Despite the distance, all the living evidence suggests that sometime during the first millennium, a small group of Austronesians crossed the Indian Ocean via canoe to become some of the first people to settle one of the last uninhabited places on Earth. The author and environmental historian Jared Diamond calls it "the single most astonishing fact of human geography for the entire world."
On the other hand, ancient evidence for this unlikely story of Madagascar's colonization has been almost impossible to find. For years, archaeologists have sought to uncover remnants of the island's early settlers that would connect them to ocean-faring Austronesians — but those efforts always came up empty or inconclusive. Meanwhile, in the past decade, their colleagues have found signs that African hunter gatherers were there more than 2,000 years ago.
"Linguistic and genetic studies … only give us the modern picture," Alison Crowther, an archaeologist at the University of Queensland, wrote in an email. "Archaeology is important for providing direct evidence of how, when and why past people made this journey across the Indian Ocean, including where else in eastern Africa they might have landed."
So Crowther and her colleagues decided to investigate a new line of evidence: ancient plants. Excavations of thousands of gallons of sediments from 18 sites across Madagascar and other islands off the east African coast revealed that Madagascar and the Comoros archipelago were dominated by ancient Asian cultivars, especially mung beans and rice.
"These crops provide the first, to our knowledge, reliable archaeological window into the Southeast Asian colonization of Madagascar," the authors reported Tuesday in the Proceedings of the National Academy of Sciences.
Past genetic studies have suggested that settlers from the east arrived in Madagascar about 1,200 years ago — perhaps against their will. In 2012, scientists reported that they'd traced the lineages of several hundred Malagasy people back to a group of just 30 women of mostly Indonesian descent. This is exceptionally small for a founding population (even though New Zealand probably was colonized by a similarly tiny group), and there isn't much evidence that ancient Indonesians attempted to start other trading colonies.
It may well be, the scientists speculated, that the first people to reach Madagascar were passengers on a boat that had blown way off course. There's plenty of precedent for this theory: During World War II, a survivor of a shipwreck near Java showed up on Madagascar days later, hungry and exhausted but still alive. It's also possible that the earliest Madagascans were refugees, forced to sail from their homes in southeast Asia in search of a new place to live.
Whatever their motivation for crossing the Indian Ocean, Crowther's study helps parse out what happened when they arrived at the other side. Analysis of charred, microscopic remnants of seeds and other plant parts showed that they arrived first in the Comoros, and then in Madagascar, carrying with them the crops of their homelands.
This is something of a surprise, because the people of the Comoros islands today have no obvious linguistic, ethnic or genetic connection to southeast Asia. But the archipelago has had a tumultuous history, marked by centuries of trading and slave raiding, which could have diluted the initial Austronesian influence.
And the find could help explain how modern people from Madagascar acquired such a mix of African and southeast Asian ancestry.
"This makes sense in many respects," Crowther said, "as it has long been suggested that the Comoros were stepping-stone islands between the African mainland and Madagascar."
She and her colleagues hope to conduct more genetic analyses to clear up this mystery. They also plan to continue their excavation of the islands in search of more botanical data, especially from Asian plants such as banana and taro that don't produce seeds, and are harder to find records of. That involves scraping microscopic remains from shards of pottery and the tarter on ancient teeth. (Yum.)
"This would show us that people were actually processing and eating these foods as well," Crowther wrote.