One of the enduring mysteries of world history is whether the Americas had any contact with the Old World before Columbus, apart from the brief Viking settlement in Newfoundland. Many aspects of higher civilisation in the New World, from the invention of pottery to the building of pyramids, have been ascribed to European, Asian or African voyagers, but none has stood up to scrutiny.

The one convincing piece of evidence for pre-Hispanic contact has been the humble sweet potato, which is of tropical American origin but widely cultivated across the Pacific islands. Until a few years ago it was assumed that this was the result of Spanish transmission, dating to the early colonial period, but archaeological discoveries in the Cook Islands show this to be wrong: excavations at Mangaia yielded carbonised remains of sweet potato (Ipomoea batatas) dating to AD1000, five centuries before Europeans entered the Pacific Ocean.

The question then arose as to whether the diffusion of this useful crop was the result of Amerindians sailing west to Polynesia, as the late Thor Heyerdahl always claimed, or whether it came about because Polynesians exploring on “the road of the winds” beyond Easter Island came to the South American mainland, and took back with them the hardy and nutritious root crop which is today fifth in importance in developing countries.

The lack of evidence for Native American seafaring and the reputation of the Polynesians as navigators inclined most scholars to the latter thesis: but a new simulation study suggests that either the Amerindians or nature may have been responsible: Alvaro Montenegro and his colleagues in the Journal of Archaeological Science argue that computer experiments demonstrate that accidental drift voyages could have been responsible.

The experiment was set up to investigate two transfer theories, by accidental voyages from the American mainland that reached Polynesia, and drifting of Ipomoea seed capsules. Deliberate voyaging was not included.

Starting positions were in a series of “departure bins” defined off the Central and South American coastlines from 50 degrees south to 30 degrees north — roughly from southern Chile to northern Mexico.

The drifter point was located at the centre of each bin, and thus some distance offshore. The various Pacific island groups were designated as targets, and the probable drift of vessels the size of a large canoe under the influence of the known winds and currents simulated over a six-month period; the drift of seed capsules was simulated for a full year.

The most probable canoe crossing to score a “hit” was from Central America to the Marshall Islands, with a likelihood of 11.5 per cent. The much shorter crossing from Ecuador to the Galápagos was second, at almost 10 per cent, followed by the central Polynesian island groups of Tuamotu and the Marquesas at 7.4 and 5.7 per cent repectively. Most other targets scored very low, although Hawaii had an almost 3 per cent chance of being encountered.

The drifting seed capsules had a 17.4 per cent chance of reaching the Galápagos, only 600 miles off Ecuador, with the Marquesas at 2.7 per cent the next most likely hit. Hawaii cultivated the crop before European contact, and probably got it from Mexico on the basis of the simulation, but there was no further onward dispersal. This route might well have been used in the putative Polynesian-Californian contacts recently proposed (The Times, November 21, 2005).

The fact that 16 of the 23 target areas were hit with at least 1 per cent probability indicates “that vessel drifts provide many access routes from South America into Polynesia”, with hits on a particular island group coming from drifters starting on specific stretches of American coastline. These could have informed Polynesians of lands to the east, making two-way traffic possible.

The date by which all this happened remains debatable. Expansion east out of Tonga and Samoa may have begun as early as AD1, but perhaps not much earlier than AD1000, when the sweet potato is attested in the Cook Islands.

Easter Island seems, on the latest evidence, not to have been settled until around AD1200, so it could not have played a part in the initial transmission. In the end what this simulation experiment tells us is that purposeful voyaging, in either direction, was not necessary for this first, tenuous contact between the settlers who had moved out of Asia and around the Pacific rim to settle first the continent of America and then, much later, the ocean wastes of the Pacific.

Journal of Archaeological Science 35: 355-367.

Antarctic addendum

In my last column, I noted that “an Archaeology of Antarctica is on the brink of being written” (The Times, February 25th): in both Britain and Brazil scholars have now begun to do so. Andrs Zarankin and Mara Ximena Senatore have just published Historias de un pasado en blanco: Arqueologéa historica antartica (Stories of a white past: Antarctica historical archaeology; published by Argumentum, Belo Horizonte, Brazil). Mike Pearson of the University of Lampeter has written Professor Gregory’s Villa and piles of pony poop: early expeditionary remains in Antarctica, forthcoming in Contemporary Archaeologies: Excavating Now, edited by Cornelius Holtorf, of the University of Lund, and Angela Piccini of Bristol University (Peter Lang Publishers). My article gave Whale Island rather than Elephant Island as the place where Shackleton’s crew awaited rescue in 1916 — sorry.