COLUMN ONE : When the Forest Is Carved Up : Giant rain forests are being sliced into ever smaller fragments. In the Amazon, biologists are studying what happens to the remaining plant and animal life--and the signs are not encouraging.

A researcher squeezed into a truck for a jostling, two-hour ride from the rain forest back to town and insisted on clutching on his lap a bulky, plastic-covered pile of leaves and stems stacked between newspapers.

As a young biology student, he had been admonished by a professor to guard “above all else” such specimens from the Amazon collected for identification and preservation. The truck’s cab where personal belongings had been stashed was not to be trusted.

“He told me they would be in a museum 200 years from now,” said Christopher Dick, who will be entering biology graduate studies at Harvard University this fall. “They will be all that is left.”

Such pessimism, or perhaps realism, is driving scientists to come here and observe how life within a forest fares as the ecosystem is carved into smaller and smaller fragments.

The answers are important because forests throughout the world, including those in the Northwest United States, are increasingly fragmented as they are cut for timber, fuel or agriculture.

Hopeless of saving all or even much of what remains, scientists here want to know how to manage wildlife within forest fragments. They are trying to learn how plants, animals and insects in an isolated reserve interact with surrounding lands.

Can corridors connecting small reserves keep animal populations healthy by giving them access to more forests? Can certain kinds of trees be planted to make cleared lands productive again? Can wildlife be successfully reintroduced into reserves they had abandoned for lack of food or space?

The 25-year experiment here in Brazil’s central Amazon began in 1979, when scientists persuaded cattle ranchers to leave fragments of forest within tens of thousands of acres they were clearing for grazing. The experiment benefited from a spottily enforced Brazilian law that requires ranchers to leave half of their land as forest.

Scientists are studying 12 fragments, ranging from 2.5 acres to 247 acres, at a site about a two-hour ride from the city of Manaus. Given current rates of forest loss, much of what is eventually left may be similarly isolated patches.

“In the future, most rain forest is probably going to be chopped up into small pieces about the size of what we are studying,” said Rob Bierregaard, director of the project, run by the Smithsonian Institution and Brazil’s National Institute for Research in the Amazon.

“What we find in these small, isolated reserves is that things tend to fall apart,” the scientist said. “The web of life begins to unravel.”

The threat is obvious on the turbulent ride into the forest over barely passable dirt roads. Hills covered with clumps of grass and trees broken like snapped pencils testify to the cattlemen’s campaign to transform the forest into pasture. The soil is too poor to sustain many cattle, and more forest must be burned to grow grass for the herds.

The untouched forest farther into the interior is refreshingly cooler and noisy with birds. Rain falls in sudden, drenching bursts, kicking up a wind that could be treacherous as trees or branches are knocked over.

In the pre-dawn hours, distant monkeys howl in a spooky chorus that sounds like the wind. Spiders here are social, working as teams in huge webs. Macaws and toucans are sprinkled sparsely among the trees. Jaguars and mountain lions are reclusive, their presence betrayed only by their tracks and their roars.

Researchers had hoped that the project would tell them how large a reserve had to be to sustain most of the forests’ species. One fragment was to be nearly 25,000 acres.

But some cattle ranchers went bankrupt after government grazing subsidies were withdrawn, and the largest parcel was never isolated. Ironically, the scientists have resorted to paying workers to slash and burn some land around the smaller fragments to maintain their isolation and keep the experiment on track.

The project has turned up disturbing signs of deterioration. Within the fragments, the canopy of trees is more open. Trees have been knocked down by the wind because there is little buffer around the reserves’ edges. The infusion of sunlight dries up the forest and changes the mixture of plants. Trees die in greater numbers in the small reserves.

A break in the forest as small as a few hundred feet has been enough to restrict the movement of birds that live under the canopy and insects that, like the birds, pollinate the plants.

In cooler climates such as the United States, forests reproduce by seeds spread by the wind. But in tropical forests, native trees of the same species are divided by great distances and depend on animals for seed dispersal and pollination. Once those animals disappear, so do the plants.

“If you are missing these birds and insects, it suggests big problems for the plants in the community,” Bierregaard said.

Fragmentation is particularly perilous for tropical forests because they contain such a diversity of interdependent life. Scientists on this project have found about 800 kinds of trees here. In all of North America, there are about 650 tree species.

Because small fragments do not contain all the tree species normally represented, animals or insects associated with those that are missing will succumb.

About half the monkeys vanished from one of the smaller forest patches because there was not enough fruit for them to feed on. Wild pigs also could not make it, and with them went tree frogs dependent on the pigs’ wallows.

