AMERICAN HISTORICAL ASSOCIATION. THE SURROUNDINGS AND SITE OF RALEIGH'S COLONY. TALOOTT WILLIAMS, OP PHILADEiSPHIA, pa. (From the Annual Report of the American Historical Association for 1695, pages 47-61.) I WASHINGTON: GOVERNMENT PRINTING OFFICE. 1806. AMERICAN HISTORICAL ASSOCIATION. THE SURROUNDINGS AND SITE OF RALEIGH'S COLONY. TALCOTT WILLIAMS, OF PHILADELPHIA, PA. (From the Annual Report of the American Historical Association for 1895, pages 47-fil.) WASHINGTON: GOVERNMENT PRINTING OFFICE. 1896. Y 22^ ."Wn2- By Tranatof. MAY 17 1910 THE SURROUNDINGS AND SITE OF RALEIGH'S COLONY. By Talcott Williams. A strange Nemesis has attended every original site of coloni- zation on the North American continent. Not one has waxed great and prospered. Most stand to-day desolate. Plymouth is but a village. The island which the Dutch first occupied below Albany is as empty of men as when the first blockhouse was built upon it. Jamestown is an open field. First and earliest of all, the scene of Ealeigh's colony, Roanoke Island, has to-day a population probably not much larger than when it was discovered, and the site of the colony itself has held no dwelling for three centuries. The occupation of civilized man has left its visible mark and change on most of our coast, but the shores of the two great sounds of North Carolina, Albemarle and Pamlico, and the various islands which separate them from the ocean — the waters Avhich first received English keels and the lands which were first occupied by English-speaking men — are to-day, for leagues together, as they were first seen. Nothing has altered. The long, low island, pictured by De Bry from White's draw- ing is still a better sketch of Eoanoke than any published since — far superior to the somewhat ridiculous print repeated in school histories from a modern magazine. 1 The rows of white swan still rise at a shot as they rose at the report of Barlowe's arquebus. The flat line of the horizon, the amazing luxuriance of vegetation (1,800 species in a single pocoson' 2 ), the wilderness of bird life, the wine-colored waters of Albe- marle, the shifting shoals which connect it with Pamlico, the tempestuous ocean without and the calm sounds within — these all still repeat in minute detail the narratives of Lane, of Hariot, of Amadas and of Barlowe, and the sketches of John White. 'Harper's New Monthly Magazine, 20:730 (May, 1860), "Louugings in the footprints of the pioneers," by Edward C. Bruce. 2 Goldthwaite's Geographical Magazine, 11:373 (May, 1892), "Physiog- raphy of a pocoson," by Charles Hallock. 47 48 AMERICAN HISTORICAL ASSOCIATION. This permanence of physical conditions, untouched and unchanged by man, lends a singular interest to this forgotten corner of the continent aud sheds a special illumination on the narratives of the expeditions which we owe chiefly to the care of Hakluyt. History the colony of Raleigh has had in abun- dant measure, particularly in the last decade, the last and fullest account having been read before the American Historical Association by Dr. Stephen Beauregard Weeks at its meeting in December, 1890. 1 Into the narrative of the colony which begins with the voyage of Amadas and Barlowe in 1584 2 , ends with the return of White in 1590, and is prolonged by the search for the colony in 1602, 1008, and 1010, I do not propose to enter ; but I have twice devoted the scanty recreation of a journalist to a visit and examination of the site, once in No- vember, 1887, and again in November and December of this year ; I have sailed over the waters of the region in an open boat from Edenton to Hatteras and I have given the physical con- ditions of the region and the j)resent state of its remains a direct and practical examination with sail, lead, and spade, supplemented by a study of the physiography of the region which may, I trust, collectively throw light on the written record. For it can not be too often repeated or too well remembered that the current of history flows in channels furnished by the earth's surface and that every narrative, however full, however accurate, however near, and however remote, needs for its fall comprehension the study of the region in which its events took place, its institutions were formed, its greater figures produced, and its battles decided. Without this background and foun- dation, history is but a succession of shadowy and shifting- scenes " whose worth's unknown although their height be taken." 1 American Historical Association V : 107. The Lost Colony of Roanoke : Its Fate and Survival by Prof. Stephen Beauregard Weeks. 1 ( Magazine of Am. History, 29 : 459 May-June, 1893. Raleigh's New Fort in Virginia, 1585, by Edward Graham Daves.) (New England Magazine, N. S., 11:565, Jan- uary, 1895. Raleigh's Lost Colony, by James Phinney Baxter.) (Canadian Magazine, 4:500, April, 1895, Lost Colony of Roanoke, by E. Y. Wilson. 