Where does the wind go after it blows?

To La Ventana, "The Window," where the echoes of distant storms roar across the continental divide of Monteverde with racing fog and horizontal rain. Cool, wet, and furious, this funnel of wind produces a bizarre "elfin forest" of tempest-shaken dwarf trees, hidden pools, and swirling mist where an astonishing animal was found and then lost.

We're here to see the scene of its probable extinction.

The One World Journeys team is slogging the undulating, knife-like spine of Costa Rica between the Pacific and Atlantic, slithering down an overgrown trail of greasy mud and clambering up the next hump by clinging to roots and shafts of bamboo. Leaves as broad as sombreros crouch in the lee of the ridge. Invasive beds of impatiens create mini-gardens where there is a patch of sky. Mud sucks our boots until water tops them. To windward, gusts estimated at 50 kilometers per hour rattle us with rain.

Trees are rocked so hard that their root pads rise and fall like the chest of a breathing animal, creating dark pockets underneath where amphibians might hide.

We're wet, tired, and exhilarated to be here. Our guide is J. Alan Pounds, 48, an ecological scientist who has spent the last 20 years documenting extinctions and migrations in Monteverde. Each year at mating season he searches for the golden toad, a brilliant-orange-colored amphibian just two inches long.

And each year Pounds is disappointed.

First discovered in 1963, the last golden toad was seen in 1989 after a severe El Nino drying episode. It's one of 20 species of frogs and toads that collapsed in Monteverde after the winter of 1987-88.

"The golden toad is just the poster child for a larger phenomena," the scientist said. "It's sad to live in a place where there are all these interesting and beautiful creatures and then see them vanish."

The harlequin frog and Holdridge's toad are gone from Monteverde. Populations of red-eyed stream frogs and terrestrial-breeding rain frogs have crashed repeatedly during recent El Ninos. The cloud forest anole lizard, similar to a chameleon, has disappeared from Pounds' study site and retreated to the wet area around a nearby river. Meanwhile, lowland species, at the rate of about 19 per decade, are moving upward; these include reptiles, amphibians, and birds such as golden-crowned warblers, lesser greenlets, and Keel-billed toucans.

Does any of this prove the threat of global climate change caused by human pollution? Pounds is too careful a scientist to say that. He notes a strong correlation between species collapses and strong El Nino episodes that dry the cloud forest. Thirty years of climate data from neighbor John Campbell, now deceased, indicate that temperatures in Monteverde have risen and rainfall patterns have shifted. But fellow ecologist Bob Lawton has suggested that maybe lowland deforestation is responsible, not greenhouse gases. Just to complicate things, reforestation of old pastures has increased the forest cover around Monteverde by 19 percent.

"The more we learn, the more questions there are to answer," Pounds says, soaked but barely winded by our clambering. Misery doesn't daunt him: the wetter he gets, the drier his wit. He cheerfully leads as far as photographer Gary Braasch asks to go. Panting, I'm glad when Gary finally says enough is enough.

Extinction is not simple. Higher temperatures or shifts in precipitation may only indirectly affect plants and animals. For example, a drying spell in 1983 forced frogs to bunch near the mist of waterfalls where they were attacked by parasitic flies. As a result, scientists found 40 dead or dying frogs near waterfall pools.

One idea is that a fungus has exploded worldwide, supporting disease organisms that are decimating amphibians. Can that fungus be blamed on global warming?

Even proving the golden toad is gone is difficult. Pounds says he is 95 percent sure but quotes the aphorism, "The absence of evidence is not evidence of absence." While the toads were seen in the same places for 17 years, indicating they don't migrate, only time and consistent observation can prove their extinction. In 1987 there were 1,500 golden toads counted and in 1988 and 1989 just one. Since then, zero.

While scientists sift evidence, Costa Ricans would love to have the tourist-drawing toad back. Some enterprising "Ticos" tried to sell scientists their land, complete with resident golden toads, but the animals turned out to be a common variety covered with orange paint.

Certainly the Monteverde community of researchers didn't set out to strike a global alarm. Pounds had to switch his doctoral thesis from harlequin frogs to lizards in the 1980s because the frogs didn't breed when expected to, but no one realized this might be early evidence of global change. Yet throughout the 1980s, oddity after oddity occurred. "By 1990, we knew something was awry."

