Historical mysteries · Roanoke, 1587 · Leading candidate

What became of the Roanoke colonists?

The story

On 27 August 1587 John White sailed from Roanoke Island, which lies behind the Outer Banks, the chain of sand islands along the coast of what is now North Carolina, to fetch supplies from England. He was the governor of a colony of about 115 men, women and children that Sir Walter Ralegh, who held the queen's licence to settle the coast, had sent out that spring. It was the first English settlement in America to include families, and it followed a first attempt, about a hundred men under Ralph Lane, that had given up and gone home the year before. White's daughter Eleanor and her husband Ananias Dare stayed behind with their nine-day-old child Virginia, the first English child born in America. The colonists had asked White to go, because they were short of supplies and judged him the man most likely to bring them back.

War with Spain kept him in England. Shipping was held in port in 1588, the year of the Spanish Armada, his one attempt that spring was turned back by French privateers, and he did not reach Roanoke again until 18 August 1590, Virginia's third birthday. He found the settlement empty. The houses had been taken down, the site was enclosed by a high wall of tree trunks, the colony's boats were gone, and heavy iron and lead lay in the grass. On a post at the entrance, in capital letters, someone had carved the word CROATOAN. It was the name of an island about fifty miles to the south, now the southern part of Hatteras Island, and of the Indian people who lived there and had been the colony's friends. Before White left, he and the colonists had agreed a code: if they moved, they would carve the name of the place on a tree or a post, with a cross over it if they were in trouble; the word on the post had no cross. White took the word as a sign that they were safe at Croatoan, but a storm cost his ships their anchors before he could reach the island, and the voyage was given up. No Englishman is recorded as having seen any of the colonists again.

Why it is a mystery, and what each side says

No search reached the colonists: Ralegh sent ships in 1602 and 1603 that never got to the island, and Jamestown, the English colony founded on Chesapeake Bay in 1607, about a hundred miles to the north, sent two parties into the interior in 1608 and 1609 that came back with rumours. No house, grave or bone of the colonists has been identified since. Everything said about them after 1590 is hearsay: reports from Indians, heard at Jamestown through interpreters and written down by officers of the Virginia Company, the London company that financed Jamestown and had a commercial interest in finding survivors (section 3).

The two things the colonists left behind point in different directions. Before White sailed they had told him they meant to move "50 miles further vp into the maine", that is, fifty miles inland, which from Roanoke Island means west across Albemarle Sound, the shallow water behind the island, to the mainland where the Chowan and Roanoke rivers enter it. The word they carved names an island fifty miles to the south. Four answers have been argued. David Beers Quinn, the historian who edited the colony's documents in 1955 and wrote about it for thirty years, held that the main body went north to Chesapeake Bay, where the colony had been meant to settle in the first place, lived there for twenty years, and was killed by the people of Powhatan, the paramount chief of the Indian nations around Jamestown, at the moment the Jamestown ships arrived in 1607. Since 2010 the usual reading, set out most fully by the historian James Horn, has been that the colony divided: a small party stayed at Croatoan to watch for White, the rest crossed to the mainland at the head of the sound, and both groups were absorbed over the years into the Indian peoples around them. The First Colony Foundation, an archaeological group, has been digging on that mainland shore since 2012 and announced in June 2026 that it had found the colony. Scott Dawson, a Hatteras native who founded a local archaeological society, and the archaeologist Mark Horton, who have been digging on Hatteras since 2009, hold that everyone went to Croatoan and that the colony was never lost at all. A fourth possibility, that the colonists tried to sail home in their own boats, is weighed with the others. A carved stone that surfaced in 1937, known as the first Dare Stone because it claims to be a message from Eleanor Dare, has its own defenders and is judged on its own in section 7.

How the case was judged

The method was to read the whole surviving record and to rank the possibilities against it. The record is small: everything the English wrote about the colony before 1607 fills fifty pages of Richard Hakluyt's Principal Navigations, the collection of English voyages printed in 1598 to 1600; the Spanish side is a handful of documents; the Jamestown reports of 1607 to 1625 are about a dozen passages, each traced here to the person who first told it; the later testimony is four texts from 1650 to 1888. All of these were read in full, in the original spelling, from the editions named in each section. The archaeology of three sites was checked against what has been published, as distinct from what has been said in interviews, as of September 2026. The modern books were read in extracts and reviews, not cover to cover, and section 9 says which.

Section 6 then sets the possible fates side by side in a ledger: a table in which each group of evidence is given a weight for how well it fits each possibility, and the weights are multiplied through. The numbers in it are judgements, not measurements, and the table was run again nineteen ways with different weights to show how much the result depends on them. On that weighing the mainland at the head of Albemarle Sound comes first as the main body's principal seat, with about {{P_ALBEMARLE}} chances in 100, and between {{LO_ALBEMARLE}} and {{HI_ALBEMARLE}} across the nineteen alternatives. Croatoan is second at {{P_CROATOAN}} (range {{LO_CROATOAN}} to {{HI_CROATOAN}}). Chesapeake Bay and the sea are each below 10 in every case. The result is a ranking only: no source names the place, no structure or grave has been found, and the two excavations that could decide the question have not been published.