2 Philip Amadas and Arthur Barlowe sailed in 1584, and discovered Roanoke and the Carolina sounds. Sir Ricbard Grenville in 1585 lauded over 100 persons on tbe island who were brought back by Drake in 1586. In the same year Sir Richard Grenville left 15 persons on the island who were all slain by the Indians. Raleigh's second colony, 117 souls, went out in 1587, and settled at Roanoke. When tbe island was visited in 1500 by Ralph Lane, their governor, all had disappeared, and it is about the deserted site that there centers tbe interest which still attaches to Roanoke and the first English colony on this coutinent. SURROUNDINGS AND SITE OF RALEIGH'S COLONY. 49 The long* rampart of sandy islands which shut in the two sounds of .North Carolina, between which lies the island of Itoauoke, constitutes, taken together, the physical feature on the Atlantic Coast, whose conditions have changed less and whose hydrography has altered more than any stretch as long- on the continent. Now, as then, two rivers, the Roanoke and the Neuse, and numerous lesser streams, pour their waters into these shallow sounds. 1 Now, as then triple forces, the (iulf Stream, the prevailing northeast and southwest winds, and these rivers, heap these sand-bars and fill with silt the space behind them. But while this process, in progress from the earliest days of the current emergence of the coast along the line of the Cincinnati upheaval, produces the same condi- tions and leaves the same general outline of coast and the same low horizon of sand-dune swamp and wooded islands, the outlines of the coast steadily alter .as land and sand en- croach on the sea. Nowhere else are general outlines more permanent. Nowhere else are specific boundaries and physical features more transitory. Much ingenuity has been expended, particularly by those who have never visited the region, in determining the exact course followed by the voyagers of three centuries ago ; but as it is morally certain that no one of the inlets now open was open then, with possibly a dubious excep- tion at the southern end in Ocracoke — if this was Wokokok — the attempt to decide this question absolutely is a fruitless labor. The utmost which can be done is to reach approximate conclusions. In our own brief day, Hatteras — opened in 1840; in 1860 the accepted gateway of the entire system of sounds — has begun to close, and can no longer be entered even by schooners of moderate size. Without the coast, off Hatteras, Diamond Shoals alter so rapidly that their rapid changes have thus far baffled the most astute and experienced of light- house builders and submarine engineers, Capt. John F. Anderson, who, in 1892, lost some $100,000, by the destruction of his caisson, to learn that the soundings of one year on this tem- pestuous elbow of the continent are all altered by the storms of the next winter. Steadily the winds carry the sands grind- ing along the coast from Cape Cod to Florida, until the char- acteristic detritus of the New England coast can be traced a 'The area, of these streams is: Nens<- 5,299 and Roanoke 0,237 square miles. Tbe area of the sounds is approximately 3,500 square miles. H. Doc. 2dl 1 50 AMERICAN HISTORICAL ASSOCIATION. thousand miles south; steadily the ocean currents and storms together dig breaches in the sandy rampart, and as steadily are they filled by the southward march of sand and the out- ward flow of fluviatile deposits; but in no ten years together do these varying conditions produce exactly the same result in bar, inlet, and channel. When it is once clearly understood that inlets have been and still are opening and closing, like doors on a hotel corridor, along the entire line of this coast for three centuries, it will be seen what a fruitless labor it is to endeavor to determine by exactly which inlet Amadas, Bar- lowe, and their successors entered by applying the uncertain record of the successive navigators from 1584 to 1590 to our still more uncertain knowledge of the region then and our none too certain acquaintance with it now. Very nearly every inlet 1 now upon our maps has been credited with furnish- ing an entrance to the voyagers during the period, now approaching two centuries, in which the subject has been under active discussion. But of the ten inlets which have been open at intervals into these sounds since 1580, only one, Ocracoke, has been open through that period, and it is not improbable that this was closed during part of the seventeenth century from a reference made to its navigation. 2 1 The creation of the beaches and tidal marshes of the Atlantic Coast has been luminously discussed by Prof. Nathaniel Southgate Shaler. (Report United States Geological Survey, X, 147, and in National Geographic Mon- ographs, I, 137-168.) 