What's at stake goes far beyond a little toad. "Amphibians may be telling us something about our own future," Pounds explains. "Declines in reptiles and amphibians as well as different distributions of birds may be an indication of environmental change that eventually may have much larger consequences for communities of animals and human kind."

Also at issue is whether existing remedies will work. "The obvious cause of extinction in the tropics is deforestation," he says. "The strategy is to protect some of the forest in parks and preserves. Unfortunately, climate change threatens those. I object to the idea that if you (set aside a preserve) you can do whatever you want to the rest of the land."

Monteverde's importance to science is something of a coincidence. Local biological diversity and beauty happened to draw researchers like Pounds. They in turn have produced a baseline of species population data over the last three decades, something rare in the tropics. Now this data is invaluable.

There are no fat research grants here and no glittering labs. Pounds, who chose the freedom of going it alone in Costa Rica instead of the security of an American university position, drives a 1973 Land Rover, doesn't have the money to finish his modest home lab, and recently took Costa Rican citizenship. Yet he has assembled some of the strongest evidence that something is wrong with our climate.

Another of the dozen or so key scientists associated with Monteverde is Richard LaVal, 63, the "bat man" of Monteverde. We visit him at night, of course, descending a dark path to his self-built, unpainted home in a forest hollow, its gate marked by a sign warning "perro bravo," or brave dog.

On LaVal's wooden porch are flowerpots, an old rusting water heater, a plank on cinder blocks that serves as lab table, and a rugged laptop computer purchased on e-Bay for $250 to which Richard has fitted a polished wooden handle. Next to it is a small sonar detector, ordered from Australia. As bats skitter across his garden, their sonar cries of echolocation create a scrawl on his computer screen. In just half an hour some 60 bat "signatures" have been logged, and all-night sampling over the past year has accumulated a data bank of 100,000 bat visits.

It's an impressive record for a gentle and friendly PhD scientist whose ponytail and beard make him look like an aging hippie. He has to support his own research with teaching, guiding, and wildlife photography, but colleagues respect him enough to have named a Peruvian bat after him.

He greets us with a curious orange apron tucked into his waist and surprises us in mid-interview by reaching inside. "I have a bat right here."

It's a nectar bat with long snout and hummingbird-like tongue, soft as velvet and clean as a cat, caught in gossamer "mist nets" in his banana garden. Another one later painfully nibbles the hand of Toby Malina, our sound technician (who endures it stoically). With a release of LaVal's fingers, this undeniably cute female mammal flies away.

LaVal has been studying bats for more than 40 years and has lived in Monteverde for 22. While the abundance of bats in Monteverde has not changed since he began counting in 1973, at least 22 new bat species from the lowlands have moved into the cloud forest, correlating with measurements that show mean minimum temperatures in Monteverde have risen about 2 degrees Centigrade.

"Over and over, species come by that have never appeared in Monteverde before," he explains. His 100,000 recordings will be analyzed for clues about the speed and pattern of bat migration. Meanwhile, he wishes he had funding to set up more bat monitoring stations.

It's Monteverde's rich biological diversity that makes these rapid changes so disturbing. What a treasure trove is in peril! At the community's butterfly garden, naturalists Jim Wolfe and Stephanie Smith show us cockroaches the size of small clams, butterflies whose open wings look like owls, metallic-colored cocoons that mimic raindrops, and jeweled scarab bugs that camouflage themselves by reflecting their surroundings like mirrors.

On another night we walk the dripping forest trails of the Monteverde Preserve with guide Ricky Guindon, using flashlights to see creatures that come out only at night.

We find "walking stick" insects that look like twigs, a katydid that looks like a leaf, a Blunt-headed tree snake swallowing a lizard, and earthen banks dotted with the burrows of tarantulas, which locals consider fairly harmless. (Many "Ticos" tolerate scorpions in their homes to keep down the cockroaches.)

Frogs the size of fingernails pose on huge leaves, and crickets look as pink as boiled lobsters. This dark, wet, windy, and voraciously competitive world surprises me by seeming more comforting than threatening. It's the product of millions of years of evolution and a storehouse of genetic ingenuity, marvelous in the cruel and beautiful logic of predation and defense and more intricate than any computer.

The only truly dangerous creature out here, I feel, is me.