Smokey the gardener: fire begets flowers, but heat often has nothing to do with it.

By the end of the 1980s, the human race could grow golf course greens in the middle of a desert, breed bell peppers to look like chocolate, and raise pumpkins that weigh hundreds of pounds. But all that amassed knowledge wasn't telling graduate student Hannes de Lange, then at the University of Stellenbosch in South Africa, how to grow the plant he wanted to study for his dissertation--or even how to get the seeds to sprout. His subject, false heath (Audouinia capitata), resides in the botanical wonderland called the fynbos on the west side of the southernmost tip of Africa. More than two-thirds of the plant species there, including seven entire families, grow nowhere else. Putting all those seeds in the one habitat basket proves to be a risky strategy as human development encroaches, and false heath ranks as a vulnerable species.

Protecting false heath appeared difficult because, 2 decades ago, only Mother Nature could reliably grow it. The seeds don't sprout under normal greenhouse conditions, and the only recorded findings of seedlings came from the fynbos after a wildfire.

In a variety of ecosystems around the globe, fires trigger a burst of new plant growth. The difficult question is, What is it about fire that does the trick?

The answer has a bearing on a wide range of endeavors, including farming and conservation. Research had been smoldering for decades on various aspects of the problem, but de Lange ignited the field.

Earlier work had established that for some species, such as certain beans, it's the heat that counts. However, in the late 1980s, de Lange reported that heat alone was unlikely to activate false heath seeds.

Another possible factor in fire-induced germination emerged from experiments by California researchers. They had boosted germination in local fire-adapted seeds by exposing them to charred wood or even just to water in which charred wood had soaked. But when de Lange applied a tea made from plant ashes to another fynbos species, the false heath seeds' germination didn't improve.

Then, he had an idea that no scientific experiment had tested. Where there's fire, his theory went, there's smoke.

He burned bits of plants in a barrel and blew the smoke through pipes, where it had a chance to cool, to nine tents, each of which covered a square of cleared ground. Later, de Lange removed the tents.

Visiting the site 7 months later, he observed dozens of false heath seedlings poking up in the squares of soil that the tents had covered. No false heath had sprouted in the surrounding ground.

This finding touched off a race among labs around the world to find the active component or components in smoke. These tests revealed dozens of kinds of seeds that smoke can kick into action.

But just what that special activator or activators might be has proved maddeningly hard to pinpoint. Seven years ago, a team working with North American seeds proposed an answer, but the other contenders dispute it. This summer, an Australian group put forth a completely different solution to the mystery.

THE BIG SLEEP At the heart of questions about smoke-influenced germination is seed dormancy, the suite of adaptations that keeps the brainless nuggets of tissue from sprouting at inappropriate times.

To fit the myriad habitats on the planet, seed chemistry varies widely. At one extreme are plants that essentially give live birth. Mangroves, for example, bloom and produce embryonic plants that start to expand their roots and shoots before separating from their mothers. Offspring are already growing seedlings by the time they drop off the ancestral twig.

Seeds in a fax greater number of species, however, stop their embryonic development and just wait. Seed biologist Marc Alan Cohn of Louisiana State University's AgCenter in Baton Rouge calls such dormancy "roughly equivalent to a state of self-imposed suspended animation."

He's not talking about the seeds that sit quietly in a seed packet until they're tucked into the ground. The domesticators of a crop often manage to avoid genuine seed dormancy. Before planting, such seeds are merely what Cohn calls quiescent. As soon as they encounter the necessities, such as water, they're off and growing.

In contrast, a gardener can put a truly dormant seed in appropriate soil and give it a hearty watering, and it will remain just a seed. What the seed is waiting for depends on the species.

In one broad class, dormancy ends when a waterproof coating finally gives way, says Cohn. Water floods in, and the embryo starts pushing out roots and leaves. Many wild members of the bean family experience such a shrink-wrapped start to life. Reports on some of these suggest that fire can jolt seeds into action by rupturing their waterproof coats.

