IN FUNCTIONAL TERMS A FLOWER IS NOTHING MORE or less than a plant’s way of initiating sex. Even though many plants are equipped with both male and female parts, and are therefore quite capable of having sex with themselves, most species fare better when sex occurs easily between two individuals. Genetic variety is what enables a species to adapt in a changing world; when two individuals are involved in reproduction—biologists call it cross-fertilization—the genetic permutations are much more various than when an organism reproduces itself.

But plants have a special problem where sex is concerned. A plant cannot roam the countryside looking for a suitable mate; it cannot sidle up to a likely prospect in a saloon. Plants are rooted; they know no moves. They must enlist the aid of go-betweens to shuttle the male sex cells of one individual to the female sex organ of another. Sometimes the go-between is wind or water, but usually it is an animal, in which case the plant must generally give something, a reward, in return for the assistance. To advertise the availability of the reward, most plants put out a show of color, and many add a whiff of perfume—in a word, a flower. And this is the flower’s role, its reason for being: it calls out to passing birds and insects, saying, Come here, come in, I’ve got something you want.

Primitive flowers like anemones and wild roses are not very subtle about the way they do this. The reward they offer is the very pollen that needs to be transported—sort of like paying the driver of an armored car by saying help yourself to some cash. Moreover, these primitive plants hold their pollen out where it is relatively accessible to whatever bee or beetle bumbles along. But as flowering plants have evolved they have become more sophisticated and less promiscuous. Most sophisticated of all, probably, are the orchids.

Consider for example the classic case of the European orchid Orchis mascula, whose reproductive habits were first described by Charles Darwin. (By the way, the genus name Orchis and the common name “orchid” come from the Greek word for “testicle”: the tuberous roots of orchid plants reminded the ancients of a human scrotum.) Orchis mascula produces a spiky cluster of very small blooms that are quite unlike the extravagant corsage flowers that most of us think of as orchids. However, it shares with those corsage orchids, and almost all others, three important anatomical characteristics: First, each flower has a single structure, called the column, that bears both the male and female sex organs, known respectively as the anther and the stigma. Second, the orchid’s thousands of pollen grains are contained in two structures called pollinia, each one a rounded packet of pollen attached to a base; the base is partly covered with a coat of sticky, viscous liquid. Third, one of the flower’s three petals is markedly different from the others, having developed over the eons into a broad “labellum” or “lip” that serves as a landing platform for visiting insects.

In the case of Orchis mascula the insect is a nectar-seeking bumblebee. (The flower doesn’t actually contain any nectar, but the bee doesn’t know that. Neither did Darwin.) As the bee positions itself in the flower to probe for nectar, its head butts against the sticky bases of the pollinia. One or both of the bases detach, whereupon the sticky stuff hardens and glues the pollen packet(s) to the bee’s head. To oversimplify it a bit, picture a bee costume with one of those silly antenna head pieces. The pollen packets stay with the bee until they come in contact with a female organ, a stigma, which has a tacky surface to which the pollen can adhere. When pollen touches stigma, the fertilization begins.

Darwin studied this process in minute detail, using a pencil point as a surrogate bumblebee. He championed the idea that self-fertilization is detrimental to plant species, so he was no doubt pleased to observe that even though Orchis mascula’s stigma lies mere millimeters behind the anther, where the bee picks up the pollen, its position is such that the bee cannot normally deposit a load of pollen onto the stigma of the flower from which it came. Rather, the bee departs with the pollen stuck to its head and, if all goes according to nature’s plan, takes it to another flower of the same species.

But wait a minute: if the pollen cannot strike the stigma of its own flower, how does it strike the stigma of another, identical flower? Darwin wrote, “This is effected by means of a beautiful contrivance.” He found that the base of each pollinium is “endowed with a remarkable power of contraction . . . which causes the pollinium to sweep through an angle of about ninety degrees, always in one direction . . . in the course of thirty seconds on an average.” In other words, about thirty seconds after being attached to the bee’s head, the silly antennae bend forward. Darwin could not prove it when he first wrote this, but he suspected (and a colleague later confirmed) that thirty seconds was time enough for a bee to finish its business on one flower and proceed to another. And, of course, after sweeping through an arc of ninety degrees, the pollen packets were in precisely the right position to strike the stigma of the new flower as the bee entered. So even though the flower cannot in the normal course of events mate with itself, it does mate readily with others of its kind. Darwin wrote:

A poet might imagine that whilst the pollinia were borne through the air from flower to flower, adhering to an insect’s body, they voluntarily and eagerly placed themselves in that exact position in which alone they could hope to gain their wish and perpetuate their race.

