Fort RaleighArchaelogicalProject

1994/1995Survey ReportNicholas M. Luccketti

VirginiaCompany

Foundation

December 1996

Association for the Preservation

of Virginia Antiquities

Graphics by Jamie E. MayDesign and Production by Elliott Jordan

© 1997 by The Association for the Preservation of Virginia Antiquities and Nicholas M. Luccketti

Introduction .................................................. 1Context ......................................................... 5Historical Background ....................................................................... 5Archaeological Background............................................................... 6

Research Design ............................................ 81994 Objectives ................................................................................ 81995 Objectives .............................................................................. 101994/1995 Methods ....................................................................... 111994/1995 Expected Results ........................................................... 12

Fieldwork Results ......................................... 151994 Area Excavation ..................................................................... 151994 Survey Test Pits ...................................................................... 191995 Shovel Test Hole Survey ......................................................... 201995 Test Pits .................................................................................. 21Soil Analysis .................................................................................... 22

Summary..................................................... 23Conclusions ................................................. 25Future Research ........................................... 28Bibliography ................................................ 29Appendix A ................................................ 31Finds List ......................................................................................... 31

Contents

1

The Virginia Company Foundation (VCF)is a non-profit organization devoted to the archae-ology and research of sites dating to the first yearsof English settlement of Virginia. It has been con-ducting archaeological research at Fort RaleighNational Historic Site, Roanoke Island, NorthCarolina, since 1991 (Figure 1). The first two yearsfocused on the excavation of a metallurgical anddistilling area, a “science center,” located adjacentto the west side of the reconstructed earthen fortand associated with the 1585-86 colony of RalphLane. This work is the subject of a separate reportauthored by Ivor Noel Hume that will be submit-ted to the National Park Service (NPS) in early 1996.

In 1994, the VCF commenced three area ex-cavations in the vicinity of the reconstructedearthen fort to reexamine features that were firstrecorded by J.C. Harrington during his survey ofthe park in 1947-48 and possibly related to En-glish occupation from 1585-87. The VCF alsoconducted limited testing away from the immedi-ate vicinity of the reconstructed earthen fort in1994. No European artifacts or features associatedwith the 16th-century English settlements atRoanoke Island were found. The survey did locatea deeply buried layer containing Native AmericanColington ceramics in the heavily wooded area be-tween the earthen fort and the Elizabethan Gar-dens which had never been previously tested.

The 1995 field season consisted of a shoveltest survey supplemented by the excavation of sev-eral larger test units to define the cultural layer dis-covered in 1994 and to determine if it also con-tained evidence of 16th-century Englishsettlement. Although no features were found, thesurvey did recover European artifacts dating to the16th century from this area.

The Project Director for the 1994-95 FortRaleigh Archaeological Project was VCF presidentDr. William M. Kelso, who provided general direc-tion for the fieldwork and report production.Nicholas M. Luccketti was the project archaeolo-gist supervising the fieldwork. Ivor Noel Hume,who was Project Director for the 1991-93 sciencecenter excavations, was an advisor to the project.Dr. David Phelps, Director of the Coastal Archae-

ology Office of the Institute for Historical andCultural Research at East Carolina University, gen-erously served as the prehistory consultant for the1994-95 projects. He identified all Native Ameri-can artifacts recovered during the excavations andmade three on site inspections, one in 1994 andtwo in 1995.

The 1994 field crew consisted of VCF per-sonnel David K. Hazzard, William M. Kelso, EricKlingelhofer, William L. Leigh, III, Nicholas M.Luccketti, Jamie May, Alastair Macdonald, IvorNoel Hume, Nate Smith, Martha Williams andNPS archaeologist Ken Wild. Members of the1995 field crew were David K. Hazzard, NicholasM. Luccketti, Alastair Macdonald, Martha Will-iams, Gail Wharton, and Robert Wharton withassistance from Dr. Paul Green, Thomas MacDonald,and Dane Magoon. Artifacts collected during thearea excavations and survey were processed byVCF curator Beverly A. Straube. She also identi-fied the European objects and edited the final re-port in her usual thoughtful manner. Dr. GeraldJohnson, Professor of Geology at the College ofWilliam and Mary, amiably provided his expertise,as he has done so many times in the past, regardingsoils of the project area and Dr. Stanley Riggs, Dis-tinguished Professor of Geology at East CarolinaUniversity kindly identified geological specimens.

Funds to support the 1994 fieldwork wereprovided in large part by the Fort Raleigh NationalHistoric Site thanks to the interest of then Super-intendent Tom Hartmann and Cultural ResourceManager Mary Collier. A significant financial con-tribution to the 1994 project was made by theRoanoke Colonies Research Office of the Institutefor Historical and Cultural Research at East Caro-lina University through the efforts of Dr. KeatsSparrow, Dean of Arts and Sciences. Financial sup-port in 1995 was drastically diminished to only aVCF grant for staff lodging which unfortunatelyreduced the normal fieldwork season by half. Ac-cordingly, a particular debt of gratitude is owed tothe 1995 field crew whose commitment to archae-ology at Fort Raleigh is such that they were not de-terred by the lack of compensation for either theirtime or expenses.

Introduction

2

The VCF/NPS archaeological work was amajor component of “The Riddle of Roanoke,” atelevision program produced by New DominionPictures for the Archaeology series on the Discov-ery channel and included footage of the 1994fieldwork as well as an interview with Ivor NoelHume.

Of course, the VCF Fort Raleigh Archaeo-logical Project would not be possible without thesupport and cooperation of Fort Raleigh NationalHistoric Site and the Southeast ArchaeologicalCenter of the National Park Service. Superinten-dents Tom Hartman, now retired, and Russ Berry,Assistant Superintendent Mary Collier, Supervi-sory Park Ranger Andrew Kling, and CuratorsBeBe Woody, now retired, and Steve Harrison haveall made substantial contributions to the ongoingresearch, ranging from their official managementduties to their continuing encouragement and will-ingness to help in whatever manner possible. Aspecial acknowledgment is due NPS Civil Engi-neer Charley Snow, whose unflagging assistancethrough the years, whether it was surveying, re-moving sidewalks, backfilling or any number ofother tiresome requests made of him and his staff,has been of immeasurable help during the course

of the Fort Raleigh Archaeological Project. TheNational Park Service Southeast ArchaeologicalCenter (SEAC) regional archaeologist BennieKeel, who participated in the science center exca-vations, provided much needed advice, whileSEAC archaeologist Ken Wild was an extremelyvaluable member of the 1995 team. The author isindebted to the Association for the Preservation ofVirginia Antiquities (APVA) for graciously allow-ing this report to be produced at its archaeologicalfacility at Jamestown Island and particularly to theAPVA Laboratory Manager Elliott Jordan whoskillfully designed and produced this report. Theconfluence of the VCF mission and the APVA’s de-cision to expand its archaeological program re-sulted in a formal consolidation of the two organi-zations in 1996. Finally, the author is most gratefulto Ivor Noel Hume, William M. Kelso, and DavidK. Hazzard for allowing me to continually pesterthem for their insights on the archaeology at FortRaleigh.

The 1994 fieldwork was conducted fromOctober 15-October 29 and the 1995 fieldworkfrom November 6-November 12. All errors are theresponsibility of the author.

3

Figure 1: Roanoke Island and vicinity (Harrington 1962).

