Every English document on the colony before 1607 is in Richard Hakluyt's Principal Navigations, the collection of English voyages that Hakluyt, a clergyman and geographer, printed in 1589 and enlarged in 1598 to 1600. The 1600 edition, vol. III, pp. 246 to 295, carries the report of Arthur Barlowe, a captain on the reconnaissance voyage of 1584; the journal of the 1585 voyage; the account of Ralph Lane, governor of the first colony, about a hundred men, in 1585 to 1586; the relief voyage of 1586; the report of Thomas Harriot, the scientist of Lane's colony (1588); White's narratives of the 1587 and 1590 voyages; the list of colonists; and White's letter of 1593. The 1589 edition alone carries White's failed relief voyage of 1588 (pp. 771 to 773). All of them are read here in full and in the original spelling, from Edmund Goldsmid's nineteenth-century transcription checked against the Glasgow (MacLehose) edition of the early 1900s and the scan of the 1600 volume. The Spanish side comes from Irene Wright's translations of documents in the Spanish archives, published in 1951 by the Hakluyt Society, which prints editions of voyages, and from the Relación of Luis Jerónimo de Oré, a Franciscan friar who wrote up the Spanish voyages along this coast, in Maynard Geiger's translation (1936). The searches of 1602 and 1603 are in a tract by John Brereton (1602), Ralegh's letter of 21 August 1602 and the collection of voyages that Samuel Purchas, Hakluyt's successor, printed in 1625.
On 22 August 1587 the colonists pressed White to go home for supplies. He refused at first, and one of his reasons is the only statement of the colony's intentions that survives:
"Also he alleaged, that seeing they intended to remoue 50 miles further vp into the maine presently, he being then absent, his stuffe and goods might be both spoiled, and most of them pilfered away in the cariage".
White, 1587 voyage, Hakluyt 1600, III, p. 285; Hakluyt's margin: "Their meaning to remoue 50 miles into the countrey."
The move was to be "presently", that is at once, and by carriage of goods. Fifty miles from Roanoke Island west across Albemarle Sound reaches the mouth of the Chowan River and the shore of Salmon Creek, which enters the Chowan there (straight-line distances are in section 2). It does not reach Chesapeake Bay, which Lane had put at 130 miles and which had been the seat Ralegh ordered in writing (III, p. 281: "the Bay of Chesepiok where we intended to make our seate and forte, according to the charge giuen us among other directions in writing, vnder the hande of Sir Walter Ralegh"). Simon Fernandes, the pilot of the 1587 voyage, had refused to carry the planters further than Roanoke.
The colonists' own only surviving words are the testimony of 25 August 1587 asking White to go "for the present and speedy supply of certaine our knowen and apparent lackes and needes" (III, p. 286). They were short of supplies from the start. White sailed on 27 August. The pinnace, the colony's own small sailing vessel, and the ships' boats are never recorded as leaving with him, and in 1590 he looked for "their botes or Pinnesse" and found none; the colony had therefore kept its vessels.
"as we entered vp the sandy banke vpon a tree, in the very browe thereof were curiously carued these faire Romane letters C R O: which letters presently we knew to signifie the place, where I should find the planters seated, according to a secret token agreed vpon betweene them and me at my last departure from them, which was, that in any wayes they should not faile to write or carue on the trees or posts of the dores the name of the place where they should be seated; for at my comming alway they were prepared to remoue from Roanoak 50 miles into the maine. Therefore at my departure from them in An. 1587 I willed them, that if they should happen to be distressed in any of those places, that then they should carue ouer the letters or name, a Crosse + in this forme, but we found no such signe of distresse."
White, 1590 voyage, Hakluyt 1600, III, p. 292 (the printed text reads "in the yeere 1586" for 1587 a few lines earlier; the error is in the 1600 original).
"we found the houses taken downe, and the place very strongly enclosed with a high palisado of great trees, with cortynes and flankers very Fortlike, and one of the chiefe trees or postes at the right side of the entrance had the barke taken off, and 5 foote from the ground in fayre Capitall letters was grauen CROATOAN without any crosse or signe of distresse; this done, we entered into the palisado, where we found many barres of iron, two pigges of lead, foure yron fowlers, Iron sacker-shotte, and such like heauie thinges, throwen here and there, almost ouergrowen with grasse and weedes. From thence wee went along by the water side, towards the poynt of the Creeke to see if we could find any of their botes or Pinnesse, but we could perceiue no signe of them, nor any of the last Falkons and small Ordinance which were left with them, at my departure from them."
Hakluyt 1600, III, pp. 292 to 293.
