation. Small patches of primary forest may maintain surprising
numbers of plant species for decades or centuries in the face of various threats (Dodson
and Gentry, 1991; Turner et al., 1994); this time window may prove to be a critical help for
conservation rescue e€orts. However, this paper has shown that the current level of non-
deforestation impacts on forests has already caused substantial biodiversity losses in many
tropical localities, and that worse is to come.
The causes of forest degradation are dangerous in part because their biological e€ects
are invisible from above. This makes them dicult for scientists to monitor and gives them
little public salience; therefore they are easily overlooked. We must not forget that `the
functioning of rainforests depends on innumerable intricate relationships .... Tropical
foresters and reserve designers can not a€ord to ignore them.' (Leigh et al., 1993). It is
these intricate relationships, more than simply their lush vegetation and tall trees, which
help give tropical forests their special aura of mystery and vitality. As tropical forests
degrade, these relationships quietly dissolve and simplify, and so eventually forests may be
destroyed from the inside out rather than from the outside in. We must therefore make a
critical distinction between monitoring deforestation, which mostly requires coordinated
application of remote-sensing technology, and monitoring biological degradation, which
will require long-term human and ®nancial investments in ®eld research at a large network
of sites. Furthermore, in seeking to improve human standards of living while slowing
outright deforestation (for example, by supporting `extractive reserves' and indigenous
peoples' land tenure rights), the conservation movement should not lose sight of the need
to invest in large protected areas that are insulated from as many threats as possible.

Acknowledgements
This paper has bene®ted from suggestions by Alan Grainger, Sam Rose and Angus Da-
vies. I thank Enrique OrtõÂz for permission to mention his unpublished observations on
Brazil nut trees.

References
Atkinson, I.A.E. (1977) A reassessment of factors, particularly Rattus rattus L., that in¯uenced the
decline of endemic forest birds in the Hawaiian islands. Paci®c Sci. 31, 109±33.
The changing ecology of tropical forests 307
BaleÂe, W. (1989) The culture of Amazonian forests. In Resource Management in Amazonia: In-
digenous and Folk Strategies (D.A. Posey and W. BaleÂe, eds) pp. 1±21. New York: New York
Botanical Gardens.
BaleÂe, W. (1994) Footprints in the Forest. Ka'apor Ethnobotany ± The Historical Ecology of Plant
Utilization by an Amazonian People. New York: Columbia University Press.
Bazzaz, F.A. (1990) The response of natural ecosystems to the rising global CO2 levels. Ann. Rev.
Ecol. Syst. 21, 167±96.
Beerling, D.J. and Woodward, F.I. (1993) Ecophysiological responses of plants to global environ-
mental change since the last glacial maximum. New Phytol. 125, 641±8.
Bierregaard, R.O., Lovejoy, T.E., Kapos, V., Santos, A. and Hutchings R. (1992) The biological
dynamics of tropical rainforest fragments. Bioscience 42, 859±66.
Bush, M.A., Piperno, D.R., Colinvaux, C.A., De Oliviera, P.E., Krissek, L.A., Miller, M.C. and
Rowe, W.L. (1993) A 14,300-yr paleoecological pro®le of a lowland tropical lake in Panama.
Ecol. Monogr. 62, 251±75.
Castro, F.de, Williamson, G.B. and Moraes de Jesus, R. (1993) Radial variation in the wood speci®c
gravity of Joannesia princeps: the roles of age and diameter. Biotropica 25, 176±82.
CebriaÂn, J. and Duarte, C.M. (1995) Plant growth-rate dependence of detrital carbon storage in
ecosystems. Science 268, 1606±8.
Chapman, C.A. and Chapman, L.J. (1995) Survival without dispersers: seedling recruitment under
parents. Conserv. Biol. 9, 675±8.
Clinebell, R., Phillips, O.L., Gentry, A.H., Stark, N. and Zuuring, H. (1995) Prediction of neo-
tropical woody plant diversity from climatic and soil data. Biodiv. Conserv. 4, 59±90.
Condit, R., Hubbell, S.P. and Foster, R. (1996) Changes in tree species abundance in a neotropical
forests ± impact of climate-change. J. Trop. Ecol. 12, 231±56.
Condon, M.A., Sasek, T.W. and Strain, B.R. (1992) Allocation patterns in two tropical vines in
response to increased atmospheric CO2. Funct. Ecol. 6, 680±85.