4

5

Historical BackgroundThe history of the English settlements at

Roanoke Island has been researched in great detailand extensively published, therefore it is unneces-sary to provide a shallow repetition of this materialin this report. Nevertheless, it is important forevaluating archaeological investigations to brieflydescribe those events that may have left archaeo-logical remains at Roanoke Island in the 16th cen-tury.

In 1584, Sir Walter Raleigh received a char-ter from Queen Elizabeth to colonize part of NorthAmerica in hopes of establishing an English Em-pire there. A reconnaissance voyage to NorthCarolina that same year returned with a promisingreport of the country, which was named Virginiaafter Elizabeth, the Virgin Queen. Raleigh’s firstcolony, consisting of 108 men under the leadershipof Sir Richard Grenville and Ralph Lane, landed atthe north end of Roanoke Island in July of 1585. Aletter written by Ralph lane to Richard Hackluytwas addressed “From the Newe Forte in Verginia, the8th daye of September:1585” (Quinn 1991:214).The records also suggest that the colonists built atown separate from the fort (Noel Hume1994:37). In addition to exploring the surround-ing country, observations and experiments wereconducted by scientist Thomas Hariot and byJoachim Gans, a metallurgist or ‘mineral man’ ashe was called at the time. Archaeological research,discussed below, indicates that a scientific labora-tory for Hariot’s and Gans’ work was set up apartfrom both the town and fort. When supply shipscommanded by Sir Richard Grenville, who had re-turned to England in August of 1585, were late inreturning to the colony in 1586, the discouragedcolonists abandoned Roanoke Island with SirFrancis Drake who had arrived from a Caribbeanexpedition in June of that year.

The second colony is a consequence of thelate arrival of Grenville’s resupply. In August of1586, Grenville returned to England after finding

the settlement abandoned, however, he left a garri-son of 15 men on Roanoke Island to hold it for theQueen. A captured Spanish pilot, Pedro Diaz, al-though he was not allowed to leave the ship, re-ported that the settlement had “…a wooden fort oflittle strength…” (Quinn 1991:790).

Raleigh dispatched another colony toRoanoke Island in 1587. The 117 men, women,and children arrived to find the 1585 fort ruined.The old houses were repaired and new cottageswere built; all protected by a new fort apparentlyconstructed of posts set side-by-side in a trench.Governor John White, apparently an ineffectiveleader, was persuaded to return to England for sup-plies, leaving on Roanoke Island his daughterEleanor and her husband Ananais Dare, the par-ents of the fabled Virginia Dare, the first Englishchild born in Virginia. White’s return to RoanokeIsland was delayed due to war between Englandand Spain. When he arrived back on Roanoke Is-land in 1590, White, of course, found no one onthere. He did find, in his words

…the place very strongly enclosed with a highpalisado of great trees, with curtains andflankers very Fort-like, and one of the chiefetrees or posts at the right side of the entrancehad the barke taken off, and 5 foote from theground in faire Capitall letters was gravenCroatoan without any crosse of distresse, thisdone, we entered into the palisado, where wefound many barres of Iron, two pigges of lead,four iron fowlers, Iron saker shot, and such likeheavie things, thrown here and there, almostovergrown with weeds and grass (Quinn1991:614).

Despite several attempts to locate the miss-ing settlers, they were never found and became thelegendary “Lost Colony.” A small earthen fort,known as Fort Raleigh, is located at the site, butwho built it, and when, is uncertain.

Context

6

ArchaeologicalBackground

The archaeological history of the search forthe Cittie of Raleigh began in earnest in 1947 and1948 when J.C. Harrington excavated a series oftest trenches at Fort Raleigh National Historic Site(Figure 2), and later in 1953 on the adjoiningElizabethan Garden property. Harrington exca-vated at least 38 separate five-foot wide trenches(depending upon how you count them) for a totallength of 3,320 linear feet. Harrington found only18 artifacts that could be attributed to the 16th-century and concluded that the habitation, ortown, site was elsewhere. The remains of theearthen fort were excavated in 1950 (Harrington1962). Later in 1965, Harrington excavated whathe termed an outwork: a 9-ft. square structure con-structed of post or taproot molds and log stainsjust outside the entrance of the reconstructedearthen fort (Harrington 1966). Within the struc-ture were three pits, one of which contained brick,tile, burned clay, Indian pottery, and sherds of aNormandy flask.

Fort Raleigh National Historic Site’s PhillipEvans, a dedicated student of the Roanoke settle-ments, noticed a similarity between Harrington’soutwork and flankers of a c.1619 wooden fort ex-cavated by Noel Hume at Martin’s Hundred inJames City County, Virginia (Noel Hume 1982).Evans suggested that perhaps the outwork was abastion for a wooden fort that preceded the con-struction of the earthen fort. This suppositionprompted a survey by NPS archaeologists in 1982.John Ehrenhard, William Athens, and GregoryKomara examined the area west of the recon-structed fort using proton magnetometry, soil re-sistivity, infrared photography, and soil chemicaltests. Subsequent field testing of selected anomalieswere conducted in the fall of 1982 and in 1983.Only one of these areas, near the entrance of thefort and Harrington’s outwork, produced 16th-century artifacts that eventually proved to besherds of crucibles, flasks, and English or Nether-landish tin-enameled ointment pots. No structuralfeatures related to houses or fortifications werepositively identified (Ehrenhard, Athens, andKomara 1983; Ehrenhard and Komara 1984). Thenature of the outwork remained unresolved. Alsoin 1982, Phillip Evans discovered a barrel and a

Figure 2: Location of test trenches excavated 1947-1948 (Harrington 1962).

7

hollow log in the shallows of Roanoke Soundabout 600 ft. north of the reconstructed earthenfort. Both objects had a C-14 date range that en-compassed the 16th century and they may havebeen the bottoms of wells (Noel Hume 1994:88-89).

The Virginia Company Foundation initiatedthe Fort Raleigh Archaeological Project in 1991under the guidance of Project Director Ivor NoelHume and President Dr. William M. Kelso. NoelHume’s efforts to puzzle out the nature of the firstforts at Jamestown included reviewing earlier En-glish fortifications which, in turn, led him back tothe Roanoke Island earthwork, Harrington’s out-work, and Phil Evans observation on the resem-blance of the outwork to the Wolstenholme Towneflankers.

Test trenches dug in the spring of 1991 relo-cated the area of Harrington’s outwork and of the1983 NPS excavation area that yielded numerous16th-century ceramics. In the fall of 1991 and1992, a team of 9 VCF archaeologists workingwith SEAC archaeologist Bennie Keel, 6 assistants,and several NPS rangers and interpreters volun-teering as soil sifters, opened up an area west of thereconstructed earthen fort entrance. The purposewas to determine whether the appendages extend-ing off of the Harrington outwork were in fact sec-tions of palisaded walls of a wooden fort. The exca-

vations found no evidence to suggest that the out-work was a flanker to a wooden fort, but instead,discovered the partially preserved remains of a sci-entific workshop floor. More than 100 artifactswere found, consisting principally of cruciblesherds, Normandy flask sherds, fragments ofchemical glassware, worked and unworked copper,antimony (an essential ingredient in separating sil-ver from copper), and sherds of delftware ointmentpots. These all point to the distilling experimentsof Thomas Hariot and the metallurgical work ofJoachim Gans who were members of only the LaneColony. The relationship of the earthen fort to thescientific workshop is still a matter of debate; oneinterpretation is that the earthen fort was con-structed to protect the workshop while a compet-ing argument believes that the earthen fort waserected on top of, and therefore, after, the work-shop area was deserted. A detailed field report onthis work was submitted to the NPS in 1996,though much of this research is discussed in TheVirginia Adventure, From Roanoke to JamesTowne: An Archaeological Odyssey (Noel Hume1994).