The coat-cracking scenario is easy to visualize, maybe too easy. As Cohn puts it, "People think it's like Betty Crocker physiology--you add water, you stir, you get a plant." He emphasizes that seeds can rely on many other cues.

A huge number of other seeds, he notes, have no waterproof coats yet linger unresponsively, regardless of nice, damp surroundings. In these seeds, water seeps in during wet periods and out during droughts but doesn't trigger growth.

Many of these well-watered embryos await a temperature trigger, explains agricultural seed specialist Kent Bradford of the University of California, Davis. Some seeds need a long stretch of heat followed by cold, or the reverse, or even heat and cold rapidly alternating at just the right amplitude. Since temperatures fluctuate more dramatically near the soil's surface than deep underground, that last requirement probably protects a seed from sprouting when it's too deep in the soil, Bradford explains.

And then there are the chemical cues. Finding and explaining them may keep researchers at work for decades.

SMOKE SIGNALS The search for active chemicals in smoke has kept several labs busy. Kingsley Dixon of Kings Park and Botanic Garden in West Perth, Australia, describes the 11 years that he's worked on the problem as an "odyssey with lots of dead ends and U-turns"

He and his colleagues had shown by 1995 that smoke in fire-prone areas in Australia, as in the fynbos, boosts the germination rates for some native species. Identifying the dormancy-breaking compounds, however, remained a tough challenge. Smoke can easily contain 4,000 individual substances.

The evolution of a chemical germination signal in smoke, Dixon reasoned, would probably require a widespread compound. He decided to work with smoke generated by burning cellulose, which would be present in any fire that consumes plants.

The researchers proceeded by capturing a particular portion of the chemicals in cellulose smoke, testing whether it affected the germination of seeds such as the common ornamental plant called kangaroo paw. If it did, researchers subdivided the active portion and repeated the germination test to zero in on active ingredients.

About 2 years into the last, intensive phase of the research, the experiments began to go badly, recalls Dixon. The kangaroo paw seeds germinated more and more readily, regardless of exposure to smoke samples. Finally, Dixon and his colleagues realized that seeds, as they linger in storage, become ready to sprout even without a cue. They seem to have internal clocks. "It's a great discovery, but it cost us a year's work," Dixon says.

Meanwhile, Johannes van Staden and his colleagues in South Africa had been testing compounds on their local flora, and Ian Baldwin, now at the Max Planck Institute for Chemical Ecology in Jena, Germany, focused on Nicotiana attenuata, one of the wild tobacco relatives native to the U.S. Great Basin Desert. Baldwin's team, like Dixon's, approached the problem by testing sets of the compounds in smoke and then testing subsets of the active samples.

"Some years ago, it seemed to be a very simple question," says biochemist Bernd Krock, who works with Baldwin in Jena. By now, he calculates, the group has tested at least 170 components of smoke for activity. So tar, no luck.

The first claim to an active ingredient in smoke came from a group led by Jon Keeley, now with the U.S. Geological Survey and based in Three Rivers, Calif. When Keeley and C.J. Fotheringham tested natives of California's shrubby chaparral areas, they found species with seeds that responded to fire. In 1997, they published a paper contending that nitrogen dioxide from smoke breaks the dormancy of seeds of the California chaparral wild-flower known as whispering bells.

When the researchers exposed seeds in a test: chamber to nitrogen dioxide for as little as 1 minute, the seeds sprouted at high rates that matched those of seeds exposed to actual smoke from sprigs of the common chaparral shrub called chamise. Without smoke or nitrogen dioxide exposure, the seeds didn't sprout at all.

At the time, discussion arose over the prevalence of nitrogen dioxide as an air pollutant. Keeley warned that dirty air might highjack germination cues, prompting seeds to sprout at times when seedlings couldn't succeed.

A few other laboratories have run tests exploring nitrogen oxides as germination cues. Cohn's lab tested smoke-induced germination in red rice, which springs up uninvited in commercial rice paddies in the southern United States. Cohn and his student Lucia Doherty had documented a considerable boost in germination rates after treatment with a smoke extract, actually Liquid Smoke from the grocery store. When Cohn and Doherty analyzed the chemistry of Liquid Smoke, though, they found no sign of nitrogen oxides or its breakdown products.