Five chests, three of them White's, had been dug up "long sithence" and their contents spoiled. White wrote that this "could bee no other but the deede of the Sauages our enemies at Dasamongwepeuk, who had watched the departure of our men to Croatoan" (III, p. 293). Dasamonquepeuc was the Indian town on the mainland shore opposite the island, hostile to the English since 1586. That sentence is White's inference; he names no informant.
The site reads as an orderly and unhurried removal by water. Houses were dismantled, a fortification was built after August 1587, everything portable and every boat and light gun was taken, and only the heaviest iron and lead was left behind. The code told White two things: the planters had named Croatoan as the place where they were or meant to be seated, and they were not in distress when they cut the word. White took it as "a certaine token of their safe being at Croatoan" (III, p. 293). He never landed on the island, and no later witness records finding them there. The word did not say whether all of them went, whether they arrived, how long they stayed, or where they went afterwards.
On 30 July 1587 Edward Stafford, one of the colony's captains, and Manteo, the Croatoan man who had twice been to England and served the colony as interpreter, went with twenty men to Croatoan, Manteo's mother's island. The people there "desir[ed] vs not to gather or spill any of their corne, for that they had but little" (III, p. 282). In the famine of spring 1586 Lane had used the same island only as a station for Stafford and twenty men "to feede himselfe and his company, and also to keepe watch if any shipping came vpon the coast" (III, p. 262). A hundred and sixteen extra people could have lived on the island for a season; nothing in the record suggests it could have fed them for years. Croatoan was also the natural place to watch for White's ships, and the Croatoans were the only people on the coast who were friends. The record therefore fits Croatoan as a first station, and the mainland move as the sequel the colonists had announced, better than it fits either alone.
Spain, which claimed the whole coast from its garrison at St Augustine in Florida, looked for the English settlement three times. Pedro Menéndez Marqués, the governor of Spanish Florida, sailed the coast to 37 degrees in May and June 1587, two months before White's colonists landed, and found no one (Wright 1951, doc. 61). In June 1588 Vicente González, a captain sent north by Menéndez Marqués, explored the Chesapeake to its head, and on the way south, about 30 June or 1 July, ran into a shallow bar at about 35 and a half degrees. Geiger's translation of Oré reads: "In the interior of the region where they were, there was a ship-yard, indicated by the presence of sloops, while on the land there were some remains of English barrels. Besides, there were other signs of debris, indicating that people had been there" (Geiger 1936, p. 48). They saw no people and did not go to the island. Quinn's translation, which could not be checked here against the Spanish, has a slipway and wells made of English casks. In March 1589 Pedro Díaz, a Spanish pilot whom Sir Richard Grenville, the commander of the 1585 fleet, had captured in 1585 and who had sailed on two of the English voyages before escaping, guessed that the settlers "will have died of hunger" (Wright 1951, doc. 65). A Spanish expedition of 300 men to 35 and a half degrees was ordered in 1589 and cancelled in 1590 (Geiger, pp. 50 to 52). No Spanish force ever reached Roanoke or Croatoan.
White gave up in Ireland in February 1593, still writing of "my discomfortable company the planters in Virginia" as people to be relieved (III, p. 288). Hakluyt wrote to Sir Robert Cecil, the queen's secretary of state, on 24 October 1599 that of the English colonies in Virginia "one as yet remaineth, for ought we know, alive in the countrey" (epistle to the second volume). Brereton's tract of 1602 records that Ralegh had sent "fiue seuerall times at his owne charges" to find "those people which were left there in the yeere 1587", and that the bark (a small ship) of Samuel Mace, the captain Ralegh sent in 1602, stayed a month forty leagues south-west of Hatteras and never searched the port it was sent to (Brereton 1602, p. 14). Ralegh told Cecil on 21 August 1602 that neither of his barks "spake with the peopell" and that he would "yet live to see it an Inglishe nation". Bartholomew Gilbert's voyage of 1603 aimed at "the Bay of Chesepian ... to seeke out the people for Sir Walter Raleigh left neere those parts in the yeere 1587", and Gilbert was killed on landing (Purchas 1625, IV.viii.12). No Englishman is recorded on Croatoan after 1 August 1587. White anchored off its north-east point on the night of 12 August 1590 without landing, and the lost anchors of 19 August ended the attempt to go there.
Nothing in this record names Powhatan, the Chesepians (the people of the south shore of Chesapeake Bay, among whom a party of Lane's men had wintered in 1585 to 1586), or the inland towns of Ocanahowan and Ritanoe. Those names enter with Jamestown, in section 3.