Despite this remarkable discovery, no re-mains of the people, houses, or forts of either the1585 Lane Colony or John White’s 1587 “LostColony” have yet been found.

8

1994 ObjectivesOne objective of the 1994 survey was to re-

examine several 1947-48 Harrington test trenchesthat contained features that might be postholes.Test excavations in these locations endeavored todetermine if the features recorded by Harringtonwere postholes, the date of the postholes, and whatkind of structure(s), if any, they represented.

One Harrington trench extended northfrom the east bulwark of the earthen fort (Figure3). The presence of possible postholes in this areawas explored in light of the following theory. The“science center,” with its furnaces containing po-tentially dangerous fires and offensive fumes,would have been located outside, but not far fromthe village. The barrel well found on the beach is aremnant of the severely eroded village site andthus, the direction of settlement is north from the“science center” toward the barrel well. Accord-ingly, it is plausible that if the features in theHarrington trench were actually postholes, theymay represent peripheral village structures.

A Harrington trench that contained a linearsoil stain and located approximately 100 ft. west ofthe earthen fort’s entrance was also reevaluated(Figure 4). Within the past decade, excavations atfour sites in Virginia, including most recentlyJamestown Island, have uncovered evidence ofwooden fortifications. The Nansemond Fort site,the Nicolas Martiau Fort, the Middle Plantationpalisade (Kelso, Luccketti, and Straube 1990), andthe Jamestown Island first settlement site all pro-duced unmistakable archaeological footprints ofpalisade fortifications that were constructed by set-ting small posts or trees and split rails into a nar-row slot trench. These features closely resemble thelinear soil stain unearthed by Harrington (Kelso,Luccketti, and Straube 1995). If the Fort Raleighfeature is indeed a palisade slot trench, it wouldhave major implications regarding the theory thatthe earthen fort with is western entrance, is a bul-wark for a much larger wooden fort extending tothe west.

Research Design

N

possible postholes

ReconstructedEarthwork

Figure 3: Plan of Harrington’s excavations showing possible postholes north of fort (FRNHS).

9

NReconstructedEarthwork

“outwork”

Vicinity of 1585-1586scientific workshop

Possible slot trench

Possible postholes

Figure 4: Plan of Harrington’s excavations showing features west of the reconstructed earthwork (FRNHS).

10

A third Harrington trench, about 175 ft.northwest of the earthen fort entrance, was se-lected for investigation because it had four post-hole-like features in a line suggestive of an earth-fast structure (see Figure 4). It also yielded severalof the few 16th-century artifacts that were foundby Harringon during his testing.

The second objective of the 1994 survey wasto determine whether there was any archaeologicalevidence of 16th-century settlement between thefort and the location of the previously mentionederoded barrel well(s) (Figure 5). A transect of 5-ft.squares was excavated north from the earthen forttoward the base of the sand dunes. The goal was todetermine whether there are any 16th-century arti-facts, undisturbed strata, or features present withinthis area.

Lastly, the bulk of the previous archaeologi-cal work at Fort Raleigh has concentrated on theimmediate environs of the reconstructed earthenfort with little exploration of the heavily woodedland toward the Elizabethan Gardens (see Figure2). Selected test units were excavated in this latterarea in 1994.

1995 ObjectivesThe prime objective of the 1995 survey was

to survey the wooded area between the Elizabethan

Garden and the earthen fort to determine the ex-tent and nature of the site identified by the 1994survey, and particularly if it contained any evi-dence of 16th-century English occupation.Colington Plain and Simple Stamped ceramicswere recovered from a buried organic stratum in a1994 test square in this area, about 150 ft. fromRoanoke Sound and 450 ft. west of the earthenfort, indicating that this site dated to the protohis-toric period. The stratum appears to be identical toa buried A soil horizon located by Bennie Keelwhile monitoring the excavation of sewage pipetrenches in the septic field at Fort Raleigh. Dr.Keel recovered three sherds of Colington SimpleStamped pottery from the A horizon, which heidentified as a dune-covered ground surface datingto the late prehistoric-early historic period. Noshell, bone, midden debris, or features were found(Keel 1991).

It is possible that this area, apparently firstinhabited by Native Americans, was reoccupied insome part by the Roanoke colonists, who allegedlysettled not far from the Indian village at the northend of Roanoke Island, which conventional wis-dom places at Dough’s Point. By means of a sys-tematic shovel test survey and excavation of se-lected larger test units, the project endeavored todetermine whether there were any 16th-centuryEuropean artifacts and/or features present.

Figure 5: Location of barrel well(s) along north shore of Roanoke Island (FRNHS).

11

1994/1995 MethodsThe 1994/1995 projects employed the same

grid that was previously established for the 1991through 1993 VCF projects with one modifica-tion. The previous VCF grid was composed of 50ft. blocks and each block was subdivided into 1610-ft. squares separated by alternate 2 ft. and 3ft.wide balks. The revised grid contains the same da-tum points, baselines, and blocks, only the blocksare now comprised of 25 10-ft. squares with nopredesigned balks. All shovel test holes and testunits were assigned an individual Fort Raleigh Ex-cavation Register (FRER) number, continuingwith the sequence used in 1991-1993, and a 10-ft. square was thelargest test unit to receive anFRER number. All test units weremapped on a site plan at a scale of1/4"=1'. All shovel test holes weredocumented on a shovel test regis-ter form while each test unit wasdocumented on a VCF excavationregister form. A plan and profilefor each test unit were drawn at ascale of 1/2"=1'. All features weremapped. Black-and-white photo-graphs and color slides were takento record features and profiles ofeach test unit. Elevations of eachtest unit were taken in reference tothe FDR monument as in the1991-1993 projects. The locationof test units and elevations was re-corded in 1995 field season using aNPS total station.

All shovel test holes and testunits were excavated followingnatural stratigraphy. All soil fromeach shovel test hole and test unitwas screened through 1/4 in. mesh

and all artifacts were collected. Soil samples weretaken for chemical analysis from strata that hadcultural significance. While the principal goal ofthe 1994/95 projects was survey, the area excava-tions included testing of features that might havebeen 16th century. Paleoethnobotanical, pollen,and phytolith samples were collected if the ProjectArchaeologist in consultation with the NPS ar-chaeologist determine that such sampling was war-ranted. If present, C-14 samples from features orundisturbed strata were taken. A lacquer peel of se-lected stratigraphic profiles was prepared for analy-sis under laboratory conditions.

All artifacts recovered during the survey werewashed, labeled with their respective FRER num-

Figure 6: Plan of earthfaststructures from c. 1618 Mainesite (Outlaw 1990).

12

ber, identified, and stored in archival bags andboxes. All artifacts were catalogued on the NPSAutomated National Cataloguing System and thisinformation will be provided to the NPS in hardcopy in addition to excavation register records,original site plans and profiles, color slides andnegatives. Prior to backfilling all the shovel testholes and test units were lined with black plastic. Apenny dated either 1994 or 1995 in a film canisteralso was placed into each backfilled shovel test holeand test unit.

