They motored noisily over black Amazonia lake waters where orchids swelled from dead tree trunks and yellow-winged marsh birds rose flapping from the floating grass.

A giant antenna stood in the middle of their little boat, which made them look suspicious. At the prow was a compass mounted on a wooden tripod. It was tricky getting at the compass because a sloth was sleeping on it.

Roberto H. de Salvo Souza, the young researcher from Rio de Janeiro, nosed the boat toward a pale dead tree and killed the motor. There was a quick rush of silence, and then the full high chorus of the birds: chattering kingfisher, light soprano swallows, raspy-throated limpkin. A snail kite, dark and gliding cleanly like a gull, soared overhead with a snail between its claws.

Fernando Weber Rosas, the young researcher from the southern Brazilian city of Porto Alegre, stood in the center of the boat and put headphones over his ears. He turned the antenna some and stopped, and then turned it a little more.

"Sixty-five," Rosas said.

Robin Best, the zoologist from British Columbia, gently disengaged the sloth from his compass and bent over to look more closely. "Which animal is it?" he asked.

"Thirty," Rosas said.

Best made a note in a small yellow book. Thirty would be a female, 140 pounds or so, particularly fond of that part of the lake. She would be moving underwater with her whiskers quivering and her wide flippers propelling and her thick gray body meandering along with that ponderous slow grace. Animal number thirty was an Amazonian manatee -- a "sea cow," in the folklore, and one of the gentlest and oddest-looking creatures in the whole benumbing array of Amazonian wildlife.

For the last five years, Robin Best has studied Amazonian manatees, which are part of the now-endangered family of big slow-moving water mammals whose sighing suckling of their young is said to have made ancient lonesome sailors imagine mermaids. He is a tall, broad-shouldered, nordic-looking man, and he does his work in warm Amazonian river waters, in a cluttered Manaus laboratory, in field houses where the only beds are cotton hammocks swinging from wall hooks. He has dived with thick fishing nets to catch manatees live, lowered manatees into laboratory tanks to measure the content of the air they exhale, wrestled radio transmitter belts around their powerful flailing tails, bottle-fed and soothed baby manatees until they fell asleep sucking his index finger.

The work he is doing, like much of what goes on at the Manaus-based National Institute for Amazon Research (INPA), is nearly unique. His colleagues are nutritionists who take blood samples from jungle children who have never seen a stranger before, wood specialists who cut slices of elaborately-grained hardwood unheard of outside the Amazon, agronomists who travel the jungle by canoe in search of new edible fruits that only local people know. There are botanists, ecologists, fish experts, tropical disease experts -- all of them set loose, just a few years ahead of the cattlemen's bulldozers and the colonists' machetes, in perhaps the richest unexplored extravagance of plant and animal life in the world.

Ever since the 19th-century English naturalists Alfred Wallace and Henry Bates booked passage to Belem and began selling mounted insects to London for threepence apiece, the Amazon basin has stunned biologists with the sheer numbers of species that thrive in its rivers and forests. During Bates' 11 years in Amazonia, he collected almost 15,000 species of mammals, birds, reptiles, and insects, more than half of them entirely new to science. Wallace managed during his four years of Amazon work to develop the radical notion that species evolved through natural selection rather than divine creation -- and his paper on the subject, written some years after he left the Amazon, was presented at an 1858 London conference alongside that of the highly controversial scientist Charles Darwin.

Bates finally returned to London more than a hundred years ago, but Amazon scientists are still stumbling across so many new species of plants, birds, and insects that the total numbers can only be guessed at. Two and a half million species of insect, 2,000 of fish, and 3,500 species of usable woods -- "Those are the known," said Henrique Bergamin, the bushy-eyebrowed Brazilian chemist who directs INPA. "In a few hectares of Amazon forest you have a larger number of species than the total flora and fauna of Europe."

With no harsh season to kill off weaker species, and one of the world's most diverse feeding grounds -- thousands of different kinds of trees and plants scattered all over the forest, with none of the natural single-species forest that grows in temperate zones -- the animals of Amazonia have specialized themselves and adapted to the particular demands of the jungle.

There are fish, their mouths especially designed to crush rubber seeds, that swim rivers at their annual flooding to eat the seeds dropped from rubber trees. There are ground beetles with claws ideally shaped for grabbing tree leaves to escape from flooding and torrential rains. There are howler monkeys, whose piercing territorial shriek resonates through the forest for three miles, and giant red anteaters with long prying snouts, and swamp deer with membrane on their hooves to help keep them from sinking into the mud.

Jorge Arias, the Panamanian INPA medical insect expert who specializes in the study of an ulcer-producing disease carried by sand flies, once on impulse sent an Australian colleague 2,000 specimens of booklice, a small insect that got caught easily in his sand fly traps. They were throwaways for the purposes of Arias' experiments, but the Australian scientist was astonished. Nearly half the 84 species he identified were totally new -- nobody had even known they existed. In one common insect family he found 31 species, and only four of them had ever been identified.

