Making scents of flowers: it's time for science to close its eyes and sniff.

Take a deep, sweet lungful of jasmine fragrance--and admire its courage. Its swaggering machismo, even. For a strongly scented flower takes terrible risks, says scent biologist Robert Raguso of the University of South Carolina. The same perfume that draws invited guests to the flower can also tip off pests, thieves, and killers. "It's like a peacock's tail," he says. Raguso has had to turn to the science of visual signals for a classic example of inconvenient exuberance. There's barely been a science of biological displays of scent, he laments. In textbooks or seminal papers, "odor is either ignored or treated as insolubly complex," says Raguso.

"This is a human bias," he argues. We're such visual animals that, of course, the rainbow of floral colors and flowers' myriad shapes fixed scientists' attention first. Also, the techniques for studying scents haven't allowed much precision--until recently.

Now, Raguso says, advances in collecting and identifying volatile compounds are opening a new world. At last, evolutionary biologists can examine the same issues--such as attraction powers, evolutionary history, and exposure to predators--for olfactory displays as for visual ones. Harnessing modern genetics to the new scent chemistry is revealing the genes and enzymes behind the scenes. There's even talk about boosting scents in crops and commercial flowers for bigger yields, more sniff appeal, and perhaps swaggering courage, too.

SCENT MATTERS Nature uses plenty of perfume. By the early 1990s, scent biologist Jette Knudsen of Goteborg University in Sweden had tallied some 700 volatile compounds wafting from 441 kinds of plants. While nobody has kept an exact count, newly discovered substances join the list every year.

The complexity of the chemistry depends on the species. Snapdragons and petunias release relatively simple blends of 7 to 10 compounds. Some orchids, however, secrete scents with around 100 ingredients.

Scent alone can do the vital job of attracting customers to a flower. In 1876, Charles Darwin described wrapping flowers in muslin so insects couldn't see the blooms. Pollinators showed up, anyway, so Darwin concluded that odor was the lure.

In a 1986 refinement on the protocol, Olle Pellmyr, now of the University of Idaho, led an experiment covering the yellow, tightly packed flowers of the western U.S. skunk cabbage, Lysichiton americanum, with glass so insects could see them but not catch the scent. On other skunk cabbages, the researchers blocked the view but not the smell. Beetles ignored the glass-covered plants but swarmed toward the scent.

An even more vivid demonstration focused on the snakeroot, Cimicifuga simplex, which grows in three forms, each with its own habitats, pollinators, and scents. In 1990, Pellmyr and his colleagues reported that fritillary butterflies seek the scent of only one form, probably because of its irresistible methyl anthranilate and isoeugenol. The researchers spiffed up flowers of the other forms by adding dashes of those compounds, and the butterflies started visiting them, too.

That's not to say that looks don't matter. Some pollinators, if forced to choose, will respond to visual cues rather than scented ones. And in the real world, scents and sights interact. Pellmyr's group found that the number of beetles alighting on a skunk cabbage flower more than doubled when the researchers permitted insects to see the flower as well as smell it.

Raguso says that intense whiffs of the perfume of the night-blooming datura he studies drive hawkmoths to start probing at patches of white. Lucky moths quickly find the white trumpets of the flowers, but "often a moth will poke at my socks," says Raguso. If he walks far enough away from the strongest odors, though, the moths lose interest in his ankles.

There's a dark side to this power of alluring scent. Some Australian orchids release scents that trick male wasps into ridiculous positions and, according to a new study, threaten the social lives of female wasps. The orchid Chiloglottis trapeziformis belongs to a group of about 300 species that lure pollinators by mimicking a female insect. Males grappling flower parts in attempts to mate pick up pollen and then rub it onto another orchid during the next delusional encounter.

The C. trapeziformis flowers don't look much like female wasps beyond their dark green-to-brown color range. However, scent seems to be the most important advertisement, say Bob Wong of Australian National University in Canberra and Florian Schiestl of ETH Zurich in Switzerland. When they hid a real female of the wasp Neozeleboria cryptoides and an orchid in identical opaque chambers in the woods, allowing only the scents to waft out, males showed no sign of being able to distinguish the odors.