"I'm not saying Jon Keeley is wrong," Cohn comments. "There's no doubt that nitrogen oxides [in some forms] will break dormancy." What he and other researchers do dispute, Cohn says, is whether these oxides are the main compounds that give natural smoke its power.

Fotheringham, who's now at the University of California, Los Angeles, responds, "We never said nitrogen dioxide was the thing. In fact, we said quite the opposite: that it worked with some species but not with others."

Another group including Baldwin took a different approach in work described in the March 2004 Seed Science Research. This team, including Catherine Preston now of the Agricultural Research Service in Gainesville, Fla., observed germination of whispering bells and wild tobacco in response to smoke components, but not to nitrogen oxides dissolved in water. The paper concludes that "unidentified cellulose combustion factors, rather the nitrogen oxides, are likely to be the ecologically relevant germination signals."

Fotheringham objects that the study by Baldwin's team is "not sufficient to cast doubt on our findings." For one thing, Fotheringham says, the team didn't test nitrogen dioxide, and the concentrations of the other nitrogen oxides in that study were an order of magnitude below the other concentrations she and Keeley reported as effective.

In the meantime, Baldwin's team has found a different twist in the puzzle of smoke germination. In patches of vegetation that are near a fire but don't burn, wild tobacco doesn't sprout, even when smoke drifts by. That makes ecological sense, because seeds wouldn't have the optimal growing conditions that arise after a fire has burned away the vegetation, thereby releasing nutrients and reducing competition. Seeds sometimes have to wait 150 years for the right scorching, Krock notes.

So, how does a smoke-sensitive seed tell whether its neighborhood has burned or smoke was just drifting by? The wild-tobacco seed is tipped off by the absence or presence of substances that leach out of fallen, unburned plants, Krock and his colleagues reported in 2002. Some terpenes and abscisic acid that seep into the soil can override the "go" message from smoke. However, when a fire sweeps through, this litter burns up and the seeds respond freely to the smoke.

GOT IT? Big news came this summer, when Dixon, Gavin Flematti, and their colleagues reported in the Aug. 13 Science that they have isolated and synthesized a smoke component that dramatically enhances germination in several species, including one that doesn't come from fire-adapted ecosystems. Deserting the fickle kangaroo paw, Dixon's team switched to two other Australian wild-flower species, a cottontail and the queen trigger plant, and an old commercial variety of lettuce called Grand Rapids. As Flematti of the University of Western Australia in Crawley closed in on smaller and smaller subsets of smoke components, his tests eliminated all the plentiful components of cellulose smoke. Working with the minor components was difficult because it required burning large amounts of cellulose to get enough of a chemical to test.

The substance that finally worked, one of the compounds called butenolides, had been just a "tiny little bump" in the analysis of complete smoke, says Dixon. The butenolide acts at concentrations below 1 part per billion, equivalent to one-third of a teaspoon dissolved in a swimming pool. Flematti and the other chemists on the team spent more than 4 years isolating the compound, figuring out its chemical structure, and making a synthetic version.

Dixon's group has a provisional patent on the substance, which he says is stirring interest among agricultural and conservation specialists.

Although Dixon expected plant scientists to take an interest in his claim, he was startled to hear from a beer maker. He learned that the barleys for premium beers have been selected for flavor rather than germination rate and that some of the seeds are temperamental.

"I was very excited to read the [Dixon] paper in Science," says Baldwin, who has spent 14 years chasing smoke cues. "In 1996, we isolated and synthesized a structure very similar to the published structure but lacking a methyl group, but it was inactive."

He's not giving up the search. "We have known for years that there are at least four active compounds in wood smoke," he says.

Keeley agrees there's probably no single compound in smoke that triggers seeds of all plants. He also questions whether the butenolide is ecologically relevant.

Cohn is reserving judgment for the moment on the number of active compounds, but he welcomes the find. He says, "It's new chemistry [for seed germination] to my knowledge, and that alone is satisfying."