2 Beginning at the north the varying authorities are: Byrd, opposite Collidon Island, now Collington Island; Welsh and Weeks, Caffeys Inlet; Hawks and Dover, New Inlet; Ruffin, Roanoke Inlet; Bancroft, Abert, and Moore, Ocracoke, identified with Wokokok. Of these the last appear to me the least and the first the most probable. The three principal dis- cussions of the physiography of this region in connection with this subject are: Bulletin of the Essex Institute, XVII, Nos, 1-3. An account of the cut- ting through of Hatteras Inlet, North Carolina, September 7, 1816; also through which inlet did the English adventurers of 1584 enter the sounds of North Carolina, and some changes in the coast line since their time, by William L. Welsh. Appendix G of the Annual Report of the Chief of Engineers for 1876, being annual report upon the improvement of rivers and harbors in the District of Columbia, Virginia, and North Carolina, in charge of S.T. Abert, United States civil engineer. Ex. Doc. H. R., Forty-first Congress, third session (January 18, 1871), Engineers' Report on Certain Rivers and Harbors, contains, pp. 52-59, report of J. H. Simpson, colonel engineers and brevet brigadier-general, United States Army, on Roanoke Inlet and its proposed reopening. Mr. Welsh's contribution, while brief, is the most importantof all, because it was the first to grasp the fact of frequent changes and to note that " Hat- arask" and " Hatteras" are miles apart. Mr. Abert's report is the fullest SURROUNDINGS AND SITE OF RALEIGH'S COLONY 51 While a specmc determination of all the places and inlets mentioned in these early itineraries is now impossible, a gen- eral comprehension of the coast as they found it is as important and perhaps more instructive. To-day there is no entrance to the sounds north of Oregon Inlet, fast filling-, and there will soon be none above Ocracoke, at present the only practicable ship channel. Both sounds have made considerable progress toward their ultimate destiny of land-locked waters slowly fill- ing up to the condition of the Dismal Swamp, or, better, drained and turned into fertile lands. In Byrd's time the most north- ern arm of these sounds, Currituck, 1 could be entered by vessels and most satisfactory discussion of the physical condition of the region, moved by the not unnatural circumstance that he was a better engineer than historian, and failed to note past changes while studying the current situation. Colonel Simpson's paper summarizes the physical history of the region immediately about and opposite Roanoke Island. His conclusion that the voyagers could not have entered at Roanoke Inlet (now Nags Head) is probably 7 accurate, but this omits the still more important fact that an inlet undoubtedly existed just above Nags Head by which they did in all probability enter. Nearly every historian of North Carolina has made an attempt to answer the geographical questions involved in the accounts of these voyages, most of them by resorting to the charts of their own day, with little comprehen- sion of the physical history of the region, its unceasing change, and its early 7 condition. The first indispensable apparatus for the study of this problem are the early narratives, the charts of White and Hariot, and the Coast Survey charts of the region. The gap between the outline of 1586 and the coast as it is to-day can only be filled by a careful study of inter- vening charts, nearly every one of which throws some light on the problem. These consist of three classes — the outline sketches of early navigators exteudiug over the first century, colonial surveys over the next century, and modern charts over the past one hundred years. The first have become familiar in facsimile, and it is unnecessary to specify them. Exact knowl- edge begins with the accounts and maps of John Lawson, surveyor-general, 1708, and William Byrd, of Westover. Wimbler, 1730, republished by act of Parliament, and Emanuel $owen, 1763, are the most important of the colonial charts. Modern surveys and charts may be fairly said to begin with Daniel Dunbibin, 1764. This was superseded by the State surveys made with a view to a canal in the early years of the century. (Murphey, 1816, and Hamilton Fuller, 1818; North American Review, January, 1821.) The reports of army engineers begin with that of Col. W. K Armistead, December 15, 1820, and come down to the present time. The Coast Survey charts cover the last half century, and their comparison is important. 'Currituck Inlet was closed in 1828 (Ruffin, 116), but countless maps still carry it, and even the "Map of the United States and Territories," 1882, issued by the Land Office, has the familiar gap for Currituck Inlet. There is probably nothing so lasting as a geographical error, except a fictitious historical anecdote. 