Connell, J.H. (1971) On the role of natural enemies in preventing competitive exclusion in some
marine animals and in rain forest trees. In Dynamics and Numbers in Populations (P.J. den Boer
and G.R. Gradwell, eds). Proceedings of the Advanced Study Institute, Osterbeek Wageningen,
The Netherlands, 1970. Centre for Agricultural Publication and Documentation.
Cronk, Q. and Fuller, J. (1996) Plant Invaders: the Threat to Natural Ecosystems. People and Plants
Conservation Manual No. 2. London: Chapman and Hall.
Dale, V.H., Pearson, S.M., O€erman H.L. and O'Neill, R.V. (1994) Relating patterns of land-use
change to faunal biodiversity in the Central Amazon. Conserv. Biol. 8, 1027±36.
Denslow, J.S. (1987) Tropical rainforest gaps and tree species diversity. Annu. Rev. Ecol. Syst. 18,
431±51.
Dirzo, R. and Miranda, A. (1990) Contemporary neotropical defaunation and forest structure,
function, and diversity ± a sequel to John Terborgh. Conserv. Biol. 4, 444±7.
Dirzo, R. and Miranda, A. (1991) Altered patterns of herbivory and diversity in the forest under-
storey: a case study of the possible consequences of contemporary defaunation. In Plant ±
Animal Interactions: Evolutionary Ecology in Tropical and Temperate Regions (P.W. Price, T.M.
Lewinsohn, G.W. Fernandes and W.W. Benson, eds), pp. 273±87. New York: Wiley.
Dodson, C. and Gentry, A.H. (1991) Biological extinction in western Ecuador. Ann. Missouri Bot.
Garden 78, 273±95.
Emmons, L.H. (1987) Comparative feeding ecology of felids in a neotropical rainforest. Behav. Ecol.
Sociobiol. 20, 271±83.
FAO. (1993) Forest Resources Assessment, 1990: Tropical Countries. FAO Forestry Paper 112.
Rome: FAO.
Flenley, J.R. (1979) The Equatorial Rainforest: a Geological History. London: Butterworths.
308 Phillips
Foster, R.B. (1982) Famine on Barro Colorado Island. In The Ecology of Tropical Forest: Seasonal
Rhythms and Long-Term Changes (E.G. Leigh, Jr, A.S. Rand and D.M. Windsor, eds), pp. 201±
12. Washington DC: Smithsonian Institution Press.
Foster, S.A. and Janson, C.H. (1985) The relationship between seed size and establishment condi-
tions in tropical woody plants. Ecology 66, 773±80.
Gentry, A.H. (1983) Dispersal ecology and diversity in neotropical forest communities. Sonderbd.
Naturwissenschaftlichen Vereins in Hamburg 7, 303±14.
Gentry, A.H. (1991) The distribution and evolution of climbing plants. In The Biology of Vines (F.E.
Putz and H.E. Mooney, eds), pp. 3±49. Cambridge: Cambridge University Press.
Gentry A.H. (1995) Patterns of diversity and ¯oristic composition in neotropical montane forests. In
Biodiversity and Conservation of Neotropical Montane Forests (S.P. Churchill, H. Balslev, E.
Forero and J.L. Luteyn, eds), pp. 103±26. New York: New York Botanical Garden.
Gentry A.H. and VaÂsquez, R. (1988) Where have all the Ceibas gone? A case history of misman-
agement of a tropical forest resource. Forest Ecol. Manage. 23, 73±6.
Glanz, W.E. (1982) The terrestrial mammal fauna of Barro Colorado Island: censuses and long-term
changes. In The Ecology of a Tropical Forest: Seasonal Rhythms and Long-term Changes (E.G.
Leigh, Jr, A.S. Rand and D.M. Windsor, eds), pp. 455±68. Washington DC: Smithsonian
Institution Press.
Glanz, W.E. (1990) Neotropical mammalian densities: how unusual is the Barro Colorado Island,
Panama, community? In Four Neotropical Rainforests (A.H. Gentry, ed.), pp. 287±313. New
Haven: Yale University Press.
Goulding, M. (1980) The Fishes and the Forest. Berkeley, CA: University of California Press.
Grainger, A (1993) Rates of deforestation in the humid tropics: estimates and measurements. Geogr.