DARWIN DIDDLED MANY AN ORCHID flower—using pencils, bristles, whalebone, and his own fingers as insect stand-ins—and he took great guilty pleasure in “making out,” as he would put it, the diverse pollination mechanisms of the orchids that grew near his home, in the southeast of England, and of the exotic orchids that friends and colleagues sent him from far and wide. He confided to correspondents that his fascination with orchids was “idleness” and “more play than real work.” But in 1862, just three years after publishing On the Origin of Species, he gathered the casual research of several summers into a volume with one of those delightful Victorian titles, On the Various Contrivances by Which British and Foreign Orchids Are Fertilised by Insects, and on the Good Effects of Intercrossing. Most of what we know about the intimate life of orchids can be traced directly to what Darwin called his “very little book.”

In orchids Darwin had a natural laboratory in which he could test and bolster his controversial ideas. One of the species he discussed in the book was Angraecum sesquipedale, a native of Madagascar. This orchid offers nectar as its reward, holding it in a very narrow tube that Darwin described as a “green, whiplike nectary of astonishing length.” Darwin had a hard time making this one out. The bottom of the nectary was about eleven inches from the spot where an insect could be expected to stand, and Darwin, after much prodding with bristles and needles, became convinced that the insect could detach the pollen packets only as it stretched for the very last drops of nectar. Of course no one had ever heard of an insect with a foot-long tongue, but Darwin, believing firmly in the importance of insect-aided cross-fertilization, confidently surmised that in Madagascar must live a very large moth with a proboscis of ten or eleven inches. Entomologists were amused. The moth was found about forty years later.

Darwin’s orchid researches supported his theory of natural selection. He saw repeatedly that pollination by insects is a fine, dicey business: The insect has to stand in the flower just so, to receive the pollen packets on just the right part of its body. Sometimes the pollen must then move to a new position, but only after waiting a certain period of time. Then, at the next flower, the insect has to enter again just as it did before. This is a process in which millimeters count, and many of the “contrivances” Darwin discovered have the effect of encouraging the necessary precision. On the labellum of Orchis pyramidalis he found two “guiding ridges” of tissue that he compared to “the little instrument sometimes used for guiding a thread into the fine eye of a needle”; these ridges virtually force the proboscis of the pollinating insect, a moth, to touch the sticky base of the pollinia. In Serapias cordigera, an orchid from the south of France, Darwin noted that the stigma is narrow, and that the pollinia are directed into it by two “guiding plates.” In Plantarhera hobkeri, an American species, he found separate pollen structures so far apart from each other that any normal size moth would be able to reach the nectar without picking up any pollen at all—except that the labellum was structured in such a way as to divide the flower roughly in half, forcing the moth to sip the nectar from one side or the other, where its face was almost certain to make the right kind of contact. (Darwin noted, by the way, that moth-pollinated orchids seemed always to deposit their pollinia on either the insect’s proboscis or its eyes. Most of the rest of a moth is covered with scales that are easily detached—not an auspicious surface on which to paste the hope of future generations.)

Observations like these were important to Darwin, because some of his critics were inclined to see the fine details of flower structure—the little plates and ridges and curlicues and so on—as the flourishes of a Master Designer, ornamental touches placed deliberately by God. As long as such structures appeared to be useless, Darwin’s opponents had what seemed a persuasive argument against his theory of natural selection. “See here,” they could say, “if the structure of this orchid has been determined by impersonal natural laws, then where did this little ribbon of tissue come from? What impersonal natural law made this doodad over here curve outward and turn purple at the end?” In his orchid book Darwin not only proved that such apparently trivial details can indeed serve important functions, he also was able to show that some of them had evolved from structures that at one time had served completely different functions. Further, he showed that some structures were vestigial: useful at some point in the evolutionary past but not any longer. All this added up to an important argument against the concept of special creation by a divine decorator. The famous biologist Stephen Jay Gould put it this way:

The theory of natural selection would never have replaced the doctrine of divine creation if evident, admirable design pervaded all organisms . . . . Orchids are Rube Goldberg machines; a perfect engineer would certainly have come up with something better.