52 AMERICAN HISTORICAL ASSOCIATION. drawing 10 feet, and there was a succession of inlets along the coast. Earlier the number of inlets was still greater. The bear- ings of the Long Shoal indicate its early existence under condi- tions similar to those which now create the Diamond Shoal. The general coast line has probably been traveling to the eastward, working out under aeriel and aquatic influences, mod- ified by tbe slow secular change which once elevated and is now probably depressing the entire region. During the cen- tury in which we have definite information we know that the inlets have been closing from north to south, and the waters just inside of the bars steadily shoaling. The first maps show scattered and not continuous islands. 1 Even White's map, which is extraordinarily accurate, shows the inclosing islands wider and the inlets broader than to-day. It is not improbable that in 1654 there were a series of islands of considerable size, separated by inlets, which, at Trinity Harbor, just north of Eoauoke, gave a broad entrance and an anchorage safe from any but southeast winds, and represented now by the fresh-water lakes north of Nags Head, the channels about Collington Island, and the remains and memory of Caffeys Inlet, closed in 1800, and Eoanoke, closed in 1806. In all the maps of the middle of the last cen- tury there were not one but two inlets here, or one divided by an island, giving wider and easier access than at any other point. 2 In addition, while the inlets below and near Cape Hat- teras are shut off from ocean approach by shoals extending far to sea, of which early mention is made, 3 these shoals dis- appear north of the turn of the cape. Once inside Ocracoke, also, while there is a broad stretch of water apparent, the expanse is shallow, the channel through the swash inside is not to be readily found, and this passed, the work of finding a way even for a pinnace from the south to Eoanoke Island would not be easy. Inside and outside, therefore, the advan- tages of navigation are all in favor of an entrance north of Eoanoke and against an entrance below. 1 This is particularly true of a map of 1666 (Winsor, v, 338), Mordeu, 1687, and Powers, 1763, and John Mitchell, 1755, give the same impression. At the latter date Hatteras Island was six or seven times larger than to-day. "This is true of Joshua Fry and Peter Jefferson, 1755; Robert de Van- goudy, 1755; Emanuel Powers, 1763; William Faden, 1793, and the map with Thomas Jefferson's Notes on Virginia. y Pedro Meuendez, Morquez, 1573. SURROUNDINGS AND SITE OP RALEIGH'S COLONY. 53 I doubt if anyone could go over the coast without and within and not reach the conclusion that the most probable landfall of the first navigators drawing up from the south would be the shoals which make off from Capes Fear and Hatteras, and that these would hold them too far away to make any inlet until they reached the distinct break in the coast line marked and given the name, with good reason, as Trinity Harbor on the coast, and that this lay at a point cut by various inlets, of which Oregon, far to the south of the old opening, is to-day the solitary representative, but which then was in the condition whose traces are now apparent in the region just described. To Mr. W. L. Welsh must be given credit for call- ing attention to this point first named by William Byrd. 1 The crucial argument in favor of entrance north of Eoanoke is that the island is always approached from this direction. It was at the ''north end thereof" that Barlowe found " a vil- lage of nine houses." If he had approached from the south he would have noted the other Indian village, whose remains are to-day abundantly visible on the island back of Round Ten Oak Island. It was "round about the north point of the island" that Ralph Lane sought his colony. Moreover, the upper end of this entrance was 35 miles (7 leagues) from Roanoke Island. Oollington, then of larger size, furnishes the first island of Amadas and Barlowe. Approached from the sea it would seem the mainland, and on it the tradition repeated by Byrd places the scene of taking possession of the land. More- over, starting from this point, with the prevalent wind of the region, it would be easy to run to the mainland "20 miles'' away, the Alligator River, Occam, nearer here than Roanoke, and from thence to seek Roanoke. Coining from the south, Roanoke would be almost certainly the first landing made in the Sound. "Kendrick's Mounts" are, in all probability, the conspicuous sand hills near ISTags Head, the highest on the coast, 100 feet high, and to day marked objects and fronting dangerous shoals. Nor must it be forgotten that while in this century and in the last half of the eighteenth the inlets of Pamlico Sound have been the chief channels of commerce, in the seventeenth century and the first half of the eighteenth '"Not far from Roanoke Inlet. They ventured ashore near tfaat place upon an island now called Colleton, where they set up the arms of Eng- land." (William JJyrd, Notes, p. 12.) 