The males do learn from bad experiences, though. The researchers set out orchid flowers at 2-minute intervals to mimic a colony blooming in the woods. The first blooms attracted visits and mating attempts from male wasps, but within 20 minutes, the males largely ignored them.

That's hard on the females, Wong and Schiestl report in an upcoming paper in Proceedings of the Royal Society of London B. Because they never grow wings, the females require that the males fly them over to food, such as a honeydew-rich cluster of scale insects, and later provide transport to a good place to lay wasp eggs. Yet when researchers placed females among orchid blooms, males' visits to these potential mates were far less frequent than to females not among flowers. When the male wasps wise up to the floral deception, it's hard luck for the female wasps.

SWEET SYNDROMES Wielding analytic tools capable of identifying scent components, today's scientists are revisiting questions raised about pollination at least a century ago.

One long-standing issue has been how precisely a flower's lures target particular pollinators.

Knudsen is testing the idea that flowers might create a specific bat lure. She has collected scent compounds from bat-pollinated tropical flowers in South America and finds that many of them contain sulfur. That's not common among floral scents overall, she says.

The charms of sulfur also struck Otto von Helversen of Friedrich-Alexander-Universitat Erlangen-Nurnberg in Germany and his colleagues. They reported in 2000 that in lab tests, small flower-visiting bats of the genus Glossophaga preferred dimethyl disulfide to several nonsulfurous substances. Tests outdoors in Costa Rica's lowland forest produced similar results, with bats showing significant interest in artificial flowers scented with two specific sulfur-containing compounds.

Whether bats outside the Americas respond the same way remains to be seen, Knudsen cautions. She and her student Stefan Pettersson have found aromatic sulfur-compounds in only some of the plants visited by African bats.

Other researchers have been testing fragrances of flowers pollinated by hawkmoths. In the October 2001 Phytochemistry, Rachel Levin of the Smithsonian Institution in Washington, D.C., and her colleagues reported a survey of 20 species of flowers called four o'clocks, many of which bloom at night and attract hawkmoths. The majority of the hawkmoth-pollinated flowers emitted spicy aromatic alcohols and esters, as scientists had predicted, but so did a few relatives that attract other pollinators.

For really specific targeting, pick a fig, says Martine Hossaert-McKey of Centre d'Ecologie Fonctionnelle et Evolutive in Montpellier, France. Each of the 750 known fig species has its own species of tiny wasp for pollination. To a human nose, "some fig flowers smell very sweet, some are like green tea--there are many smells," says Hossaert-McKey. In the February Journal of Chemical Ecology, she, Laure Grison-Pige, and their colleagues reported testing the fragrances of four species of tropical fig trees.

The researchers found that the fig species they tested release scents that share several main ingredients. Each species has a distinctive aroma with different proportions of the shared ingredients, and some have a dash or two of unique compounds.

In lab tests, three of the wasps responded strongly to whiffs of their own fig species but not significantly to the others. Yet there was a problem with the wasps that the scientists had thought would pollinate the fourth fig. Those insects turned up their antennae at all four fig aromas. So the researchers began to rethink the taxonomy. They suspect that the trees they had considered varieties of the fourth species were actually members of two other species. The wasps seem to be more discerning than human taxonomists.

There's more to attractive scent than a pretty flower though. Raguso, Levin, and Lucinda McDade of the Academy of Natural Sciences in Philadelphia, remind floral biologists to stop and sniff the foliage. Raguso has found that hawkmoths searching for alluring odors respond to leaf fragrances as well as to those of flowers. Levin's recent survey of four o'clocks revealed that four species in the genera Acleisanthes and Selinocarpus emit up to 80 percent of their total volatile emissions not from the flowers but from the leaves. The leafy allure even includes hefty proportions of sesquiterpenoids, which are among the ingredients that have been proposed as hawkmoth attractants.

Evolutionary biologists are also waking up and smelling the flowers. Because plants use scents to flirt with their pollinators, the odors strongly influence the complex dynamic that determines species.