RUBE GOLDBERG WOULD HAVE LOVED the astounding orchids of the American tropics: Coryanthes speciosa, for instance, one of the “bucket orchids” of Central America. The lip of this orchid is a grotesquely modified petal whose tissue, which is quite tough, forms a sac, or bucket, roughly the size of a shot glass. Two small “faucet” glands hang over the bucket and fill it, a drop at a time, with a watery liquid.

As its reward this orchid offers a perfume that is highly prized by male bees of the tribe Euglossini—in English we call them Euglossine bees, or more commonly “orchid bees.” They collect the perfume by mopping it up with their hairy front feet; then they hover, transferring the aromatic compounds to storage organs on their hind legs. They do this with such maniacal focus that early observers thought the insects were being intoxicated. Evidently the bees use the perfume the same way we do, for courtship. Years before this was confirmed a botanist told me, “We’re almost sure it has something to do with their sex life, because very few animals would put that much effort into something that didn’t.”

The bees, in their turn, are an important part of the orchid’s sex life. Often several bees can be seen around the “bucket” of this flower, jockeying for position near the spots where the perfume is most abundant. The bucket’s rim is slippery. Once in a while a bee will lose his footing, or be bumped by another bee, or in some other way fall into the liquid. Now his wings are wet and he cannot fly. His only way out is to crawl through a narrow passage near the bottom of the bucket . . . and guess what happens there! The roof of the passage is formed by the flower’s column, the structure that bears its male and female organs. As he squirms through the passage—a struggle that’s been known to take as long as forty-five minutes—he detaches the pollen packets and gets them pasted to his back. Later, when escaping in identical fashion from another flower, he leaves the pollen on the flower’s stigma, and thus begins a new generation of bucket orchids.

NOW AS WE CONTINUE, please don’t be confused if you catch me referring to male flowers and female flowers. As I’ve said, most orchids are both in a structural sense, but each instance of cross-fertilization involves a supplier of pollen, a male, and a supplier of ovules, a female.

And the male has a lot riding on the back of his insect go-between. Because remember, his thousands of pollen grains are bound together in just two packets. If a mishap occurs—say the pollinating insect is eaten by a bird before reaching another plant of the same species, or leaves the pollen with a female of the wrong species—at least half of the male’s reproductive potential is squandered in one fell swoop. (All of it if both pollen packets are involved.) Common flowers have loose pollen that can be picked up a few grains at a time by a succession of visiting insects; these flowers have many chances to pass their genes into the next generation. But an orchid flower gets only one or two chances. Again, accuracy is of the essence.

To ensure the necessary accuracy, it behooves an orchid to know its pollinators well—to develop a relationship with a small group of insects, or even a single species, that will interact with the flower in regular ways and will be highly motivated to visit orchids of the same species again and again. One South American orchid attracts its pollinator by mimicking a female fly of the genus Paragymnomma; it even has a small, glistening spot that resembles the female fly’s genital orifice. Male flies of this genus pick up and deposit the orchid’s pollen as they attempt to copulate with the flower. Their incentive to visit other plants of the same species could hardly be stronger. At the same time, given the nature of the attraction, the odds are low that any other insect would pick up the pollen by mistake.

Many of the orchid family’s most marvelous pollination contraptions are found among the relatively small number of species pollinated by the abovementioned Euglossine bees. These species, all native to Central and South America, are arguably among the most highly evolved and specialized of orchids, and in terms of reproduction they are almost certainly the most thoroughly researched. The plants, the bees, and the relationships between them have been studied by several generations of American botanists—the colleagues and students and students’ students of Calway H. Dodson, who began the work at the University of Miami.