54 AMERICAN HISTORICAL ASSOCIATION. Albemarle was the more accessible sheet of water. The balance of evidence of record, of tradition, and of physical conditions is therefore all in favor of an entrance at some inlet, which even in Byrd's time had disappeared, north of Roanoke Island and not far from Colliugton Island. This approach decided the point at which the two colonies planted by Raleigh were established on the higher ground at the northern end of the island, where a low, quadrangular mound has been identified since 1654 as its site. This island, which is now 1_! miles by 3 and was three centuries ago 16 miles long and a half mile or more wide, is entered at the northwest, on the side toward Croatau Sound, by the Alder Branch below Weirs Point, and on the eastern face by Shallow Bag Bay, on which now stands the little village of Manteo, and about which Daniel Baum and other early settlers lived when the island was reoccupied. This harbor is the natural site for a settlement on the island. If another was earlier selected, it was for reasons due to the approach and the site of existing Indian villages. Indian remains are numerous on Roanoke Island and their careful study would probably do much to solve two important problems, the advance of the early red man along the coast and the first contact of his last descendants with the white man. Four recent Indian village sites w^re examined by me, one at the southern end, where extensive shell mounds have been reduced in extent by their use as an easy fertilizer, but on which a clearly marked mound, 600 by 200 feet in size, fronts on an old canal or waterway cut through the swamp for a mile. Another extensive Indian deposit is on Baums Point, 1 most of which has been eaten away by the encroachment of the sound on the island, which a few years ago, about 1870 to 1S75, laid bare a number of Indian skeletons. Opposite this point, across the mouth of Shallow Bag Bay, is Ballast Point. Off this marshy projection is a mass of stones under from 3 to 4 feet of water, covering a space about 'This is the point now known as Dolhys Point, on which Martin (1: 35) places "the stump of a live oak said to have been the tree on which this word (Croaton) was cut, was shown as late as the year 1778 by the people of Roanoke Island. It stood at the distance of about 6 yards from the shore of Shalon-bas-bay, on the land then owned by Daniel Baum. This bay is formed by Ballast Point and Baums Point." Baum's descendant, Mr. B. F. Meekiu, now holds the site of the original Baum farm. SURROUNDINGS AND SITE OF RALEIGH'S COLONY. 55 40 by 20 feet. These stones tradition for over a century has alleged to be ballast from Baleigh's first two vessels. They may possibly have reached this point. I procured a quantity of this ballast. It is made up of round and angular stones of quartzite, porphyritic rock, and greenstone of a few pounds of weight, some cleaved. It looks extremely like the raw mate- rial of an Indian workshop for the manufacture of arrowheads and stone axes, but I suspend judgment awaiting a competent mineralogical determination of the material. A third Indian settlement is on the northeastern angle of the island, much of which has also been gradually swept away by the sound and the shifting sand dunes. The most important Indian remains are, however, the mounds on the Alder Branch, which stand about 100 yards south of the corduroy bridge, thrown over the creek during the war, and part of "Burnside avenue." 1 This low, but clearly artificial, mound contains closely packed in a sitting posture a great number of skeletons so decomposed that no bones can be extracted and only the general outlines of the skull vertebrae and femora traced. A single trench of several opened, 2£ feet square, showed twelve of these skeletons on its four faces. White's sketches show that the Indians of the region kept their dead in huts, where they were exposed to smoke, as was the case in Florida, and it is interesting to have this corrobo- rative proof that in addition the bodies were packed closely together and heaped about with sand. It is also an interest- ing circumstance that the Alder Branch, at an early date and one apparently anterior to white occupation, had been cut to a straight course and the earth heaped on its southern bank. Similar artificial waterways are to be found in east Florida. This mound probably marks the neighborhood of the Indian village found by Barlowe, as the Alder Branch makes a natural boat entrance to anyone approaching as he did from the main- land opposite. Besides these surface remains, there are on the northeastern shores of the island, where careless denuda- tion of forest has set the sand in motion, two earlier horizons of Indian occupation, one 8 and the other about lo feet below 1 This is upon the laud of Charles rvttigrew Meekiu, near the "Indian hole," a large artificial cavity, 20 yards across aud 30 feet deep, mentioned in deeds for many years and an early landmark. Neither this nor the mounds can he due to operations in the war. 56 AMERICAN HISTORICAL ASSOCIATION. the present surface, the latter clearly marked