1994/1995 ExpectedResults

It was anticipated that any surviving remainsof either the 1585/86 or 1587 settlements willconsist of earthfast structures, possibly irregular

Earthenware, EnglishNorth DevonBorderwareDelftwareEast AngliaStaffordshire blackwareMidlands purple butter pot

Earthenware, EuropeanDutch coarsewareDutch delftwareItalian majolicaBerettinoMontelupoSaintongeSpanish majolicaGerman refractory wares

and shallow postmold patterns similar to thosefound at the c.1618-1625 Maine site (Figure 6)just outside Jamestown (Outlaw 1990). Thewooden fortifications of the two settlements mayresemble the post-rail-plank fort excavated by IvorNoel Hume at Martins Hundred (Noel Hume1982) or possibly a palisade of small posts and splitrails set into a slot trench similar to those of theearly 17th-century forts excavated by the VCF atthe Nansemond Fort site in Suffolk, Virginia andYorktown, Virginia (Kelso, Luccketti, and Straube1990).

Archaeological research on early 17th-cen-tury sites in Virginia, including a current project atJamestown to search for the first fort and settle-ment, has resulted in a knowledge of artifact as-semblages, particularly ceramics, that might befound at either of the two Roanoke settlements.The ceramic types that might be associated withthe settlements include:

Earthenware, IberianMeridaSevillian olive jar

Slipware, EuropeanDutch

Stoneware, EuropeanBeauvaisFrechenRaerenMartincamp costrel

Chinese Porcelain

13

Figure 7: Location of 1994 excavations.

14

Figure 8: Location of 1995 test pits.

15

The 1994 fieldwork consisted of three areaexcavations and nine survey test pits (Figure 7).The 1995 survey included the excavation of 14shovel test holes, 3 small test pits, and 4 large testpits (Figure 8).

1994 Area Excavation

East Excavation Area:Harrington Trench, North ofthe Reconstructed Fort

Since the precise location of this Harringtontrench was unknown, the 1994 excavation beganwith a series of continuous 3 ft. wide trenches run-ning west to east to intersect the Harringtontrench. The first two 10-ft. trenches (FRER 63 and64) were unsuccessful. The third 10-ft. trench waspassed over due to the presence of a large tree, andthe fourth 10-ft. trench (FRER66) also failed tofind the Harrington trench. Consequently, the

skipped third trench (FRER65) was excavated and,as luck would have it, the Harrington trench wasfound in it.

Two tree root stains/possible posthole fea-tures were exposed during the excavation of the40-ft. long test trench and in the relocated Har-rington trench, so the excavation was enlarged todetermine if there were more features present andif they made any pattern (Figure 9). The 3-x-10-ft.test trench FRER65 was expanded into an 8-x-10-ft. excavation unit, still avoiding the large tree,while the adjoining test trench FRER66 was en-larged to a 10-ft. square. Both altered test units re-tained the FRER number assigned to their originaltest trench, respectively. A new 5-x-10-ft. unit(FRER67) was opened to uncover more of theHarrington trench.

The stratigraphy was consistent throughoutthe East Excavation Area: a 3-5 in. thick upperlayer of sod, root mat, and loose grey sand, a 6-9in. thick layer of compact light grey sand thatsealed a bottom layer of 2-3 in. of mottled brownsand. Several treehole/possible posthole features

Fieldwork Results

Figure 9: Plan of EastExcavation area.

16

were uncovered all having the same general ap-pearance: amorphous shapes with very indistinctlines and a large core of black sand with a halo verylight grey-to-whitish sand. Two such features,FRER65F and G, were sectioned, revealing irregu-lar and tapering profiles and changing soil colors,

all of which suggest that the features are not man-made (Figures 10 and 11).

No artifacts were found in either the strataor features, further evidence of the lack of culturalactivity in this area.

Figure 19: Section of FRER 65G, facing south. This feature was intentionally overdug to examine fill and subsoil.

Figure 11: Section of FRER 65F, facing west.

17

Central Excavation Area:Harrington Trench with Ditch

The south half of square FRER72 and northhalf of square FRER73 were first excavated in anunsuccessful effort to relocate the Harringtontrench in this area. Subsequent excavation of thenorth half of FRER72 did find the Harringtontrench, which was further exposed by excavatingthe southern halves of squares FRER74 and 75.The narrow linear soil stain or narrow trench wasfound in FRER75, though apparently it had beencompletely excavated within the Harringtontrench. Several other features were uncovered in-cluding 2 tree root stains/possible posthole featuresresembling those found in the East ExcavationArea, 2 small round postmold-like features, halvesof two larger features with similar fill, and a largetreehole or small pit (Figure 12). The only featurestested were the linear soil stain (FRER75D) andthe treehole/pit (FRER73D,E).

The stratigraphy across the Central Excava-tion Area generally corresponded to that found inthe East Excavation Area. The humus and root matwas about 3 in. thick with an underlying 2-3 in.thick stratum of dark grey sand that sealed a layerof mottled sand.

The linear soil stain was trowel-cleaned todiscern if it contained any postmolds set side-by-side that would be characteristic of a palisade, andnone were apparent (Figure 14). A test cut of thefeature revealed that it was very shallow, the bot-tom of the trench was only 8 in. below moderngrade and did not even penetrate into the subsoil.The trench did have a well-defined contour withnearly vertical sides and a smooth bottom. No arti-facts were found in the mixed sand that filled thetrench. No postmolds were noted in the trench norwas their any indication that sharpened posts had

Figure 12: Plan of Central Excavation area.

Figure 13: Unexcavated linear soil stain, facing east.

18

been driven into the bottom of the trench. The ab-sence of postmolds and the shallowness of thetrench indicate that it cannot be a palisade trenchand more likely is of modern origin.

The large pit-like feature was sectioned andfound to contain two layers, an upper stratum ofmottled grey and light grey sand (FRER73D) thatcovered a layer of dark brown sand (FRER73E)(Figure 15). No artifacts were recovered from thesection. The irregular sides and tapering deep bot-tom suggest that the feature is a treehole.

No artifacts were found in the Central Exca-vation Area.

West Excavation Area:Harrington Trench withPossible Postholes

This area was reinvestigated by the VCF firstin 1991 when two test pits were excavated, one 10-x-10-ft. and one 5-x-10-ft., to uncover the possiblepostholes located by Harrington. The 1991 workwas inconclusive, so additional excavation of thisarea was conducted in 1994 by opening 3 10-ft.squares (FRER 68, 76, 77), 2 half squares (FRER70,77), and 1 three-quarters square (FRER69) (Fig-ure 16).

There were several features (FRER68D,E,F,K and FRER69E) and other undesignated fea-tures in FRER70 that were comprised of a blacksandy nucleus ringed with whitish-grey sand, iden-tical in appearance to features interpreted as treeroots in the East Excavation Area. The general

Figure 14: Profile of linear soil stain.

Figure 15: Section of pit-like feature FRER 73D&E, facingnorth.

Figure 16: Plan of West Excavation Area.

S N

19

stratigraphy was consistent with the other two ex-cavation areas; beneath the humus and root matwas a layer of grey sand that covered a bottom layerof mottled light grey sand.

Features FRER68H,J, and K were sectionedand while FRER68K proved to be a treehole, fea-tures FRER68H and J appeared to be postmolds,though of unknown period. FRER68H and J bothhave well-defined edges in plan and were filledwith brown sandy loam rather than the black andgrey sandy loam that was found in treehole fea-tures. Also, both features had smooth, regular sidewalls and bottoms that are more characteristic ofpostmolds than treeholes (Figures 17 and 18).