J. 159, 33±44.
Grainger, A (1996) The degradation of tropical rain forest in Southeast Asia: taxonomy and ap-
praisal. In Land Degradation in the Tropics (M.E. Eden and J.T. Parry, eds). London: Mansell
Publications.
Grainger, A., Rose, S., Trisurat, Y. and Brockelman, W.Y. (1995) Mapping spatial variation in
tropical rain forest biodiversity using GIS techniques. In Proceedings of the IUFRO Symposium
on Measuring and Monitoring Biodiversity in Tropical and Temperate Forests, 28 August±8
September 1994, Chiang Mai, Thailand (T.J.B. Boyle and B. Boontawee, eds), pp. 335±54.
Bogor, Indonesia: CIFOR.
Graybill, D.A. and Idso, S.B. (1993) Detecting the aerial fertilization e€ect of atmospheric CO2
enrichment in tree-ring chronologies. Global Biogeochem. Cycles 7, 81±95.
Guilderson, T.P., Fairbanks, R.G. and Rubenstone, J.L. (1994) Tropical temperature variations
since 20,000 years ago modulating interhemispheric climate change. Science 263, 663±5.
Hartshorn, G. (1992) Possible e€ects of global warming on the biological diversity in tropical forests.
In Biological Diversity and Global Warming (R.L. Peters and T.L. Lovejoy, eds), pp. 137±45.
New Haven: Yale University Press.
Hegarty, E.E. and CaballeÂ, G. (1991) Distribution and abundance of vines in forest communities. In
The Biology of Vines (F.E. Putz and H.E. Mooney, eds), pp. 313±36. Cambridge: Cambridge
University Press.
Horn, S.P. and Sanford, R.L. (1992) Holocene ®res in Costa Rica. Biotropica 245, 354±61.
Howe, H.F. (1984) Implications of seed dispersal by animals for tropical reserve management. Biol.
Conserv. 30, 261±81.
Hubbell, S.P. and Foster, R.B. (1990) The fate of juvenile trees in a Neotropical forest: implications
for the natural maintenance of tropical tree diversity. In Reproductive Ecology of Tropical Forest
Plants (K.S. Bawa and M. Hadley, eds). Paris: Parthenon Publishing Group.
Janzen, D.H. (1970) Herbivores and the number of tree species in tropical forests. Am. Naturalist
104, 501±27.
The changing ecology of tropical forests 309
Janzen, D.H. (1983) No park is an island: increase in interference from outside as park size de-
creases. Oikos 41, 402±10.
Janzen, D.H. (1986) The eternal external threat. In Conservation Biology: the Science of Scarcity and
Diversity (M.E. SouleÂ, ed.), pp. 286±303. Sunderland, MA: Sinauer Associates, Inc.
Johnson, H.B., Polley, H.W. and Mayeux, H.C. (1993) Increasing CO2 and plant ± plant interac-
tions: e€ects on natural vegetation. Vegetatio 104, 157±70.
Jorgenson, P.M., Ulloa U., C., Madsen, J.E. and Valencia R., R. (1995) A ¯oristic analysis of the
high Andes of Ecuador. In Biodiversity and Conservation of Neotropical Montane Forests (S.P.
Churchill, H. Balslev, E. Forero and J.L. Luteyn, eds), pp. 221±38. New York: New York
Botanical Garden.
Kapos, V. (1989) E€ects of isolation on the water status of forest patches in the Brazilian Amazon. J.
Trop. Ecol. 5, 173±85.
Kubitzki, K. and Ziburski, A. (1994) Seed dispersal in ¯oodplain forests of Amazonia. Biotropica 26,
30±43.
Leigh, E.G. Jr, Joseph Wright, S., Herre, E.A. and Putz, F.E. (1993) The decline of tree diversity on
newly isolated tropical islands: a test of a null hypothesis and some implications. Evol. Ecol. 7,
76±102.
Lorence, D.H. and Sussman, R.W. (1986) Exotic species invasion into Mauritius wet forest rem-
nants. J. Trop. Ecol. 2, 147±62.
Lovejoy, T.E., Rankin, J.M., Bierregaard, R.O., Brown, K.S., Emmons, L.H. and van der Voort, M.
(1984) Ecosystem decay of Amazon forest remnants. In Extinctions (M.H. Nitecki, ed.). Chi-
cago: University of Chicago Press.