and extending for a mile or two, with frequent fire pits and pottery fragments of the fish-net type. When the colonists landed under Grenville and under Lane they came from the north, as Barlowe had. They would natu- rally enter at the same creek and push their boats up the mile or more which it was navigable down to the memory of men now living, 1 shipbuilding on a small scale having been carried on at its head in the last century, as the oak chips, blocks, etc., which can be turned up show, although the stream is now a dense marsh of* tall reed. They sought for their new home the highest ground on the island, upon which stand such remains as are left, a site carefully designated, it may be noted, by White by. a mark O, distinct from that used to indicate the Indian villages on the island. That this was the approach to the colony, and not by Shal- low Bag Bay, as the traditional oak tree would have indicated, appears from Lane's account of his return to the island. Com- ing from the sea side, he first "espied toward the north end of the island the light of a great fire." Landing at daybreak, he " went through the woods to that part of the island directly over against Das am on guepeuk " — that is, the western or Croatan Sound side — and "from thence we returned by the water side round about the north part of the island until we came to the place where I left our colony in 1586," which would be the nearest approach from the eastern side of Boanoke Island to the existing fort. It was here on the "sandy bank" that he found the tree "in the very brow of which were curiously carved these fair Roman letters, 0. B. O." From the fort he "went along by the water side to the point of the creek," which is more likely to be the Alder Branch, half a mile off, than the creek of Shallow Bog Bay, 3 miles distant. "Presently," con- tinues Lane, "Captain Cooke and I went to the place, which was in the end of an old trench, made two years past by Captain Amadas." "Two years past" must be a misprint for six years past, when Captain Amadas visited the place in 1584; and as it is difficult to see what digging he coald have done on his flying visit, it does not appear a forced construction to take "made" in the seafaring sense of "found," and the "old trench," the canalled stream of which I have already spoken. 1 1 owe much in these details to the kindly interest and the local knowl- edge of Mr. Walter Dough, long the owner of the fort site. SURROUNDINGS AND SITE OF RALEIGH'S COLONY. 57 Into the details of Lane's melancholy visit, whose pathos must have touched every reader, I do not enter, because I propose to confine myself to the topographical aspects of the history. On the site itself, while the colonists were "left in sundry houses" originally built by Grenville's colonists, Lane found "the houses taken down and the place very strongly inclosed with a high j)alisade of great trees, with curtains and Hankers very fortlike." It is this for which the low, square mound, still preserved, now stands. Few sites are better established by tradition. In 1654 Travis Yardley records the visit of "a trader for beavers," in September, 1653, to Koanoke Island, where he was shown "the ruins of Sir Walter Baleigh's fort" by friendly Indians. 1 The island was bought from the Indians by Yardley, and in 1676 became the property of a New Englander. A gap of a century leaves it without record. The local tradition runs back clearly authenticated to the middle of the last century, and there were then living those who could by one or two removes have heard the Indian tra- dition noted by Yardley. When I visited the site in November, 1887, I could find no record of any description since that made in 1860 by Mr. Edward 0. Bruce, to whose article allusion has already been made. Judging from his account, it has seen few changes in thirty-five years, though all the brick and mortar he mentions is gone. It is a quadrangular embankment whose angles lie due north and south and east and west, so that the faces front southeast, northeast, northwest, and southwest. The mound, which is perfectly clear around the entire inclosure, is 2 feet 4 inches high above the ditch at its most prominent point. The eastern angle has a slope of 23 feet on the angles and about 15 feet on the curtains, and is broken by what was apparently a sally port crossing the southwest angle, the one turned toward the creek already mentioned. The four faces measure: The southeastern, S4.3 feet; the southwestern, 77.6 feet; the north- western, 63.3 feet, and the northeastern, 73.9 feet. As the mound is irregular, these measurements are necessarily ap- proximate. By measuring from points on the irregular slope farther in or farther out, different dimensions would be secured, but it was probably originally a square of 25 yards. The eastern angle is a right angle, without any signs of a bastion whatever. Each face is broken by an angle about 15 1 North Carolina Colonial Records, 1 : 18. 