The rewards offered by these tropical orchids—the fragrance compounds that the bees collect so ardently—make possible an extraordinary degree of specificity in the relationship between orchid and bee. Each species of orchid produces its own distinctive fragrance, and each fragrance attracts only certain species of orchid bee; some of the fragrances seem to attract only one or two species. For a plant whose tiny seeds are dispersed by wind through the forest canopy, the usefulness of such an arrangement can hardly be overstated. Imagine that a bee picks up a load of pollen from an orchid that is 150 yards away from the nearest flower of the same species; imagine further that both plants are nestled in tiny niches near the tree tops, and that the area is home to a great variety of tempting flowers. What would happen if the bee, like a common honeybee foraging for nectar, were depending on visual clues to guide him to his next stop? Or if he were merely hungry, and inclined therefore to visit any old flower that would give up a drink? The orchid pollen might never reach its destination. But this is a bee with a mission: he is on the prowl for fragrance compounds, and only one particular scent will do. Sometime after he emerges from the male flower, with its pollen packets stuck firmly on his back, he picks up the odor of the female, wafting toward him on the jungle breeze. He makes a beeline, and the orchids are mated.

This focused relationship between bee, fragrance, and flower may encourage “speciation,” the branching off of new orchid species. That would make it a rare phenomenon indeed. The chemistry of fragrances is such that a small change in the composition of an aromatic compound can cause a relatively large change in aroma. In the orchids pollinated by Euglossine bees, a change in aroma can mean a change of pollinator. Researchers working at the University of Florida have proposed, therefore, that in these orchids a relatively small amount of genetic change could isolate a plant from its species mates over a very short span of evolutionary time—a couple of generations, perhaps, or conceivably even one. These researchers may have found in the forests of Central America at least a few examples of what they call “sympatric speciation,” one of the Holy Grails of evolutionary biology.

Mainstream biological thought holds that speciation usually occurs as a result of some sort of geographic separation: A seed is carried by a bird to an island, and there, changing a little at a time over the course of many generations, the plant evolves into something completely different from its progenitors. Or a mountain grows up from the bottom of a lake and separates a population of guppies into two isolated communities, each of which goes its own evolutionary way. What separates two species is known as a “reproductive barrier,” and physical isolation has generally been considered the only one strong enough to get a new species started. But the Florida researchers have wondered if the job couldn’t also be done by a sudden change in a flower’s perfume. One of those researchers, the recently departed Mark Whitten, wrote his doctoral dissertation on Gongora quinquenervis, an orchid that occurs from central Mexico to northern Brazil. Studying it and closely related orchids in central Panama, he found what appears to be evidence that it is diverging into two distinct, genetically isolated species that are “sympatric,” meaning they share the same habitat.

The flower of this orchid measures about two inches from top to bottom; it has bent-back sepals that suggest wings, and a bulbous lip that reminds me of an animal’s head. It looks like a giant, obscene, red-and-white flying insect. In Panama, Mark Whitten found growing side by side with it a flower that is nearly identical in size and shape and only slightly different in color. The two plants differ in structure and appearance much less than a brother differs from his sister, or a beagle from a dachshund—certainly not enough to be called separate species. But Whitten analyzed the fragrances produced by the two flowers and found that they are markedly different. Actually, he didn’t have to analyze them; his nose knew. “One has sort of a honey smell,” he told me, “real sweet, and the other smells just like Pine-Sol toilet-bowl cleaner.” In the field Whitten found that each flower is visited by a different group of Euglossine bees. At first he thought the flowers were identical in structure, but the results of his fragrance analysis prompted him to look closer. When he did, he found minute differences in the shape of the lip.