Native butterflies, with too few plants for feeding, declined in number and were replaced by other kinds of butterflies that thrive in disturbed habitat. An attempt to reintroduce some species of birds into a small reserve failed; the birds needed a larger range.

On the positive side, researchers discovered that corridors connecting the fragments to contiguous forest will sustain much wildlife even within the small patches.

Corridors could be vital to some monkeys that refuse to come down from the trees to walk across open land. Similarly, many birds that live under the canopy will not leave its protective cloak. If they did, falcons that patrol the borders would snatch them for a meal.

Tropical forests are important to the planet because they contain about half the world’s species. No one is really sure why, but scientists speculate that it may be because the lands near the Equator were never covered by glaciers or because tropical forests lack seasons and extreme cold and hot temperatures.

Some scientists theorize that species, vying for the same niche in the tropics, failed to drive competing species to extinction because tropical forests are regularly disturbed by rivers and the winds.

Just as one species would gain an advantage, it would be reduced in number by falling trees or other disturbances and the competition with another species would start anew. Extinction was thwarted.

Whatever the reason, scientists want to learn more about these species before most are gone. The loss of forests has declined in Brazil in recent years, but wildlife is still vanishing. Illegal hunting, often for food, can be nearly as destructive as wholesale forest clearing.

“It’s actually an underrated problem,” said Bierregaard. “There are huge areas where you fly over where the forest is intact. But if you go down there, a lot of important species are missing.”

To feed Brazil’s booming population without having to resort to decimation of virgin forests, project researchers are probing ways to make cleared lands productive.

The challenge is considerable. Near one of the project’s reserves, what had been pasture now consists of tall green grass dotted by an occasional brown, dead tree.

“It grows back, but there are only two or three species of trees in there,” said Dick, the biology student. “You don’t know how long it takes for the other 900 or so species in the forest to re-establish. Nobody knows how old tropical trees are because they have no tree rings. There are no seasons.”

In temperate climates such as North America’s, rings are formed because of varying temperatures throughout the year. But in tropical forests, the temperature is fairly constant. Rings, which help to date trees, are absent.

Knowing such things as the age of trees is important for sustainable lumbering--using trees for timber without destroying their ability to reproduce.

Julieta Benitez-Malvido, a biology graduate student from Cambridge University, is planting commercially important native trees in secondary forests to see if they can take hold.

Growing tropical trees in plantations has proved extremely difficult, she said, and some species may vanish before people know how to cultivate them.

“Maybe mahogany will become extinct before we know how to produce it commercially,” she said, sitting in one of the project’s comfortable camps, where researchers sleep on hammocks and cook on a gas stove.

Just identifying all species is enormously daunting. Researchers have collected 80,000 samples of trees. Substantially less than half have been identified.

Only one or two experts in the world are even capable of identifying some species. At least 25 tree species examined are believed to be new to science.

“It’s rather depressing when you see how universities and museums are losing support for training,” said Bierregaard. “There just isn’t a new generation of taxonomists or systematists coming along to replace the current generation. So knowledge is being lost and there is not enough reward or job openings for the army of systematists or taxonomists you would need to do the job.”

After they are identified, the tree specimens collected by Dick and others will be stored in herbariums, a kind of plant museum where samples are kept under climate-controlled conditions for research. Such specimens could one day be of interest to a pharmaceutical, food or cosmetic company or for a conservation effort.

“Obviously, we can’t conserve what we don’t know,” said Bierregaard. “One of the things we need to do is find out which species are vulnerable to the process of fragmentation.

“How many can survive in a patch of 100 hectares or 1,000 hectares? The only way we can do that is to count the number of species we find before fragmentation and after, and we can’t count the species unless we can identify them.”

Project researchers are trying to identify the forest’s insects and animals, as well as the plants. About 2 million dead insects collected from the fragments sit in a laboratory waiting to be studied. Small mammals have been trapped, with rodents being the most difficult to categorize by species. Some are distinguished only by their scent.

Many kinds of birds have been identified by their calls. Recordings of their voices are stored at Cornell University in its library of natural sound. Along with the tree specimens wrapped in newspaper and plastic, these recorded calls may one day be all that is left of their presence.

Despite the apparent rigor of their probing, the researchers admit that much of what lives in the rain forest will never be identified or preserved as samples. There is just so much.

“All you can do is rely on the manpower available and do the best you can,” Bierregaard said. “But we’re still going to be left sadly lacking. . . . We’re not going to know who is at the party here.”

Maura Dolan reported this story while on assignment in Brazil.