Decoding a Flower's Message
Floral fragrance does more than beckon pollinators
Texas gourd vines unfurl their large, flared blossoms in the dim hours before sunrise. Until they close at noon, their yellow petals and mild, squashy aroma attract bees that gather nectar and shuttle pollen from flower to flower. But “when you advertise [to pollinators], you advertise in an open communication network,” says chemical ecologist Ian Baldwin of the Max Planck Institute for Chemical Ecology in Germany. “You attract not just the good guys, but you also attract the bad guys.” For a Texas gourd plant (Cucurbita pepo variety texana), striped cucumber beetles (Acalymma vittatum) are among the very bad guys. They chew up pollen and petals, defecate in the flowers and transmit the dreaded bacterial wilt disease, an infection that can reduce an entire plant to a heap of collapsed tissue in mere days.
The gourd vine’s problem—how to attract enough pollinators but not too many beetles—is a specific case of a floral dilemma that biologists first noticed many decades ago. In 1879, Austrian botanist Anton Kerner von Marilaun published a treatise titled The Protective Means of Flowers against Unbidden Guests, in which he described glands and sticky hairs that seemed to help keep harmful insects at bay. But a century later, scientists had still barely begun to consider the contribution of a flower’s scent to those interactions; to do so would require a convergence of fieldwork, chemistry and molecular biology. “We’re just beginning to train the type of biologists who can use those tools,” Baldwin says. The resulting experiments have begun to reveal the many ways that floral fragrances may manipulate animal behavior.
In one recent study, published in the February issue of Ecology, Nina Theis and Lynn Adler took on the specific problem of the Texas gourd. Its main pollinators are honey bees (Apis mellifera) and specialized squash bees (Peponapis pruinosa), which respond to its floral scent. The aroma includes 10 compounds, but the most abundant—and the only one that lures squash bees into traps—is 1,4-dimethoxybenzene.
Intuition suggests that more of that aroma should be even more appealing to bees. “We have this assumption that a really fragrant flower is going to attract a lot of pollinators,” says Theis, a chemical ecologist at Elms College in Chicopee, Massachusetts. But, she adds, that idea hasn’t really been tested—and extra scent could well call in more beetles, too. To find out, she and Adler planted 168 Texas gourd vines in an Iowa field and, throughout the August flowering season, made half the plants more fragrant by tucking dimethoxybenzene-treated swabs deep inside their flowers. Each treated flower emitted about 45 times more fragrance than a normal one; the other half of the plants got swabs without fragrance.
The researchers also wanted to know whether extra beetles would impose a double cost by both damaging flowers and deterring bees, which might not bother to visit (and pollinate) a flower laden with other insects and their feces. So every half hour throughout the experiments, the team plucked all the beetles off of half the fragrance-enhanced flowers and half the control flowers, allowing bees to respond to the blossoms with and without interference by beetles.
Finally, they pollinated by hand half of the female flowers in each of the four combinations of fragrance and beetles. Hand-pollinated flowers should develop into fruits with the maximum number of seeds, providing a benchmark to see whether the fragrance-related activities of bees and beetles resulted in reduced pollination.
“It was very labor intensive,” says Theis. “We would be out there at four in the morning, three in the morning, to try and set up before these flowers open.” As soon as they did, the team spent the next several hours walking from flower to flower, observing each for two-minute intervals “and writing down everything we saw.”
What they saw was double the normal number of beetles on fragrance-enhanced blossoms. Pollinators, to their surprise, did not prefer the highly scented flowers. Squash bees were indifferent, and honey bees visited enhanced flowers less often than normal ones. Theis thinks the bees were repelled not by the fragrance itself, but by the abundance of beetles: The data showed that the more beetles on a flower, the less likely a honey bee was to visit it.
That added up to less reproduction for fragrance-enhanced flowers. Gourds that developed from those blossoms weighed 9 percent less and had, on average, 20 fewer seeds than those from normal flowers. Hand pollination didn’t rescue the seed set, indicating that beetles damaged flowers directly—regardless of whether they also repelled pollinators. (Hand pollination did rescue fruit weight, a hard-to-interpret result that suggests that lost bee visits did somehow harm fruit development.)
