Sweet, rich and rare, Tahitian vanilla is the gourmet's choice and second only to saffron in the league of pricey spices. For botanists, Tahitian vanilla holds a fascination far beyond its flavour: the orchid that produces it is an evolutionary enigma. Vanilla tahitensis is known only from French Polynesia and New Guinea, where it has been cultivated since the late 19th century. The puzzle is where it came from originally. All vanilla orchids with aromatic beans are native to tropical America, yet this species has never been found there. Diligent detective work and DNA analysis suggest that the plant emerged from the chocolate gardens of Guatemala, crossed the Pacific on a Spanish galleon, and found its way to Tahiti in a French admiral's baggage.

All good detective stories need a decent hunch, some intriguing clues and, these days, some sophisticated DNA analysis. If the plot revolves around an old, unsolved case, so much the better. The Tahitian vanilla mystery has all this and more: exotic locations and a trail that spans half the globe and at least half a millennium.
© Jean-Pierre Pleuchot/Image Bank/GettyTahiti is the home of the finest vanilla, but Tahitian vanilla has long been an enigma

For more than 50 years, botanists have puzzled over the origin of the vanilla orchid grown on the islands of French Polynesia in the Pacific Ocean. The first formal description of the species was based on a specimen from Raiatea, the second largest island in the archipelago and sacred to the Tahitians, so it was named Vanilla tahitensis. Yet the orchid was clearly alien to these islands.

There are some 110 species of Vanilla orchid distributed across the tropics, but those with aromatic pods are native to tropical America. Most of the world's vanilla comes from the cured beans of just one of these species, Vanilla planifolia, a scrambling vine-like plant that grows wild in the forests of Central America. Although the plant is rare in nature, the Mayan people were well acquainted with it. For thousands of years, they cultivated forest gardens to provide food for their towns and cities. Among their crops were cacao beans - the source of chocolate - and vanilla, which they used to flavour their chocolate drinks.

After the Spanish conquest, the colonists lost little time shipping both back home, and for the next three centuries Europe's vanilla came exclusively from southern Mexico. Mexico lost its monopoly in the early 19th century, when the French established plantations in Madagascar and Réunion in the Indian Ocean. Today, most vanilla comes from these islands.

The origins of V. tahitensis have proven more elusive. The only plants known are on plantations or are escapees run wild. Not one specimen has been found growing naturally anywhere in the world. This prompted some botanists in the 1950s to suggest that the orchid might be a mongrel, a species created through hybridisation. But after half a century, no one had found any evidence to support the idea, nor put forward any better suggestions.

Intrigued, plant scientist Pesach Lubinsky at the University of California, Riverside, decided to dig deeper. He soon found himself following two separate trails: one a paper trail of historical clues, the other a genetic trail that he hoped would lead to the orchid's closest relatives American Journal of Botany, vol 95, p 1040).

Historical sources say that in 1848, Alphonse Hamelin, commander of the French Pacific fleet, sailed to Tahiti in the flagship Virginie. During his visit, he gave the governor a present - some cuttings of a vanilla orchid, which he thought would make a fine addition to the governor's garden. The cuttings were said to have come from the Philippines, then a Spanish colony. Later, sailors from other French ships brought more cuttings to sell, and by the 1880s plantations had sprung up on several of the islands in the archipelago.

The trail might easily have gone cold there, for today there is no sign of V. tahitensis in the Philippines. There is, however, a tantalising reference to a "vanilla of Guatemala" in the Flora de Filipinas published in 1837. This book was the work of Manuel Blanco, an Augustinian friar who spent years surveying the islands of the Philippines and is regarded as a reliable botanist. Blanco's fleeting reference corroborates French reports that their cuttings came from the Philippines. More interestingly, it suggests that the orchid's journey may have begun somewhere no one expected: Mexico's southern neighbour, Guatemala.

According to historical records, all exports of vanilla from Spanish America came from Veracruz on Mexico's Gulf coast. But not all Spanish ships travelled the Atlantic route to Europe. For three centuries, two or three Spanish galleons a year made the long journey between Acapulco on Mexico's Pacific coast and Manila, the capital of the Philippines. A favourite target of pirates, these ships carried silver plundered from the Inca mines across the Pacific, returning to Mexico with ivory, silks and spices from Asia. They carried plants too. Many of the native American species now established in south-east Asia and Oceania arrived aboard a Manila galleon.

As Lubinsky points out, if there had been any trade in vanilla with the Philippines, or any attempt to establish the plant there, then it would make sense to acquire beans and cuttings from plants grown along the Pacific seaboard of Spanish America, in what is now Guatemala, El Salvador and the Mexican state of Chiapas. The alternative was to transport vanilla from Veracruz to the Pacific coast, a difficult journey that meant crossing the mountains of the Sierra Madre or the malarial swamps and thick forest of the Isthmus of Tehuantepec. Although there is no record of vanilla exports from the west coast, there is documentary evidence that the Mayan Indians had been cultivating the plant there long before the arrival of the Spanish. In the 1630s, English explorer Thomas Gage wrote a detailed description of how the Mayan people in the Suchitepequez region of Guatemala cultivated "orchards" of cacao trees in forest gardens and how they grew an assortment of spice plants to provide flavourings for their chocolate drinks. Vanilla was a favourite.

The historical trail pointed to Guatemala. The DNA trail also led to Central America. If the hybrid theory was right, then the plant's putative parents ought to grow somewhere in these rainforests. Lubinsky and his colleagues analysed the DNA of 40 Vanilla species distributed from southern Mexico to Peru. The two most closely related to Tahitian vanilla turned out to be the commercial species V. planifolia and a less familiar species, Vanilla odorata, which is known only in the wild. Both grow in Guatemala. The big question was: were these just close relatives, or was there a more intimate connection?

Lubinsky analysed the DNA data again to look for patterns of relatedness characteristic of a hybrid origin. This time, he scrutinised DNA from both the cell nucleus and the chloroplast, which is inherited entirely from the "mother" plant. The results confirmed that Tahitian vanilla had arisen from the hybridisation between V. planifolia and V. odorata, and that V. planifolia was its "mother". The data also indicated that there had been little divergence between the new species and its parents, meaning that it must be a relatively recent addition to the flora, perhaps only 500 or 600 years old.

To Lubinsky, the plant's recent emergence and its absence from the wild suggests that it owes its existence to some ancient gardener, most likely one of the Maya who centuries ago cleared patches of forest to plant their cacao trees and spice plants.

"In the forest V. planifolia is rare, with an estimated abundance of one individual per square kilometre or less," says Lubinsky. This makes the chances of hybridisation with another rare wild species very slim. The Maya collected plants from the forest and planted them close together in their gardens. "Species that had previously been geographically separated were then able to hybridise because they were in the same place. That's how we think Tahitian vanilla got started."

There is a chance, then, that somewhere deep in the forest, in a long-abandoned chocolate grove, a strange-looking vanilla plant awaits discovery. In January, Lubinsky is going hunting for hybrids, starting in Belize where Dawn Dean of the Maya Mountain Research Farm has identified patches in the forest where several vanilla species grow together in association with cacao trees. "We think these were probably once managed cacao and vanilla forests," says Lubinsky. It is exactly the sort of place you might expect hybrids to evolve - and one of them might be the elusive Tahitian vanilla.