58 AMERICAN HISTORICAL ASSOCIATION. to 18 feet across and projecting from the embankment line about 5 to 7 feet. The southern, western, and northern angles are bastioned. These may have anticipated the pentagonal bastion of a century later ; but this is extremely improbable of Elizabethan fort builders, more familiar with the earlier roundel, better suited for the trajectory and angle of the pro- jectiles of the period. There is nothing in the bastions themselves to show that they were pentagons, unless one reads into them a preconcep- tion based on our familiarity with this form. The interior is nearly on a level with the embankment, but at points slightly lower. The oak mentioned by Mr. Bruce still stands, though aging. The other trees are more recent, and none are of any great age. The most serious challenge which must be addressed to this ancient relic is its size. An area of 625 square yards is scant space for over 100 souls who composed the beleaguered colony for which it was built. As the houses had been taken down, it was their only dwelling, and while it is not impossible that it would hold them in leaguer, one would expect the fort would be larger. It is also rigorously fair to add that the remains have the look, slope, and appearance of smaller Indian mounds, some of which are quadrangular and are laid with reference to the four cardinal points. If this embankment were in an In- dian mound region, with no other history, it would probably be given this origin ; but with the chain of evidence which exists, broken though it be by the gap of a hundred years, there appears to be no reason for challenging its assigned source. So far as is known, the surface has been disturbed only once prior to the excavations just conducted. During the occupa- tion of the island by Federal soldiers in 1863 holes were dug in the embankment at the eastern angle and on the south- eastern face. On complaint by Mr. Walter Dough, who then owned the fort, the vandalism was promptly checked and the fort placed under military guard. It was probably at this time that the hatchet mentioned by Mr. Weeks in his paper was found. With the exception of the Indian pottery and the small iron fragments just discovered, this is the only object yet found in the inclosure. As a careful examination of the site seemed desirable, I made application to its present owner, the Roanoke Memorial Asso- ciation, and from its president, Maj. Graham Daves, and its SURROUNDINGS AND SITE OF RALEIGH'S COLONY. 59 secretary, Dr. J. B. Bassett received prompt arid cordial x>er- missiou to conduct excavations. I was careful to avoid any disturbance of the embankment and its slope, tlie surface dis- turbed was carefully returned to its original condition, the site of each trench was carefully plotted and fixed by bearings and measurements, and a minute record kept and deposited with the association, so that no injury would be done to the site and no embarrassment caused to any future explorer by his inability to know where the soil was disturbed. In. all, 13 trenches, most of them 5 by 3 feet, were opened and carried from 4 to 9 feet deep. Water, it may be premised, is reached at 15 feet, and undis- turbed sand at about 4 feet. Wherever trenches were sunk, and, it is fair to conclude, over the entire area, there was found a thin and undisturbed layer of sandy humus of 6 to 8 inches to a foot, then a layer of black, ashy earth, containing many frag- ments of charcoal and frequent fire pits. This layer rested directly on undisturbed sand, often penetrated by fire pits. If we imagine a forest surface from which the original humus had been removed to make an embankment, laying bare the sand below, this site occupied for a season and then for three centuries left to gather humus again, the condition revealed would be created. Toward the base of the black, ashy layer were found small pieces of iron, a corroded nail, a chipped piece of quartzite, and some small fragments of Indian pottery, networked. No one could reasonably expect to find any objects of importance on a site ransacked as this must have been, but I confess my surprise at the absence of small frag- ments, particularly of pottery. For a site occupied as it was, the place proved singularly barren of debris. Like its size, this circumstance has no ready explanation. The trenches opened were dug in three angles, the eastern, northern, and western — the southern being too much occupied by trees — across the center in two of the flanking bastions, and at other points where the surface was either above or below the normal level. In addition, the embankment was sounded with an iron rod for a depth of from 3 to 4 feet at intervals of from 10 to 20 feet around the inclosure. The embankment may have had logs in it which have wholly decayed, but the indications were that it was heaped sand, the dark ashy layer curving