The evidence is far from conclusive — and Whitten emphasized that it could be explained in a variety of ways—but it at least suggests a scenario of sympatric speciation: A small mutation occurs in a single G. quinquenervis plant, the sort of mutation that happens every day and generally amounts to little. But this one, instead of adding a spot of red on the underside of the lip, or subtracting a couple millimeters from the length of the flower stalk, causes a slight change in the chemical process by which some of the mutant’s offspring produce their fragrance compounds; as a result, some plants in the next generation grow up smelling like Pine-Sol instead of honey. The bees that normally pollinate this species have no interest in these new, strange-smelling flowers. They do not visit. But, as luck would have it, another species of orchid bee lives in the area, one that disdains the smell of honey but cherishes that of Pine Sol. So now the “normal” G. quinquenervis plants live side by side with the strange ones, yet because the two groups are pollinated by different bees they are genetically isolated from each other—they are two separate communities of orchids, each free to accumulate its own genetic quirks and thus to wander down its own evolutionary road. Within a few years the new group changes slightly in color. Decades later it develops an extra lump of tissue at the base of its column. In a century, perhaps, one flower will be white and the other red. In a few million years they may be as different as a rose and a cactus. The Florida researchers were excited by the possibility that this could be happening right under their noses. Norris Williams, the head of Whitten’s lab, wondered, “Have we caught it right at the stage where speciation is occurring, where the fragrances are changing, the pollinators are changing? Have we caught it right at the stage to study the actual events of speciation going on?” If so, their view was a privileged one. They were peering into the mechanism of evolution, watching the gears of a clock that normally ticks but once in a thousand years.

SOME ORCHIDS HAVE CONTRIVED to get themselves pollinated by Ernest Finney. When I met him, more years ago than I am going to admit, Mr. Finney was a compact and energetic man in his early sixties, with gray hair and a fondness for blue jumpsuits. He was the orchid breeder at Orchids by Hausermann, one of the largest commercial growers in the United States, located in the suburbs of Chicago. Early in the morning he would buzz through the company’s greenhouses watching for an orchid from which he could snatch, or on which he could deposit, a packet of pollen. A giant blue bumblebee.

In this environment an orchid needs more than perfume to attract its pollinator. Mr. Finney favored orchids that were very large and very colorful. He preferred flowers that would last for weeks instead of days, and plants that would “like to grow,” as he put it, in the homes of his customers. (Because “Some plants,” he confided tersely, “would rather die than live.”) He was also attracted to orchids that bloomed at certain convenient times, so he’d have flowers to sell at the greenhouse all year round and a big supply of corsage orchids for Easter and Mother’s Day. If survival of the fittest is the law of the jungle, survival of the sellable is the law of a commercial greenhouse.

For many years, Finney told me, orchids were marketable primarily as corsage flowers, so most commercial breeding was aimed at producing large, frilly blooms, usually white or lavender in color. What many people think of as an orchid is the triumphant result of this breeding process, an oversized, hybridized Cattleya that’s four or five generations removed from its origins in the jungle. But now orchids are also used as houseplants and fancied by hobbyists, and commercial growers are more interested than they used to be in the colorful and the exotic. Often the goal of Finney’s breeding efforts was to combine the size of the large white corsage flowers with the color or sometimes the shape of smaller, “wilder” plants.

The process was a slow one. Depositing the pollen of one orchid onto the stigma of another was simple enough—Finney used a pencil—but sometimes eight or nine years would pass before he could see the result; that’s how long some crosses need to produce a mature flower. And usually several generations of crossing and back-crossing were required to produce a flower that Finney wanted to photograph and offer for sale in the catalog. And when all that work was done—sometimes fifteen or twenty years after it began—Finney had but a single plant: a prototype. Then it had to be mass-produced.

In some cases Finney could (and in some cases he had to) induce the plant to manufacture itself: he simply placed its pollen on its stigma and raised the seeds that resulted. But many orchid hybrids are unable to produce seeds. And even when sterility is not a problem, consistency often is: Compared with plants that are commonly raised from hybrid seed—corn, for example—the orchid family is very “heterozygous,” as the botanists say. Roughly meaning: genetically diverse. The jumbling of genes inherent in sexual reproduction therefore produces in orchids a comparatively high amount of variation, and this, combined with the long flowering times, makes it very difficult for breeders to bring a hybrid to the point where it will “breed true”—that is, produce a high percentage of offspring that are identical to the parents. As Finney said, when a large part of your business is serving orchid enthusiasts by mail-order, you’d better be sure the flower that blooms in the customer’s home looks like the one pictured in the picture. The best way to be sure is to circumvent the complications of sex and reproduce the orchid by “tissue culture.” In other words, clone it.