No artifacts were found in the West Excava-tion Area.

1994 Survey Test PitsA total of nine 5-x-5-ft. test pits were exca-

vated in 1994. Three test pits (FRER60,61 and62) were excavated at 20 ft. intervals along atransect that extended from the service road thatconnects the parking lot to the amphitheater to thebase of the dunes at the north end of the island. Afourth test pit (FRER83) was excavated betweenthe parking lot and the sand dunes. One test pit(FRER78) was located just east of the fort and

three test pits (FRER79,80,and 81) were dug inthe south end of the wooded tract between the re-constructed fort and the Elizabethan Garden.None of these test pits uncovered any prehistoricor historic strata, features or artifacts. One test pit,FRER82, contained evidence of 16th-century activity.

FRER82Approximately 200 ft. west of the recon-

structed earthen fort is a north-south ridge thatseparates the open park around the fort from anuneven heavily wooded section that adjoins theElizabethan Garden. In 1947/48, Harrington ex-cavated a test trench in this area about 375 ft.northwest of the fort and about 350 ft. south ofRoanoke Sound and found two Spanish olive jarsherds in the upper sand layer. Aside from this,there has been no other archaeological testing here.

FRER82 was located about 400 ft. west ofthe fort and about 200 ft. from Roanoke Sound ina bottom that gradually slopes up to the south withsand ridges surrounding the remaining three sides.Beneath the humus were two layers of sand(FRER82B,C) consisting of an upper light brownloose sand covering a more compact stratum ofpale brown, almost whitish, sand that togetherwere 16 in. thick. Below this was an 8-10 in. thicklayer of dark brown sand (FRER82D) with char-coal flecks and Native American ceramics (Figure

Figure 17: Profile of feature FRER68H, facing northeast. Figure 18: Profile of feature FRER68J, facing northeast.

20

19). Dr. David Phelps identified the pottery as be-ing 3 sherds of Colington Plain, 4 sherds ofColington Simple Stamped, 11 residual sherds,and one fragment of a tobacco pipe stem. No arti-facts were found in the underlying brown sandlayer (FRER82F) which was an old topsoil on topof sterile sand (Figure 20). No other artifacts wererecovered from the FRER82D layer, nor were anycultural features observed.

1995 Shovel Test HoleSurvey

A total of 15 shovel test holes (STH),FRER88-102, were excavated in 1995. The testholes were a minimum of 1 ft. in diameter andgenerally placed at 25 ft. intervals, though the to-pography and nature trail necessitated changingthe spacing in some instances. The test holes wereexcavated in lines radiating out from FRER82 un-til the layer corresponding to FRER82D disap-peared from the stratigraphy.

The shovel testing revealed that layerFRER82D was quite extensive. Twenty-five ft. tothe west, test hole FRER90 located the black sandlayer 17 in. below modern grade where it was 10

Figure 20: South profile of FRER 82.

E W

Figure 19: FRER82 after excavation. The hole left of the ranging pole is a test hole, facing south.

21

in. thick. It contained charcoal bits, a shell frag-ment, and a piece of an English tobacco pipe bowl.The back sand layer was not present at either 50 ft.or 75 ft. west of FRER82. Test hole FRER91, 24ft. east of FRER82, found the black sand layer over2 ft. below modern grade. The 15 in. thick layercontained much charcoal but no artifacts. It wasalso present 62 ft. to the east in test hole FRER95,where 2 ft. below modern grade the black sand was11 in. thick. No shell, charcoal, or artifacts werepresent at this point. Thick sand ridges preventedfurther testing to the east. The black sand layer wasdetected 25 ft. south of FRER82 in test holeFRER89 at a depth of 22 in. below modern grade.The 13 in. thick layer contained charcoal bits butno artifacts. Fifty ft. to the south, the black sand,nearly 14 in. below modern grade, diminished to 8in. in thickness and was devoid of charcoal and ar-tifacts.

Sand ridges blocked testing to the north,however test holes found the black sand layer run-ning northeast from FRER82 at distances of 50 ft.,87 ft., and 100 ft. The 87 ft. test hole, FRER96,hit the black sand layer at slightly more than 2 ft.below modern grade where the layer was 8 in.thick and contained 6 small fragments of charredbone, 3 jasper flakes, 2 Colington sherds, and onesherd of Spanish olive jar. The 100 ft. test hole,FRER97, hit the black sand layer 22 in. below

modern grade where it was about 12 in. thick andyielded charcoal bits, a sherd of Colington’s simplestamped pottery, and one sherd of Spanish olive jar.

1995 Test PitsThe two shovel test holes that produced Eu-

ropean artifacts, FRER90 and FRER97, were ex-panded into 5-ft. squares; FRER103 was thesquare around test hole FRER90 and FRER105was the square around test hole FRER97. A 2.5-x-9-ft. trench (FRER99) was excavated out from thenorth end of FRER97 to examine the relationshipof the black sand layer to the nearby sand dunes. A5-ft. square (FRER104) was dug in the woodedmargin of the drain field adjacent to the amphithe-ater. A sand ridge separated FRER104 fromFRER97. The larger test units were excavated inan effort to discover whether there were any fea-tures in the black sand layer as well as to obtainadditional artifacts for dating and cultural affilia-tion.

FRER103, not surprisingly, exhibited a soilsequence identical to that found in squareFRER82, although a thin mottled black sand layer(FRER103D) was noted between the overlyingsand deposit (FRER103C) and the homogeneousblack sand layer (FRER103E) (Figure 21). No cul-

Figure 21: FRER103 after excavation, facing south.

22

tural features observed at the surface of the blacksand layer or after it was removed. TheFRER103D layer did contain 4 shell fragmentsand 2 sherds of Colington Fabric-impressed and 1Colington Residual sherd while the FRER103Elayer yielded 5 shell fragments, 4 bone fragments,2 Colington Plain sherds, 2 Colington Fabric-im-pressed sherds, 1 Colington Simple-stampedsherd, 9 Colington residual sherds, and a piece ofgrey flint. The flint has clearly been worked andmost likely is a gunspall.

FRER105 also displayed the same stratigra-phy as square FRER103, including the thinmottled black sand layer. Again, no features wereseen either cutting into the black sand layer orsealed by the black sand layer. Artifacts collectedfrom the black sand layer (FRER105E) were 4shell fragments, 36 small pieces of charcoal, 1 fire-cracked rock, 2 quartz flakes, 2 chert flakes, 6 pre-historic sherds, and 1 small fragment of a locally-made tobacco pipe stem. A lacquer peel of thesouth profile of FRER105 was made in order tostudy the stratigraphy under laboratory conditions.

FRER99 extended north toward RoanokeSound off the northwest corner of FRER105 witha 18 in. balk separating the two units. The stratig-raphy remained consistent and it was quite evidentthat the black sand layer (FRER99E) ran under-neath the sand dunes. While no features were ap-parent, the mottled (FRER99D) and black sandlayers did produce 55 fragments of charcoal, 2pieces of charred nut, 3 chert flakes, 8 prehistoricsherds, and 1 1/4 in. diameter lead ball.