MacArthur, R.H. and Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton, New
Jersey: Princeton University Press.
Meher-Homji, V.M. (1991) Probable impact of deforestation on hydrological processes. Climatic
Change 19, 163±74.
Myers, N. (1988) Threatened biotas: `hotspots' in tropical forests. Environmentalist 8, 1±20.
Myers, N. (1989) Deforestation Rates in Tropical Countries and their Climate Implications. London:
Friends of the Earth.
Myers, N. (1993) Tropical forests: the main deforestation fronts. Environ. Conserv. 20, 9±16.
O'Brien, S.T., Hayden, B.P. and Shugart, H.H. (1992) Global climatic change, hurricanes, and a
tropical forest. Climatic Change 22, 175±90.
Padoch, C. (1988) Aguaje (Mauritia ¯exuosa L.f.) in the economy of Iquitos, Peru. Adv. Econ. Bot. 6,
214±24.
Paine, R.T. (1966) Food web complexity and species diversity. Am. Naturalist 100, 65±75.
Peres, C.A. (1990) E€ects of hunting on western Amazonian primate communities. Biol. Conserv. 54,
47±59.
Phillips, O.L. (1995) Evaluating turnover in tropical forests. Science 268, 894±5.
Phillips, O.L. (in press) Increasing tree turnover in tropical forests as measured in permanent plots.
In Measuring and Monitoring Biological Diversity (F. Dallmeier and J. Comiskey, eds). Wa-
shington, DC: Smithsonian Institution Press (in press).
Phillips, O.L. and Gentry, A.H. (1994) Increasing turnover through time in tropical forests. Science
263, 954±8.
Phillips, O.L., Gentry, A.H., Reynel, C., Wilkin, P. and GaÂlvez Durand B.C. (1994a) Quantitative
ethnobotany and Amazonian conservation. Conserv. Biol. 8, 225±48.
Phillips, O.L., Hall, P., Gentry, A.H., VaÂsquez, R. and Sawyer, S. (1994b) Dynamics and species
richness of tropical forests Proc. Acad. Sci. USA 91, 2805±9.
Pimm, S. (1986) Community stability and structure. In Conservation Biology: the Science of Scarcity
and Diversity (M.E. SouleÂ. ed.), pp. 309±29. Sunderland, MA: Sinauer Associates, Inc.
Pimm, S. and Sugden, A (1994) Tropical diversity and global change. Science 263, 933±4.
310 Phillips
Putz, F.E. (1982) Natural History of Lianas and Their E€ects on Tropical Forest Dynamics. Ph.D.
thesis, Cornell University, Ithaca, New York.
Putz, F.E., Leigh Jr. E.G. and Wright, S.J. (1990) The arboreal vegetation on 70-year old islands in
the Panama canal. Garden 14, 18±23.
Redford, K.H. (1992) The empty forest. Bioscience 42, 412±22.
Roth, P. (1984) ReparticËao do habitat entre Psitacideos simpatricos no sul da Amazonia. Acta
Amazonica 14, 175±221.
Salati, E. and Nobre, C.A. (1991) Possible climatic impacts of tropical deforestation. Climatic
Change 19, 177±96.
Sanford, R.L., Jr, Saldarriage, J., Clark, K.E., Uhl, C. and Herrera, R. (1985) Amazon rain-forest
®res. Science 227, 53±5.
Saunders, D.A., Hobbs, R.J. and Margules, C.R. (1991) Biological consequences of ecosystem
fragmentation: a review. Conserv. Biol. 5, 18±32.
Sayer, J.A. and Whitmore, T.C. (1991) Tropical moist forests: destruction and species extinction.
Biol. Conserv. 55, 199±213.
Simberlo€, D. (1992) Extinctions, survival, and e€ects of birds introduced to the Masacarenes.
Oecologia 13, 663±78.
Skole, D. and Tucker, C. (1993) Tropical deforestation and habitat fragmentation in the Amazon:
satellite data from 1978 to 1988. Science 260, 1905±10.
Stute, M., Forster, M., Frischkorn, H., Serejo, A., Clark, J.F., Schlosser, P., Broecker, W.S. and
Bonani, G. (1995) Cooling of tropical Brazil (5°C) during the last glacial maximum. Science 269,
379±83.
Swaine, M.J. and Whitmore, T.C. (1988) On the de®nition of ecological species groups in tropical
rain forests. Vegetatio 75, 81±6.