What this report covers

Section 1 sets out what the printed record of 1584 to 1603 says, in White's and the colonists' own words. Section 2 measures the "fifty miles" on a map. Section 3 traces every report heard at Jamestown from 1607 to 1625 back to the person who told it. Section 4 weighs the later testimony, from 1650 to 1888. Section 5 takes the archaeology as it stands in September 2026: the digs on Roanoke Island, on the mainland shore of the Chowan River and on Hatteras. Section 6 sets the possible fates against the evidence in the ledger and shows what changes when its weights change. Section 7 judges the carved stone of 1937 on its own. Section 8 places the verdict in the modern debate and credits the historians whose arguments it uses. Section 9 says what would overturn it, what is new, and what could not be read.

The verdict

The colonists left Roanoke in an orderly removal by water. The word they cut named, by the agreed code, the place where they were or meant to be seated, without distress. White read it as a token that they were safely at Croatoan; he never reached the island, and no witness recorded their arrival. Read together, the surviving reports point to a main body that moved on into the mainland at the head of Albemarle Sound, the fifty miles west it had announced, and dispersed among the peoples there. No witness recorded that sequence; it is inferred from the reports in sections 1 to 3. Some colonists were possibly killed by Powhatan's people about 1607. The Hatteras Indians who told the surveyor John Lawson, in his book of 1709, that several of their ancestors were white people are more likely than not descended in part from them.

The figures below are the judgements of section 6; the ledger and its code are published so that a reader can substitute other weights.

{{FIG_SCENARIOS}}

Figure 1. Chances in 100, on the weighing in section 6, that the main body's principal seat from about 1588 to 1607 was each place, with the range across the nineteen alternative weightings. Made for this report from the ledger in section 6. Each figure blends in an even split across the five places, at {{MIX}} parts in 100, to allow for evidence not in the ledger, such as an unread document or an unpublished excavation. The range is the lowest and highest value across the alternative weightings: each group of evidence left out in turn, no starting preference, the evidence counted at half strength, a larger allowance for the unknown.

Separate findings, judged on their own evidence

A · Some colonists went to Croatoan

{{FA_P}} in 100

Range {{FA_LO}} to {{FA_HI}}. The carved name, in the colonists' own code and without the distress sign, names Croatoan as their seat. Arrival is inferred from it, from the kin of Manteo, the Croatoan man who had twice been to England and was the colony's interpreter and ally, from the Hatteras tradition that Lawson recorded in 1709, and from unpublished finds at Cape Creek, the site of the Croatoan town on Hatteras, since no witness records them there.

B · The colony divided

{{FB_P}} in 100

Range {{FB_LO}} to {{FB_HI}}. White's text names two destinations, Lane had split his own colony to find food in 1586, every Jamestown report counts a handful, and the two archaeological footprints lie fifty miles apart.

C · Some killed by Powhatan's people, about 1607

{{FC_P}} in 100

Range {{FC_LO}} to {{FC_HI}}. The killing is stated in May 1609, in instructions from the Virginia Company, with a date and a place. Against it are three things: one informant lies behind every telling; the Company had an interest in the story; and the "confession" exists only as a note in a margin printed in 1625.

D · Descendants among the Hatteras in 1709

{{FD_P}} in 100

Range {{FD_LO}} to {{FD_HI}}. The basis is Lawson's testimony from the Hatteras themselves; shipwrecked sailors and Lane's men of 1585 to 1586 are alternative sources of white ancestors.

F · The first Dare Stone is genuine

{{FF_P}} in 100

Range {{FF_LO}} to {{FF_HI}}. Section 7 gives the grounds: the finder could not be traced; that summer at Manteo a stone relic was hawked from a suitcase and a scheme for a bogus relic had already been proposed; every statement on the stone can be found in the printed sources; and no test has ever been run on the stone.

Label
Leading candidate: several answers remain possible and are put in order. The first is ahead of each rival, and the order can change when one judgement changes. The leading one is a reconstruction from hearsay and from digs that have not been published, and every figure depends on the weights chosen in section 6.
Earlier accounts
Quinn (1974, 1984, 1985): the main body went to Chesapeake Bay and was killed by Powhatan in April 1607. Since 2010 (Horn; the journalist Andrew Lawler; the First Colony Foundation): the colony split between Croatoan and the Chowan and was absorbed. Dawson and Horton (2016 to 2026): all of them went to Croatoan.
How this report stands to them
This report agrees with Horn, Lawler and the First Colony Foundation that the colony split and was absorbed, at lower confidence than they state. It moves the weight of the slaughter story: it is one London report of 1609, which itself dates the killing to 1607 and places it outside Powhatan's territory, so it does not support the Chesapeake. It overturns no earlier account.
What is new here
{{WHATS_NEW_SHORT}}
What could not be read
Quinn's two books, Horn's and Lawler's were read in extracts and through reviews, not cover to cover; the Zúñiga chart, a manuscript map of Virginia of 1608 known from the copy the Spanish ambassador sent to Madrid, was read in three published transcriptions, not the original; Site Y, one of the two dig sites on the Chowan shore, and the hammerscale found on Hatteras in 2024 (the flakes struck off iron when it is forged) have no published data to weigh.
Materials
The notes, raw texts, code and data behind it are published with the report.