The new results provide a reason that Texas gourd plants never evolved to produce a stronger scent: “If you really ramp up the odor, you don’t get more pollinators, but you can really get ripped apart by your enemies,” says Rob Raguso, a chemical ecologist at Cornell University who was not involved in the Texas gourd study. “It’s kind of like asking ‘What if antlers were longer or larger or heavier? Is there some threshold above which they’re actually a burden?’” For Texas gourd, there is—and Raguso, with ecologist Candace Galen and their coauthors, reached a similar conclusion in February 2011, in an American Naturalist paper on alpine skypilot (Polemonium viscosum). Too much fragrance could harm those flowers, too, but for a completely different reason.
Skypilot grows in the alpine meadows of western North America, where its purple flowers release a “complicated” odor “like grape juice floating on beer,” Raguso says. The main component of their chemical bouquet is 2-phenylethanol, and it alone is absorbed into the skypilots’ nectar. The flowers vary widely in the amount of 2-phenylethanol they produce, and, like Theis and Adler, the researchers expected that pollinators might prefer the highest levels, perhaps as an indicator of abundant nectar.
Instead, they found that blossoms with the most 2-phenylethanol didn’t necessarily have the most nectar. And when they supplemented some flowers with sugar solution plus 2-phenylethanol, boosting them to the highest naturally occurring levels of the fragrance, bumble bees were unimpressed. Freely foraging bees (Bombus balteatus) chose control flowers over highly scented ones in 18 out of 21 trials. “We were absolutely shocked,” Galen says. “We really did that expecting to see that the bumble bees would visit the 2-phenylethanol-enriched flowers.”
Part of the explanation emerged when the team investigated the effects of the odor on ants (Formica neorufibarbis)—the skypilots’ unbidden guests. The insects gather skypilot nectar, but not only do they not pollinate the flowers, they often sterilize them by biting off the style, the stalk that supports the pollen-receiving organ. The style seems to get in the ants’ way, says Galen, an ecologist at the University of Missouri in Columbia. So they “do some housekeeping and shove it right out of the flower.” In contrast to the case of the Texas gourd, however, ants avoided 2-phenylethanol just as much as bees did. Indeed, exposure to the highest natural concentrations of the compound could be lethal to ants in as little as one hour. It seemed that the plants had entered a trade-off in which—for some flowers—it was worth sacrificing pollinator visits to keep ants out.
But not all flowers have quite that much 2-phenylethanol. So in a final round of experiments, the team tested blossoms treated with an intermediate concentration. Bees weren’t as put off by those blossoms, but they still spent less time on them and guzzled less nectar. In that way, 2-phenylethanol might encourage bees to keep moving and transferring pollen between flowers—while saving the plant its investment in nectar.
That makes for “an interesting comparison” to the Texas gourd story, Galen says. Whereas the gourd flower scent advertises, the skypilot scent is all about defense: It repels destructive ants and, although pollinators tolerate it to some extent, it keeps them from lingering.
Both studies are examples of the kinds of “manipulative approaches” that are finally uncovering the functions of floral scents and the evolutionary pressures that shaped them, Baldwin says. But, he adds, there’s much more to do. By focusing on a single scent compound in each species, as these studies did, “you’re probably missing out on a lot of detail,” he says. In work that’s completed but not yet published, his team used genetic modifications to knock out individual components of floral aroma in petunia “to take apart a very complicated floral bouquet.” The results, he says, show that within the scent of a single flower, some individual compounds attract pollinators while others serve as defense.
As such results accumulate, they will not only solve an evolutionary puzzle; they will lead to “an enormous number of possible applications,” Baldwin says. Understanding a flower’s message to both pollinators and enemies “is absolutely essential,” he says, “so you can design and engineer crops that aren’t going to cause problems for themselves.”