Kim Steiner of the California Academy of Sciences in San Francisco has been investigating South African orchids and relatives of snapdragons that attract bees with the unusual lure of insect baby food. The bees don't eat the oil themselves but instead mix it into a lump of provisions for their larvae.

In March, Steiner told a conference in Ventura, Calif., that he had overlaid a family tree of these unusual species with information about their scents. For example, he found two Pterygodium orchid species growing in the same area in southwestern Africa and another species, with a different pollinator bee, far to the east. Yet all three species had a distinctive scent that Steiner calls "soapy and pungent." Chemist Roman Kaiser of the fragrance company Givaudan in Dubendorf, Switzerland, has identified eight compounds in their scents. The most abundant, with a benzene ring, has never before been isolated from a living thing.

On the family tree, the three orchid species are one another's closest relatives, so the scent may have persisted as the species diverged. In other eases, plant lineages appear to have switched their scent chemistry, taking advantage of different pollinating species.

DOLLARS AND SCENTS Shifts in pollinators may have economic, even culinary, consequences. In the early 1990s, Allen Young of the Milwaukee Public Museum and David Severson of the University of Notre Dame compared floral fragrances from nine varieties of the cacao plant that aren't closely related. In an old cacao plantation in Costa Rica, the researchers set out traps scented with the oils from the different varieties. Then, they watched to see which insects responded.

One fragrance outperformed all the rest, luring lots of midges and stingless bees. This winning concoction came from the variety called Rim-100, selected from Mexican plants that still look much like cacao's wild ancestors. The researchers speculated that over the years, cacao breeders had maximized some desirable properties of the crop but lost much of the scent. Such feeble fragrance, the researchers suggest, might explain why many plantation varieties attract few insect visitors.

Scent scientists could do farming a great service if they could put some aromas back into crops, says Natalia Dudareva of Purdue University in West Lafayette, Ind. Three-quarters of crops depend on insect pollinators, and the number of their visits influences the size of the fruit. Dudareva muses that better-smelling plants could score more pollination and produce better yields. A watermelon flower needs about a dozen visits from pollen-carrying insects to develop a real whopper of a fruit, and a strawberry needs some 25 visits to reach prime berry size.

While being bred for other characteristics, many ornamental flowers have lost their scent. After all, some modern roses smell about as exciting as iceberg lettuce. Dudareva says that people would prefer to buy fragrant flowers.

Geneticists are beginning to tease out the basic mechanisms that give flowers their scents. Eran Pichersky of the University of Michigan in Ann Arbor and his colleagues reported the first floral scent gene in 1996. It came from a fragrant, Western wildflower, Clarkia breweri. The gene encodes an enzyme critical to making linalool, an alcohol that powers many strong flower odors, including lavender.

Since then, researchers have been finding other scent genes in Clarkia as well as in snapdragons, roses, and some other species. Pichersky estimates that the worldwide score for finding scent genes hangs somewhere around 10.

Finding the genes has been the easy part. Exploiting them to enhance floral aroma is much harder, says Dudareva.

Pichersky managed to get one of the Clarkia aroma genes to work in a tomato plant to produce small amounts of the scented compound linalool in the fruit--resulting in a faint, sweet, floral smell. Enhanced fragrance might increase the appeal of store-bought tomatoes, the researchers hoped.

A team in the Netherlands put the same Clarkia gene into a petunia. Unfortunately, other compounds in the flower bound to the linalool and made it too heavy to waft away.

Even more challenges await those who seek to transplant fragrances, says Dudareva. For example, snapdragons churn out most of their odor between 9 a.m. and 4 p.m., the standard workday for bees. Even plants growing in darkness maintain this rhythm. In 2000, Dudareva and her colleagues reported that the enzyme for a compound that gives the snapdragon fragrance its snap functions around the clock. However, the benzoic acid that the enzyme works on waxes and wanes, creating the daily rhythm. Putting the enzyme gene into a plant may not mean a thing if it doesn't have enough raw material.

The secrets of fragrance may be difficult to uncover, but it's not an impossible task, says Pichersky. It's a great time to sniff the flowers.