FRER104 had the same stratigraphy as theabove three test units and was devoid of features.The mottled (FRER104D) and black sand(FRER104E) layers contained 45 pieces of char-coal, 2 shell fragments, 3 quartz flakes, 2 jasperflakes, 2 sherds of Colington Plain, 2 sherds ofColington Simple-stamped, 11 Colington residu-als, 1 crucible sherd, and 3 delftware glaze chipsthat had blue dots on a white background. Thedelftware glaze chips are too small to precisely dateand presently can be attributed only to the period1585-1800.

Three small test pits excavated in the south-ern end of the project area produced no prehistoricor historic strata, features, or artifacts.

Soil AnalysisThe soil analysis consisted of a lacquer peel,

8 sediment samples, and 3 soil chemical samplestaken from FRER105. The lacquer peel was madeof the south profile by soaking the wall with lac-quer, placing several layers of cheesecloth againstthe wall, and then painting a final coat of lacqueron the cheesecloth. The lacquer peel was only 2 ft.wide due to the extremely sandy soil. Once dry, thestiffened cheesecloth, which now had the outersurface of the profile glued to the cheesecloth, sim-ply was peeled off of the wall and placed on aboard for support. Dr. Gerald Johnson of the Ge-ology Department of the College of William andMary conducted a preliminary analysis of the peeland samples (Johnson 1996). Also, soil samplesfrom FRER105E, FRER105F, and FRER105 ster-ile were tested by a commercial agricultural labora-tory for organic matter, phosphorus, potassium,magnesium, and calcium.

FRER105E and FRER105F are the upperand lower parts of a paleosol, respectively. The pa-leosol developed on an aeolian sand and then wasburied by a single event that lasted from severalhours to several days. The upper paleosol has me-dium organic matter, the lower paleosol has loworganic matter and the sterile has very low organicmatter. The upper paleosol is also richer in potas-sium, and both parts of the paleosol contain morephosphates than the underlying sterile sand. Al-though both chemicals have been shown to be re-flective of human activities (Pogue 1988), thechemical makeup of the samples is consistent witha natural unaltered soil. The sandy substrate, how-ever, is not an effective medium for the attachmentand retention of soil chemicals.

A more complete interpretation of the ar-chaeological record and the conditions at the timeof the Roanoke settlements could be obtained bythe following:1) determining the topography of the modernground surface and the paleosol;2) obtaining three additional complete soil lacquerpeels from other locations at FRNHS; and,3) detailed textural and compositional analysis ofthe sands.

23

The main thrust of the VCF research in1994 was to determine whether there was any ar-chaeological evidence of either the 1585 or 1587colonies in two specific zones. The first zone ex-tended north of the reconstructed earthen fort tothe sand dune ridge along the north edge ofRoanoke Island toward the beach where several ar-tifacts and the remains of at least one, and possiblytwo wells, all attributable to the 16th century, hadbeen found. The survey tested the theory that thebeach artifacts and wells are the remnants of the1585/1587 settlements since lost to erosion andthat the science center was on the periphery of thevillage. The East Excavation Area, north of the re-constructed earthen fort, and the four test squaresat the base of the sand dunes produced no 16th-century artifacts, no shell or bone, and no featuresor strata related to the Roanoke settlements.

The second zone, west of the reconstructedearthen fort, addressed the speculation that theearthen fort is a western strongpoint or bulwarkfor a wooden fort by investigating features first un-covered during J.C. Harrington’s surveys in 1947and 1948. The Central Excavation Area reexam-ined a linear soil stain that resembled fortificationpalisade slot trenches that have been discovered onearly 17th-century Virginia sites. The West Exca-vation Area explored a line of possible postholes todetermine if they were part of a domestic earthfaststructure. Neither the linear soil stain nor the pos-sible postholes proved to be 16th-century culturalfeatures. The former is likely a 20th-century trenchand the latter are treeholes. No artifacts werefound in either area.

The complete absence of any evidence of the16th-century forts or villages prompted a limited

survey away from the immediate environs of thereconstructed earthen fort and toward the thicklywooded Nature Trail tract between the fort and theElizabethan Gardens. This property had neverbeen archaeological surveyed until the 1994 fieldseason when a 5-ft. square test pit was excavatedhere and uncovered a cultural stratum beneathmore than 2 ft. of sand. The black sandy loam con-tained 22 sherds of Indian pottery and numerousfragments of charcoal. All the pottery was identi-fied by Dr. David Phelps as belonging to theColington series whose temporal range includesthe late 16th century.

The 1995 survey consisted of the excavationof a series of 1-ft. test holes out from the 1994 testpit to define the limits of the artifact-bearing layeras well as four larger test pits to recover additionalartifacts and search for archaeological features inorder to determine whether buried land surfacecontained evidence of 16th-century English occu-pation. While no features were uncovered, Euro-pean artifacts were recovered from the black sandlayer including two sherds of Spanish olive jar, acrucible sherd, a lead shot, delftware glaze, a frag-ment of an English tobacco pipe bowl, and a pieceof a gunspall. Numerous sherds of Colington ce-ramics also were collected. All of this materialcould date to the late 16th century. Given manycrucible sherds found during the 1991-1993 exca-vations of the metallurgical workshop of JoachimGans, the crucible sherd likely is associated withthe 1585 Lane settlement. While it is not knownwhat occurred on this site, it appears that the thicksand layers have preserved a land surface that wasused by the English during the time of theRoanoke settlements.

Summary

24

25

What can be learned from looking at the“big picture” formed by the three major archaeo-logical campaigns at Fort Raleigh? Despite the1980’s excavations by the NPS and the 1990’swork by the VCF, the answer is the same reachedby J.C. Harrington at the end of his 1947-1950archaeological project. In Harrington’s words“…conclusion from the explorations carried onthus far is that the habitation site has not been lo-cated (Harrington 1962:38).” Most Roanoke Is-land settlement researchers will go even further to-day, believing that neither the habitation site northe fort of either the Lane Colony or the LostColony has been discovered. Harrington reachedhis conclusion in part by comparing his findings toa similarly short-lived French settlement, St.Croix. In view of the additional archaeology atFort Raleigh and Fort St. George, it seems a usefulcomparison to revisit, particularly given that theLane Colony, the Lost Colony, St. Croix, and FortSt. George are quite similar in many respects. Allwere settlements of nearly the same size (100-117people), established generally at the same time(1585-1607), and each lasted about one year.

The first attempt to settle New France wason St. Croix Island, Maine. In June of 1604, 111men under Sieur de Monts established a colonythat was abandoned in June of 1605. John L. Cot-ter (1978) has summarized preliminary archaeo-logical testing in 1950 by Wendell S. Hadlock(1950) and survey and area excavation in 1968-1969 by Jacob W. Gruber (1970). The total areaexcavated at the north end of the island is approxi-mately 4000 ft.2. The finds consist of 1015 ce-ramic sherds (some date to the 18th and 19th cen-turies, but the vast majority belong to the St. Croixperiod), 250 yellow clay brick bats, 452 glass frag-ments, 51 glass beads, 1005 complete or fragmentsof wrought nails, 5 musket balls, 5 lead shot, leadscrap, 99 copper-lead-iron objects, daub, bone,and pieces of flint.