In a small room that was sometimes grandly referred to as Hausermann’s “propagation lab,” Finney’s assistants worked with sterilized scalpels and a large magnifying lamp, cutting the vegetative buds, or eyes, of orchid plants down to “meristems,” the tiny, magical bits of tissue where cells divide and growth occurs. They have been used to mass-produce orchids since the 1960s.

The dividing cells of a meristem are undifferentiated; that is, they have not yet become leaf cells or root cells, they are merely orchid cells, not yet committed to a function. These cells can be induced to continue dividing without differentiating if the meristems are rotated in a flask containing liquid nutrients. The flask is mounted on a motorized wheel so the contents are constantly tumbling. This technique works because the dividing cells of a growing plant differentiate according to clues provided by gravity; to put it crudely, cells at the bottom become roots, and cells at the top become shoots. But cells dividing in a rotating flask don’t know what to become— they can’t tell up from down. So they just keep dividing.

After about three months growing in the flask, one bit of tissue was cut into two. After another month, the two were divided into four or six. When the bits of tissue numbered in the dozens, or hundreds, they were placed on a layer of solid growth medium in a glass bottle. There, oriented to gravity at last, each little bit of tissue became an orchid plant. And each of these dozens or hundreds of plants was identical to the single plant from which the original meristem was cut.

When Ernest Finney took me on my first tour of Hausermann’s greenhouses, he wanted to be sure I understood the import of all this. He led me to a vantage point between two different rooms of orchids. In one direction was a roomful of breeding stock—a garden of potted Phalaenopsis orchids, whites and yellows and greens and pinks of many different shapes and sizes. “Now come on up here,” Finney told me, “and turn around and look back.” I laughed out loud. Now I was looking into a room of several hundred Phalaenopsis clones, delicate white flowers with red lips and red “candy striping” in the petals and sepals. The plants and flowers were identical in height, shape, and color. Of course they all stood the same, facing the light coming through a greenhouse wall of glass. It was a battalion of orchids, a mass of conformity. “This,” Finney said proudly, “is what tissue culture does for you.” It separates the business of propagation from the messiness of mating and the complications of sex. Mr. Finney used sex only to make different genetic combinations. Once he hit on the right combination, his assistants could do the rest.

IN 1982 THE AMERICAN ORCHID SOCIETY challenged botanists to do away with orchid sex altogether. The society offered a cash prize of $50,000 to the first scientist or group to produce a so-called “somatic” orchid hybrid— one that combined the genes of two orchids by fusing their somatic cells (that is, normal body cells as opposed to sex cells) in a test tube. The idea was to apply to orchids the techniques of bioengineering, which then had been used far more successfully on animals than plants (because of basic differences in animal and plant cell structure). The development of such a technique would no doubt have implications for all of plant biology, but for orchid breeders the most intriguing possibility was that it might enable them to create flowers that nature alone could not produce: an electric blue Phalaenopsis, for example, or a Gongora measuring six inches across.

Alas, the prize was never claimed. And in general the genetic manipulation of flowering plants has not progressed as quickly as the society must have dreamed in 1982. A few years after the prize was offered I met an orchid man from Portugal, Luis Pedrosa, who hoped to earn his PhD working on the problem. He found, however, that orchids were too slow for degree work, and so turned his attention to cereal plants. They grow faster, are easier to handle, and have been more thoroughly researched than orchids have. Besides, a breakthrough in corn or wheat would be far more important to the human race than a blue Phalaenopsis.

And now genetic engineers have even more important and lucrative work to do—manipulating RNA to make vaccines, for example. But what if scientists one day succeed in breeding orchids without sex? What would the orchid lover Charles Darwin say to that? Would he see it as an evolutionary event? Would he say that our largest and most diverse family of flowering plants was branching off into two distinct and incompatible groups—one depending primarily on insects for its propagation, the other depending entirely on humans? If so, I wonder, which would he bet on as the more promising strategy for survival? The humans, or the insects?