LIKE a Gothic cathedral rising from the middle of a slum, a perfect square of
rainforest stands in a broad, weedy pasture. Bromeliads ornament the branches of
the vaulted canopy, and buttresses support the largest trunks. The square is
virgin rainforest, as the dictionaries define the term, but it has been stripped
of its context鈥揳 common fate where human activity meets the forest.
But this fragment of rainforest, just north of Manaus in the centre
of the Brazilian Amazon, is different. It is a part of one of the most famous
experiments in ecology and it is teaching some unexpected lessons.
The experiment is the brainchild of Thomas Lovejoy, an ecologist from the
Smithsonian Institution. Almost two decades ago, he set out to learn what would
happen to fragments of forest ecosystems left behind after the men with
chainsaws had done their worst. At the time, ecologists were abuzz over a theory
known as island biogeography, proposed barely a decade earlier by ecologists
Robert MacArthur and E. O. Wilson.
Advertisement
MacArthur and Wilson predicted that islands in the ocean or isolated forest
fragments should have fewer species than a comparable area of mainland or
continuous forest. On an island, there is a risk that a run of bad luck will
wipe out a particular species forever. On the mainland or in continuous forest,
even if bad luck strikes an equivalent area, other members of the same species
can always move in from other nearby areas.
In the late 1960s, Wilson and his graduate student, Daniel Simberloff鈥搉ow at
Florida State University鈥搕ested the theory when they cut the vegetation of
mangrove islands off Florida in two using chainsaws. Sure enough, the new,
smaller islands held fewer insect species.
But would the same hold true for tropical rainforest? The answer was of more
than theoretical interest. Conservationists needed to know if they had to
protect some critical minimum size of forest tract in order to preserve a real,
functioning rainforest ecosystem instead of a dying hulk.
Lovejoy persuaded several ranchers who were clearing land to leave fragments
of forest which formed squares of exactly 1, 10, or 100 hectares. Ever since, a
small army of researchers, graduate students, and technicians has been following
the fate of these fragments to try to answer the question, how big is big
enough?
Fifteen years have now passed and some clear answers have emerged, says
Claude Gascon, a Canadian ecologist who now lives in Manaus and coordinates the
project.
As everyone expected, some species were squeezed out of the smaller
fragments. Large, mixed-species flocks of insect-eating birds dispersed within 2
years of isolation in the 1 and 10-hectare fragments, leaving only a few
stragglers. Many other single-species flocks of insectivorous birds also
vanished from the fragments over the first 3 to 6 years.
Living space
And even the 100-hectare patches proved too small for species such
as army ants and the retinue of ant-following birds that depend on them.
Capuchin monkeys, which need a lot of space, also vanished. Not all species
suffered, however. Hummingbirds appeared unaffected. Small mammals increased in
the smallest fragments. 鈥淚t鈥檚 a trash habitat, and small mammals are more
abundant because there are more insects there,鈥� says Gascon.
To the researchers鈥� surprise, however, the size of a fragment turned out not
to be the most powerful force determining the fate of the plants and animals
within it. Instead, the future of most individuals depended more on how close
they were to the edge of the forest and on what sort of habitat surrounded the
forest fragment.
When rainforest trees topple to make way for pasture, they expose the naked
flanks of their neighbours that still stand. These trees at the edge of the
forest are five times as likely to be uprooted in a windstorm as trees in
unbroken forest, says Gascon. But even if they remain upright, light streams
onto the forest floor past the leafless trunks, and wind stirs the stifling,
humid air.
For the first few years, the forest floor is warmer and drier for at least 60
metres in from the edge. Gradually, though, a tangle of undergrowth seals over
the cut edge and the microclimate returns to a more normal one, according to
Valerie Kapos of the University of Cambridge and her colleagues.
This burst of light and higher temperature has a profound effect on the tree
seedlings that dot the forest floor. 鈥淎 lot of these small seedlings will die
off, and others that are adapted to drier conditions will take over,鈥� says
Gascon. For example, seedlings of Brazil nut survive much better near the forest
edge, whereas jacaranda seedlings do not.
The researchers found these 鈥渆dge effects鈥� in the normal course of their
experiments. However, it took a stroke of well-timed bad luck to reveal the
second key factor that affects the outcome of fragmentation: the nature of the
surrounding habitat. When Brazil鈥檚 economy hit the skids in the mid-1980s, the
government stopped subsidising landowners who cleared rainforest and converted
the land to pasture. At about the same time, some ranchers began to discover
that their pastures were turning to dust and becoming virtually worthless after
just a few years of hard use.
The result was bad news for Lovejoy鈥檚 experimental design. The ranchers lost
heart and stopped clearing, and nearly half the planned forest fragments were
never cut off from the rest of the forest. Several others became less isolated
as ranchers abandoned pastures and forest regrew around the fragments. This gave
the researchers an unexpected opportunity to compare the fate of fragments set
in different sorts of habitat.
They found that the nature of the surrounding 鈥渕atrix鈥� is critical to the
fate of the fragment. The more forest-like it is, the less the impact of
isolation. Fragments surrounded by tall secondary growth dominated by the
fast-growing tree Cecropia retain more species than those in areas
dominated by the spreading shrub Vismia or pasture grasses.
Even some species that disappeared soon after isolation gradually recolonised
the fragments that were surrounded by 鈥渂enign鈥� matrix. The mixed-species flocks
of insect-eating birds, for example, began to re-form after six years on forest
fragments set in Cecropia. Today, all the common flocking species have
returned to the 10-hectare fragments.
Gascon says he can even predict which species will succeed in forest
fragments by knowing how well they do in the matrix habitat. Many frogs, for
example, live in pastures just as readily as in forest.
By contrast, birds living in the understorey of the forest don鈥檛 cross even
short stretches of pasture. The result is that birds decline and frogs flourish
in fragments surrounded by pasture. This pattern is even more pronounced in
small mammals: an amazing two-thirds of the variation in the abundance of
species in forest fragments can be explained statistically by how common they
are in the matrix鈥揳 correlation that is almost unheard of in ecology.
These results have changed the way the project workers think about forest
fragments and their matrix. 鈥淚 think we鈥檙e talking more in terms of filters than
barriers,鈥� says Gascon. Most monkeys, for example, travel in treetops and need
well-developed forest regrowth to move to and from a fragment. Anything less
than that filters them out.
鈥淲e can now start looking for other correlates of abundance in the matrix.
Does it have to do with body size, or food habits, or physiology? If we can
answer that, we can predict in the future which species are likely to be
vulnerable,鈥� says Gascon.
Today, Gascon and his co-workers are no longer looking for a simple answer to
the question of how big is big enough. The effects of forest fragmentation have
turned out to be more complex than expected. But the new, more sophisticated
picture of matrix habitats and edge effects provides better tools for
conservationists who want to preserve as many rainforest species as
possible.