over its slopes. Excavations were also made in the ditch and at 60 AMERICAN HISTORICAL ASSOCIATION. various poiuts in the woods, showing there an undisturbed surface and no remains of a layer of coal and ashes below the surface. The most plausible deduction which can be made from these sparse results is that the site was occupied at an early period by those using iron, succeeded by many years in which the forest did its natural work of making soil. As a corroboration of the tradition in regard to the site, this conclusion is important. In any other view the fruits were meager ; but the fortune of excavation — of all pursuits of chance the most baffling and the most absorbing — may richly reward some successor with more time than the brief days I could devote. It is at least a profound satisfaction, for which I am most grateful to the officers of the association, to have had the jmvilege of devoting a short vacation to increasing the scanty knowledge previously recorded in regard to the earliest site associated with the history of men of our race and tongue on this continent. The site, as already remarked, is now the property of the Roanoke Colony Memorial Association. With this last chap- ter in its history, there rests the same melancholy associations as with all before, the founder of the association, Edward Graham Daves, late of Johns Hopkins University, having died within a year of its organization. In November, 1887, after my first visit to the site, when I made a compass survey of the mound and a hurried investigation of its surroundings, I stopped at Johns Hopkins University on my return, where my account of the remarkable preservation of the old fort excited the interest of both Dr. Herbert Baxter Adams, the secretary of the American Historical Association, and of Dr. Daves. The possibility of purchasing the site was discussed at the time, but no active steps were taken until March 25, 1893, when a call l for enough money to buy the fort and a farm of 250 acres on the northern end of the island was issued by Dr. Daves, a native of eastern North Carolina, to whose j>ersonal enthusi- asm as an historical student was altogether due the acquisi- tion of the site, the organization of the association, and the preservation by it of the earliest English remains on the con- tinent. The modest sum needed, $1,250, was raised before the end of the year, a large portion being the returns of author's 'This call was signed by Edward Graham Daves, Francis White, Wil- liam Shepard Bryan, A. Marshall Elliott, Bartlett S. Johnston, and Thomas J. Boykin. SURROUNDINGS AND SITE OP RALEIGH'S COLONY. 61 readings by Dr. S. Weir Mitchell of his poem, " Francis Drake," at Bar Harbor, Baltimore, Philadelphia, the residence of Mr. Frank Thomson at Marion, Pa., and elsewhere. May 22, 1893, the Boanoke Colony Memorial Association was organized at Baltimore, and the first meeting- of its stockholders was held at Raleigh October 23, 1894, a second meeting having been held last October. The present officers of the association are : Presi- dent, Maj. Graham Daves, of Kewbern; vice-president, Mr. W.D. Prndeu, Edenton, S. C, and secretary and treasurer, Dr! John Spencer Bassett, of Trinity College, Durham, M". C. The association now owns the site, with 10 acres, and a farm of 230 acres covering the northern end of the island. The associa- tion proposes to fence and preserve the site, erect a monument upon it, and draw public attention to its history. Contribu- tions for this purpose are urgently needed. With the association and its work, the history of the site closes. By little short of a miracle of accident this crumbling mound, "child of silence and slow time," has escaped destruc- tion. The elements have spared it on an island where the merest exposure of the loose, thin soil starts shifting sands rj pile dunes and level them. The plow has never passed over its low walls and it has escaped the ravages of the relic hunter. Even the war found officers who appreciated its value and guarded its outlines. A just local pride has shared in its pres- ervation, and the first sod turned by English hands in the Americas stands to-day after three centuries more clearly marked than many a later site and more ambitious structure. The low mound, scarce higher than a grave, will rear its round outline for long years to come. The beginning of the birth of a great people, it is impossible to forget that it was also the sepul- cher of the hopes, the fortune, and the future of Walter Baleigh, brightest blossom of our English renaissance. About this low heap centered once the plans of a kingdom, the promise of a principality, and the prospect of enduring fame. -the lion and the lizard keep The courts where Janishyd gloried and drank deep : And Bahrain, that great hunter— the wild ass Stamps o'er his head but can not break his sleep." \ M01 LIBRARY OF CONGRESS 014 444 015 8 Hollinger pH 83 Mill Run F05-2193