The Virginia Company, a joint stockcompany with a charter to colonize NorthAmerica, was divided into the London Companywhich established Jamestown and the less well

Conclusionsknown Plymouth Company that founded thePopham Colony on the Sagadahoc (nowKennebec) River in Maine. The Popham Colonybegan in August of 1607 when 100 settlers landedand set about constructing Fort St. George. Ap-proximately half of colonists returned to Englandin December of the same year due to diminishingsupplies, leaving 45 men at Fort St. George. Theentire settlement was abandoned in September orOctober of 1608. Jeffrey P. Brain (1995) directedrecent archaeological investigations at Fort St.George as well as reviewed earlier testing byWendell S. Hadlock. Brain dug 51 excavationunits totaling approximately 585 ft.2. The excava-tions produced 732 ceramic sherds, 27 fragmentsof green glass, 2 glass beads, 54 tobacco pipe stemand bowl fragments dating to the Fort St. Georgeperiod, at least 84 wrought nails, 5 musket balls, 2lead shot, 5 sprue, lead scrap, a few pieces of cop-per sheet metal, a seal-top spoon, and a bale seal.Large quantities of daub and shell also were found,though much of the shell could be attributed toprior Native American occupation.

Harrington’s 1947-1948 testing west ofthe reconstructed earthwork excavated approxi-mately 17,500 ft.2 and recovered 2 nails, 3 leadmusket balls, 5 crucible fragments, 3 olive jarsherds, 1 stoneware sherd, and 3 brick fragments.His subsequent excavation of the earthen fort inte-rior and ditch in 1950 produced the following Eu-ropean artifact assemblage: 1 iron sickle, 1 iron au-ger, 10 iron spikes and 2 spike fragments, 3 nailfragments, 2 U-shaped iron objects, 1 12 in. longiron bar, 1 iron hinge of questionable date, 3 brasscasting counters or jettons, 1 small brass balanceweight, 2 copper nuggets, 2 lead musket balls, 1glass bead, 22 Spanish olive jar sherds, 2 majolicasherds, a coarseware sherd, 1 brick fragment, and 1roofing tile fragment (Harrington 1962).Harrington’s excavation of the “outwork” in 1965resulted in the following collection: 25 sherds ofNormandy stoneware representing one costrel, 4whole bricks, pieces of 8-9 additional bricks, 20 al-tered brick pieces, approximately 3 roofing tiles,and sherds of Native American pottery (Har-

26

rington 1966).The 1982/83 NPS investigations adjacent to

the entrance of the reconstructed earthwork exca-vated 1120 ft.2 and found 38 European artifacts(Ehrenhard and Komara 1984), mostly sherds ofcrucibles, ointment pots, and Normandy flasks, allof which can be attributed to the 1585 science cen-ter. The 1991-1993 VCF excavations, all immedi-ately west of the reconstructed earthwork, con-sisted of digging a minimum of 3000 ft.2 andproduced 119 pieces of crucible, 9 cupel frag-ments, 17 sherds of Maiolica ointment pot, 19sherds of West of England butterpot, 136 sherds ofNormandy flask, 2 sherds of Spanish jar, 25 frag-ments of bottle glass, 5 pieces of roofing tile, 119sherds of Native American pottery, as well as 58flint chips, 106 flakes of iron scale, 1 piece of leadwaste, 1 bale seal, and 1 lump of antimony (NoelHume 1995). The 1994 VCF survey excavatednearly 1000 ft.2 in the immediate vicinity of the re-constructed earthwork, but away from the sciencecenter, and found no 16th-century artifacts.

The total area excavated at Fort Raleigh isabout 20,000 ft.2 (22,620 minus an allowance forsome overlapping excavations) versus 4000 ft.2 atSt. Croix and 585 at Fort St. George. Yet, despitethe vastly larger excavated area, the Fort Raleighartifact assemblage significantly lacks a domesticcomponent that is present at St. Croix and Fort St.George. Virtually all the artifacts recovered at FortRaleigh by Harrington and Ehrenhard almost cer-tainly were associated with the Gans/Hariot scien-tific workshop identified by the VCF and what isconspicuously absent is the refuse normally associ-ated with European domestic sites. There are nopots, pans, pitchers, bones, or shells (Noel Hume1995:63-64). There is also a striking shortage ofarchitectural materials at Fort Raleigh in contrastto the large quantities of nails, brick, and daubseen at St. Croix and Fort St. George. The limitedwork at St. Croix and Fort St. George makes itquite plain that a site occupied for only one year by100 people should have a lot of domestic artifacts,an idea proclaimed by Harrington (1984: 19)more than thirty years ago. This is simply not thecase at Fort Raleigh. Consequently, the answer toanother long-standing question at Fort Raleigh—how could so many colonists have left such a pau-city of artifacts?—is that they didn’t because nei-ther the fort or village was on the propertyimmediately surrounding the reconstructed earth-

work.The reconstructed earthwork cannot have

been a bastion for the 1585 Lane colony fort sinceit was the site of the scientific workshop of thesame expedition. The dearth of domestic artifactsalso strongly suggests that it is not the site of the1587 White colony either. In fact, Noel Hume(1994:88) has asserted that the earthwork is of18th-century vintage. However, there are two16th-century interpretations that are still conceiv-able. One is that the earthen fort was built for thegarrison left by Grenville in 1586 — not a newidea and one which has its detractors. The secondpossibility hinges upon the presence of sciencecenter artifacts inside the earthen fort and the factthat they were recovered from a disturbed context;thus it may be that the science center horizon didnot extend under the earthen fort and that the fort(or sconce) was built to protect the science centermuch in the same way that the Lane colonists builtearthen fortifications during a larcenous raid ofSpanish salt mounds on Puerto Rico (Figure 22).If one accepts the premise that the forts and settle-

Figure 22: Earthwork contructed by Lane colonists onPuerto Rico (Hulton 1984).

27

ments are somewhat removed from the sciencecenter, then the necessity for the science center tohave some form of individual protection becomesmore credible.

There were two instances where 16th-cen-tury artifacts were recovered from excavations ofthe earthwork other than from the trench. Har-rington (1962:21-22) found a nugget of copper, asherd of an Iberian storage jar, and a musket ballunder the remnant of the original parapet. Theseobjects must be related to the establishment of thescience center which therefore predated the con-struction of the earthwork. The artifacts, however,do not indicate when the earthwork was built, onlythat it was done sometime after the science centerwas set up. Nor were the artifacts definitively de-posited within the walls of the workshop; theycould simply have been part of the workshop mid-den. The same is true for the artifacts found insidethe earthwork during the VCF excavations. Con-sequently, since the earthwork did not necessarilyabolish the science center, it could have been builtsometime after the science center was functioningand before the Lane Colony was abandoned. Al-though the science center was unknown to him atthe time, Harrington (1962:53) speculated that

the earthwork could have been constructed in thespring of 1586 when the English-Indian relation-ship became increasingly hostile. David Stick(1993), building on Harrington’s conjecture, hassuggested that the earthwork was built to protectthe science center from the same threat. A smallearthwork can be erected expeditiously andcheaply— the comparable earthwork encirclingthe Puerto Rican salt mounds was built quicklywith 31 men in approximately two days (Quinn1991:184-185).

The 1994-95 archaeological work at FortRaleigh NHS reinforced the findings and conclu-sions of the 1991-93 VCF excavations led by NoëlHume, namely the remains of the 1585/1587 fortsand villages are not located within the immediatesurroundings of the reconstructed earthwork. NoëlHume has advised that archaeologists are now freeto search other areas for evidence of the Roanokesettlements (Noël Hume 1994: 88). Of course, oneunhappy prospect of this situation is that erosionhas completely consumed the sites (Dolan andBosserman 1972; Dolan, Hayden, and Bosserman1981). However, all possible alternatives should bereviewed, including areas previously surveyed inthe 1940’s.