Terborgh, J. (1986) Keystone plant species in the tropical forest. In Conservation Biology: the Science
of Scarcity and Diversity (M.SouleÂ,ed.), pp. 330±44. Sunderland, MA: Sinauer Associates.
Terborgh, J. (1988) The big things that run the world: a sequel to E.O. Wilson. Conserv. Biol. 2, 402±
3.
Terborgh, J. (1992) Maintenance of diversity in tropical forests. Biotropica 24, 283±92.
Terborgh, J. and Winter, B. (1980) Some causes of extinction. In Conservation Biology: an Evolu-
tionary ± Ecological Perspective (M. Soule and B. Wilcox, eds), pp. 119±33. Sunderland, MA:
Sinauer Associates.
Thiollay, J.M. (1989) Area requirements for the conservation of rain forest raptors and game birds in
French Guiana. Conserv. Biol. 3, 128±37.
Tuomisto, H., Ruokolainen, K., Kalliola, R., Linna, A., Danjoy, W. and Rodriguez, Z. (1995)
Dissecting Amazonian biodiversity. Science 269, 63±6.
Turner, I.M. and Tan, H.T.W. (1992) Ecological impact of alien plant species in Singapore. Paci®c
Sci. 46, 389±90.
Turner, I.M., Tan, H.T.W., Wee, Y.C., Ibrahim, Ali Bin, Chew, P.T. and Corlett, R.T. (1994) A
study of plant species extinction in Singapore: lessons for the conservation of tropical biodi-
versity. Conserv. Biol. 8, 705±12.
VaÂsquez, R. and Gentry, A.H. (1989) Use and misuse of forest-harvested fruits in the Iquitos area.
Conserv. Biol. 3, 350±61.
Vitousek, P.M. (1988) Diversity and biological invasions of oceanic islands. In Biodiversity (E.O.
Wilson, ed.), pp. 181±9. Washington, DC: National Academy Press.
von Roosmalen, M. (1985) Habitat preferences, diet, feeding strategy and social organization of the
black spider monkey (Ateles paniscus paniscus Linnaeus 1758) in Surinam. Acta Amazonica 15,
(supplement).
Wayne Polley, H., Johnson, H.B., Marino, B.D. and Mayeux, H.S.(1993) Increase in C3 plant water-
use eciency and biomass over Glacial to present CO2 concentrations. Nature 361, 61±3.
The changing ecology of tropical forests 311
Wells, T.C.E. (1971) E€ects of sheep grazing and mechanical cutting on the structure and botanical
composition of chalk grassland. In The Scienti®c Management of Animal and Plant Communities
for Conservation (E. Du€y and A.S. Watt, eds), pp. 497±515. Oxford: Blackwell Scienti®c
Publications.
Wilcove, D.S. (1987) From fragmentation to extinction. Nat. Areas J. 7, 23±9.
Williams-Linera, G. (1990) Vegetation structure and environmental conditions of forest edges in
Panama. J. Ecol. 78, 356±73.
Windsor, D.M. (1990) Climate and Moisture Variability in a Tropical Forest: Long-Term Records
from Barro Colorado Island, Panama. Smithsonian Contributions to the Earth Sciences, 29.
Smithsonian Institution, Washington, DC.
Woodward, F.I. (1987) Stomatal numbers are sensitive to increases in CO2 from pre-industrial levels.
Nature 327, 617±8.
Woodward F.I. (1992) Predicting plant responses to global environmental change. New Phytol. 122,
239±51.
Yost, J. and Kelley, P. (1983) Shotguns, blowguns, and spears: the analysis of technological e-
ciency. In Adaptive Responses of Native Amazonians (R.B. Hames and W.T. Vickers. eds), pp.
189±224. New York: Academic Press.
Young, K. (1992) Biogeography of the montane forest zone of the eastern slopes of Peru. Memorias
del Mueso de Historia Natural, U.N.M.S.M. (Lima) 21, 119±40.
Young, K., Church, W.B., Leo, M. and Moore, P.F. (1994) Threats to Rio Abiseo National Park,
northern Peru. Ambio 23, 312±4.
Zhang, K. (1986) The in¯uence of deforestation on tropical rainforest on local climate and disaster
in Xishuangbanna region of China. Climatol. Notes 35, 223±36.

View publication stats