On August 18, 1590, the English navigator John White stepped ashore on Roanoke Island, off the coast of what is now North Carolina, expecting to find the colony he had left three years earlier. He found instead an abandoned fort, a few overgrown gardens, and the word "CROATOAN" carved into a wooden post at the entrance. There was no body. There was no sign of violence. There were 115 men, women, and children missing, including White's daughter Eleanor Dare and his granddaughter Virginia Dare, the first English child born in the Americas.

White had been instructed before his departure that if the colonists left the settlement, they should carve their destination on a tree or post, and that if they had been forced to leave under duress, they should add a cross. There was no cross.

White attempted to sail south to Croatoan Island, the home of a friendly Algonquian-speaking community, but a storm prevented him from reaching it. He returned to England. He never came back to North America. He died in Ireland, probably in 1593, without learning what had become of his family.

The fate of the Roanoke colonists has been argued over for more than four centuries. The theories have included starvation, massacre, integration with neighboring Native peoples, voluntary relocation to the Chesapeake Bay area, and various combinations of these.

For most of the twentieth century, the dominant academic view favored integration. The colonists, this argument went, joined the Croatoan community on what is now Hatteras Island and were absorbed into Algonquian society. The carved word on the post was a forwarding address.

Recent archaeology has tested this hypothesis in several ways.

Recent archaeology has tested this hypothesis in several ways.

The Croatoan Archaeological Society, working on Hatteras Island since 2008 under the direction of Mark Horton at the University of Bristol, has excavated a Native village site at Cape Creek. Among the finds are sixteenth-century European objects that appear to be in domestic use rather than as ritual or trade goods. A rapier hilt, fragments of slate writing tablets, a navigator's astrolabe component, and lead seals of the type used to mark fabric bales have all been found in stratigraphic contexts dating to the late sixteenth and early seventeenth centuries.

What these objects mean is contested. They might represent traded items that the Croatoan acquired and used. They might represent the presence of English settlers living within the community. The difference is significant but archaeologically subtle.

A separate project, the First Colony Foundation, has worked on the western shore of Albemarle Sound, roughly 80 kilometers from Roanoke. A 2012 reexamination of a 1585 map drawn by John White, now held by the British Museum, revealed two patches where the map had been altered. Under infrared light, the patches showed a hidden symbol — a four-pointed star within a square — at a location near the confluence of the Chowan and Roanoke rivers. The mark appeared to indicate the site of a planned secondary settlement.

Excavations at this location, called Site X by the foundation, have produced sixteenth-century English ceramics, including Border Ware and a fragment of a tenant farmer's tankard, in contexts that predate any documented English presence in the region. The ceramics were not trade goods. They appear to have been used by English-speaking households over a period of years.

Taken together, these two sites suggest a possible bifurcation. Some of the colonists may have gone to Croatoan with the Croatoan people. Others may have moved inland to the Albemarle Sound area, possibly to join a small group of advance settlers White had sent ahead before his departure.

This reading is consistent with what is known about the colonists' situation in 1587. They had landed at Roanoke against their wishes — they had been intended for the Chesapeake — and they faced a strained relationship with the local Roanoke people. Splitting into smaller groups and joining neighboring communities for survival would have been a reasonable response to abandonment by their supply ships.

What the archaeology cannot yet establish is what happened to these smaller groups over the following decades. The Powhatan Confederacy, which dominated the Chesapeake Bay region in the early seventeenth century, included some communities with reports of European-looking inhabitants. The Jamestown colonists, arriving in 1607, were told stories by Powhatan informants of white people living to the south and west. Some of these reports may refer to Roanoke survivors. Others may refer to other Europeans entirely.

A persistent claim, dating to the early twentieth century, identifies the Lumbee people of present-day Robeson County, North Carolina, as descendants of the Roanoke colonists. Genetic studies have not supported a specific Roanoke connection, though the Lumbee are a community of complex and partly European ancestry whose origins are themselves the subject of ongoing research.

What is clear is that the colonists did not vanish. They went somewhere, and the somewhere is being slowly reconstructed by excavation, document analysis, and patient comparison of artifact assemblages.

The Roanoke story has long served as a kind of American gothic, a tale of the wilderness swallowing a community whole. The archaeology is unromantic but more accurate. Small groups of people, under pressure, made survival decisions and dispersed. Some of them probably lived for years among Algonquian neighbors. Some of them probably died of disease or hunger. Their descendants, if any, are not identifiable.

The word on the post was not a mystery. It was a note left for John White, who could not reach the place it pointed to. Four hundred years later, archaeology is finally getting close.