Scientists like Robin Best arrive at this wealth of unexplored life with a wonder that never entirely leaves them. "It's a bit overwhelming," Best said as he leaned in his swim trunks against a pillar of the small pontoon-based floating house where his researchers spend every other night.

"Right now the Amazon represents the last frontier in biological research. Africa and Asia have been well-studied in terms of biology of elephants and wild beasts and lions and so on -- whereas in Amazonia virtually nothing's been done. Some studies have been done on primates, but really it's only scratching the surface . . . . The giant otter -- there's no studies done in the Amazon. The dolphins -- there's no study. Manatees -- in five years we've basically done the only work that's been done. Things like the jaguar, the small ocelots and marguay cats -- deer, wild pigs -- nobody has studied these animals."

What that means, as Brazil's debt-ridden development forces push steadily into this frontier, is that scientists are now working at a furious pace simply to help Brazil grasp exactly what it stands to lose. What kinds of usable, potentially high-profit woods are being ground under bulldozers or drowned in hydroelectric projects like the enormous Tucurui dam? What permanent damage and erosion is forced on jungle land that is stripped all at once of its tree cover? How many species are lost every time a swath of tropical forest is burned for a cattle ranch?

In a country where environmental impact statements and pleading conservationists' proposals are sometimes dismissed as luxuries of the developed world -- whose bankers, it must be noted, hold the papers on the staggering Brazilian debt -- the work of INPA scientists has become, as one agronomist put it, "a race to what we like to call 'rational development.' "They can't leave it for its beauty and say, 'Hey, look, guys, we have the biggest and prettiest jungle in the world,'" Arias said. "It isn't realistic . . . . They're having problems, and they're trying to solve the problems before they become catastrophic."

What Best and his colleagues are hunting, then, is an economic incentive to keep the endangered Amazon manatee from disappearing to hunters who still kill them for their meat and skin, despite a 14-year-old Brazilian law against manatee hunting. Nearly every day, on a five-hour course that takes them skimming by motorboat from the head of Curua Una lake to their little floating house surrounded by half-submerged trees and unflappable black piranha fish, Rosas and de Salvo Souza track 30 radio-belted manatees that Best deliberately released there two years ago. Each manatee's transmitter emits a different signal, which lets the researchers establish a daily compass point for every manatee they can find.

Curua Una is a man-made lake, the product of Amazonia's second hydroelectric project, and like many such lakes has been troubled by the floating grasses that grow on much of the water's surface. The grass breaks up in the turbine gratings, makes the lake evaporate far too much water, and ferments as it rots, releasing a sulphide gas that pulls oxygen out of the water and suffocates all the fish. Best's work is showing so far that Amazonian manatees eat those grasses -- and they are entirely peaceable animals, apparently ideally suited to the waters of a tropical man-made lake.

To spend a working day last month with the manatee team was to glimpse in miniature the extraordinarily complicated world a scientist sees in Amazonia. The floating island that looked at first like solid grasses suddenly turned cosmopolitan with the animal life Best described -- the freshwater shrimp under leaves, the small snails crawling, the brilliant dragonflies hunting out insects, the spider weaving its grass-to-grass web, the mud-dauber waiting to paralyze the spider and carry if off as food for its young.

And when the truck carrying the manatee team to the lake suddenly screeched to a stop, all three men leapt out and ran back to a small brown heap that the truck had managed not to squash. It was a female two-toed sloth, a foot long, down on one of her rare descents from the treetops, and belly-crawling slow as a slug across the pavement. Best had been fascinated by sloths ever since he watched one crawl across his patch shortly after his arrival at INPA, and he gently lifted this one and began running his fingers through her fur.

She looked at him with the permanent dreamy smile sloths seem to wear, and dozens of tiny moths began to scramble from her fur. "A whole ecosystem living on one animal," Best said. "The fur is especially adapted to house certain types of algae . . . . The algae actually use the hair surface, probably, to take up nitrogens, which is their main food source -- and probably the benefit for the sloth, although the sloth doesn't know he's been benefited, is he gets a sort of greenish coloration which probably helps the sloth in camouflaging against predators.

"In addition, we've found what we call a guild of insects which feed on the feces of the sloth . . . . We've found two new species of beetles, two new species of moths, three species of mites and two species of ticks . . . . Some really spectacular ticks."

The sloth nestled contentedly against Best's chest while he carried her to the field house. The three men laid her on a wooden desk, picked off every moth and mite they could find, and dropped them in a small alcohol jar. Rosas scratched her neck softly and the sloth leaned back, arms spread, looking very happy. Best smiled. "You only get this in the best geisha restaurants of Japan," he said.

In the morning they took her to the lake and set her on a tall pale tree that looked like an aspen. Slowly, her long claws reaching for the tree bark overhead, she climbed up and disappeared into the treetops, and the manatee team headed back town the lake as a sudden rain came smashing down on the boat and their heads and the black lake waters all around them.