Published online 18 May 2007 | Nature | doi:10.1038/news070514-20


Drifters could explain sweet-potato travel

An unsteered ship may have delivered crop to Polynesia.

Where did these come from?Where did these come from?AddStyle

How did the South American sweet potato wind up in Polynesia? New research suggests that the crop could have simply floated there on a ship.

The origin of the sweet potato in the South Pacific has long been a mystery. The food crop undisputedly has its roots in the Andes. It was once thought to have been spread by Spanish and Portuguese sailors in the sixteenth century, but archaeological evidence indicates that Polynesians were cultivating the orange-fleshed tuber much earlier than that, by at least AD 1000. However, there's no hard evidence of people travelling between South America and the South Pacific so early in history. Most Polynesian crops have their origins in Asia, where the people are thought to have migrated from.

Some scientists have suggested that sweet-potato seeds were carried in the stomachs of itinerant birds. Others suggested that Polynesians once voyaged to the Ecuadorian coast, and — foregoing such South American staples such as maize and Phaseolus beans — they brought back a root that reminded them of Asian yams. Still others proposed that the sweet potato floated 8,000 kilometres across open ocean, either in its spherical seed pod or in a drifting vessel.

An unlikely journey

This last theory was dismissed in the 1970s, when a group of researchers ran what was then a state-of-the-art ocean-circulation simulation and concluded that the vessels would be extremely unlikely to drift into Polynesia by accident1. But Álvaro Montenegro, an oceanographer at the University of Victoria, British Columbia, says that this old model has "outlived its usefulness". Researchers now know, for instance, that the equatorial ocean currents travel much faster than previously thought; projections of drift have changed.

Montenegro and his team used data from the Estimating the Circulation and Climate of the Ocean (ECCO) experiment, which combines a circulation model with 12 years of oceanographic measurements. This approach gives a glimpse of circulation patterns with a spatial resolution of about 50 square kilometres, and allowed Montenegro to consider 160 launch sites along the South and Central American coast compared with the earlier team's six.

In the simulation, sweet-potato seed pods that started off in these American waters could reasonably hit seven different island groups, and had the best chance of landing on the Marquesas. "Among the three most likely targets that get hit, two are within the area where people believe the crop was introduced," says Montenegro. But the trip took at least four months. Even coconuts can't survive in salt water that long.

More likely, says the team, a loaded vessel was blown out to sea and landed on the islands — which could take as little as 90 days, they report in an upcoming issue of the Journal of Archaeological Science2. Montenegro notes that seed pods are transported by currents alone, but a drifting vessel gains momentum from the wind.

Rooted in fact?

With 2,000 years between sweet-potato domestication in South America and its arrival in Polynesia, there have probably been numerous such drift events. But did people survive the trip, along with the crops?


Montenegro says linguistic similarities between the Quechua word for the tuber, cumal, and the Polynesian one, kumala, suggest that humans must have been along for the ride. Patrick Kirch, an archaeologist at the University of California, Berkeley, who has worked on the issue, thinks that more extensive and deliberate contact must have taken place. "In my view, the most probable mechanism of transport was Polynesians sailing to South America," he says.

The simulations, Montenegro says, simply demonstrate the feasibility of the drift theory, but a final verdict will depend on future archaeological findings. This approach may also help us understand how other plants, such as the bottle gourd, have drifted to the distant islands. And vessel drift is not just a thing of the past either. Montenegro points out that in 2005-2006, a fishing boat from Mexico was swept to the Marshall Islands — just as his model predicts.

Visit our couldexplain_sweetpo.html">newsblog to read and post comments about this story.  

  • References

    1. Levison M., Ward R., Webb J. &Levison M.The Settlement of Polynesia: A Computer Simulation, Univ. Minnesota Press, Minneapolis (1973).
    2. Montenegro Á., Avis C. &Weaver A. J. Archaeol. Sci., Preprint at