28

There still is potential for future archaeologi-cal research related to the science center. The Na-tive American/16th-century artifact scatter locatedduring the 1995 survey is one such location. Thesurvey demonstrated that the limits of the scientificworkshop and associated areas are more wide-spread that previously believed. While the cruciblefragment from FRER103 seemingly links this area400 ft. west of the reconstructed earthwork to thescientific workshop, the two are separated by a sub-stantial artifact gap manifested by the 1994 WestExcavation Area, implying that the FRER103 vi-cinity is a related but detached element. One pos-sible explanation is that it is an undiscovered work-shop component, perhaps Thomas Hariot’s housethat may been near the scientific workshop (NoelHume 1995:107).

The vicinity of Harrington’s charcoal-filledpit is also worthy of additional investigation (NoelHume 1995:83). While making charcoal in a pit isan unusual method, there is 16th-century prece-dent for doing so (Figure 23), although how thecharcoal was made at the site remains an openquestion (Noel Hume 1995:81). Regardless ofwhether the charcoal was made in the pit or not,carbon-14 dating indicated that the charcoal datedto the late 16th-century. This strengthens the ap-parent connection between the science center andthe pit since Gans’ assay furnaces surely were fu-eled with charcoal, and it is therefore possible thatthere are additional science center features present.

Future Research

Figure 23: 1540 engraving of making charcoal in a pit (Biringuccio 1966: 178).

29

Biringuccio, Vannoccio 1966 (1540) Pirotechnia. The M.I.T. Press,

Cambridge, MA and London, England.

Brain, Jeffrey P.1995 Fort St. George, Archaeological Investiga-

tion of the 1607-1608 Popham Colonyon the Kennebec River in Maine.Manuscript on file, Peabody EssexMuseum, Salem, MA.

Cotter, John L.1978 St. Croix National Historic Site: A

Beginning in Historical Archaeology.Manuscript on file, National Park ServiceNortheast Cultural Resources Center,Boott Cotton Mills Museum, Lowell,MA.

Dolan. Robert and Kenton Bosserman1972 “Shoreline Erosion and the Lost Colony.”

Annals of the Association of AmericanGeographers, vol. 62, no. 3. pp 424-426.

Dolan. Robert, Bruce P. Hayden, and KentonBosserman

1982 “Is the Lost Colony Really Lost.” Shoreand Beach, vol. 50. pp 17-19.

Ehrenhard, John E., William P. Athens, andGregory L. Komara

1983 Remote Sensing Investigations at FortRaleigh National Historic Site, NorthCarolina. Manuscript on file, SoutheastArchaeological Center, National ParkService, Tallahassee, Fl.

Ehrenhard, John E. And Gregory L. Komara1984 Archaeological Investigations at Fort

Raleigh National Historic Site, Season 2,1983. Manuscript on file, SoutheastArchaeological Center, National ParkService, Tallahassee, Fl.

Gruber, Jacob W. 1970 The French Settlement on St. Croix

Island, Maine: Excavations for theNational Park Service, 1968-1969.

Manuscript on file, National Park ServiceNortheast Cultural Resources Center,Boott Cotton Mills Museum, Lowell,MA.

Harrington, Jean Carl1962 Search for the Cittie of Ralegh, Archaeo-

logical Excavations at Fort RaleighNational Historic Site, North Carolina.Archaeological Research Series NumberSix, National Park Service, U.S. Depart-ment of the Interior, Washington, D.C.

1966 An Outwork at Fort Raleigh. EasternNational Park and Monument Associa-tion.

Hulton, Paul1984 America 1585, The Complete Drawings of

John White. University of North CarolinaPress, Chapel Hill.

Johnson, Gerald.1996 Preliminary Analysis of Fort Raleigh

Sediment Peel, Manuscript of file,Virginia Company Foundation,Williamsburg, VA.

Keel, Bennie C.1991 Trip Report, Fort Raleigh National

Historic Site, SEAC Accession Number927, Southeast Archaeological Center,Tallahassee, FL.

Kelso, William M., Nicholas M. Luccketti, andBeverly Straube.

1990 A Reevaluation of the ArchaeologicalEvidence Produced by Project 100.Manuscript on file, Virginia CompanyFoundation, Williamsburg, VA.

1995 APVA Jamestown Rediscovery FieldReport 1994. Manuscript on file, APVA,Jamestown Island, Va.

Noel Hume, Ivor1982 Martin’s Hundred. Alfred A. Knopf, New

York.

Bibliography

30

1994 The Virginia Adventure, Roanoke to JamesTowne: An Archaeological and HistoricalOdyssey. Alfred A. Knopf, New York.

1995 First and Lost: In Search of America’sFirst English Settlement. Draft Manu-script on file, Fort Raleigh NHS,Manteo, NC.

Outlaw, Alain C.1990 Governor’s Land, Archaeology of Early

Seventeenth-Century Virginia Settlements.University Press of Virginia,Charlottesville.

Phelps, David S.1983 Archaeology of the Native Americans:

The Carolina Algonkians. Manuscripton file, Institute for Historical andCultural Research, East CarolinaUniversity, Greenville, NC.

Poque, Dennis J.1988 “Anthrosols and the Analysis of Archaeo-

logical Sites in a Plowed Context.”Northeast Historical Archaeology. Vol. 17,1988.

Quinn, David B (ed).1991 The Roanoke Voyages. Dover Publica-

tions, Inc., New York.

Stick, David1993 Letter form David Stick to Ivor Noel

Hume, December 24, 1993.

31

Finds ListER90C

1 Prehistoric pottery1 English clay tobacco pipe bowl1 Shell

ER91C1 Sand concretion

ER 961 Spanish olive jar3 Chert flakes7 Sand concretions2 Prehistoric pottery6 Charred faunal

ER971 Spanish olive jar1 Prehistoric pottery4 Sand concretions

ER97D7 Sand concretions3 Charred wood

ER97E6 Prehistoric pottery4 Shell2 Quartz flakes2 Chert flakes1 Fire cracked rock1 Pebble36 Charred wood103 Sand concretions1 Locally made clay tobacco pipe stem

ER97F239 Sand concretions63 Charred wood

ER98D1 Shell1 Charred wood

ER99D2 Prehistoric pottery4 Shell2 Charred nut1 Charred faunal4 Charred wood1 Charred sand concretion

ER99E6 Prehistoric pottery1 Lead shot3 Chert flakes51 Charred wood259 Sand concretions

ER99E-16 Prehistoric pottery1 Sand concretion1 Limonite

ER99E-22 Prehistoric pottery

ER103C2 Charred faunal2 Charred wood1 Shell

ER103D3 Shell3 Prehistoric pottery1 Faunal1 Wood

ER103E14 Prehistoric pottery4 Faunal1 Chert flake1 Pebble4 Shell93 Sand concretions

ER104C6 Prehistoric pottery2 Charred wood1 Chert flake2 Pebbles

ER104D43 Charred wood1 Charred faunal3 Quartz flake1 Pebble3 Delftware glaze chips (white with

blue dots)9 Prehistoric pottery1 Sand concretion1 Crucible2 Shell

Appendix A

THE ASSOCIATION FOR THE PRESERVATION OF VIRGINIA ANTIQUITIES

R204 W Franklin Street, Richmond, VA 23220