1

2

Tobacco-Pipe Manufacturing in EarlyMaryland: The Swan Cove Site (ca.1660-1669)

Al Luckenbach and C. Jane Cox

Situated on the banks of a relic cove off MillCreek in Anne Arundel County Maryland, the SwanCove Site (18AN934) was occupied by planter andtobacco-pipemaker Emanuel Drue from perhapsthe 1650s until his death in 1669. Drue used stateof the art, European production techniques in themanufacture of tobacco-pipes made of native clays.Drue produced two main forms, a “Chesapeake”style angular elbow pipe and a European “bellybowl.” The pipes and kiln debris recovered fromintact features at Swan Cove characterize a seven-teenth-century industry that has yet to be studied inthe New World. The evidence of pipemaking atSwan Cove challenges many assumptions aboutlocally-made pipes and undoubtedly will energizefurther studies of locally-made pipes in the Chesa-peake.

IntroductionSwan Cove is one of only eight known sites

that were part of the 1649 settlement of Providence.Over the past decade, thetown of Providence hasbeen the subject ofinvestigation and excava-tion by Anne ArundelCounty’s Lost TownsProject (Figure 1). Initialinvestigations at the SwanCove site tobacco-pipekiln are providing newinsights into the tobaccotrade, pipemaking, andsome of the earliestindustrial activity on theseventeenth-centuryfrontier of Maryland.

Scholars have long been interested in thepractice of local tobacco-pipemaking in the Chesa-peake Bay region, variously concluding thatEuropean, indigenous, and slave populations wereinvolved in the production of terra-cotta pipes inthe seventeenth century (Emerson 1988, 1994).Particularly intriguing theories associate their usewith servants and slaves, as well as African-American and Creole populations (Mouer 1993,Neiman and King 1999).

While locally-made pipes are abundant atseventeenth-century Chesapeake sites, archaeologi-cal evidence for the physical production of thesepipes is extremely limited. When found, suchevidence consists solely of damaged pipes orunusual clay “shavings” or “blobs” (TomDavidson, Ann Markel, personal communications2000). In some instances, these clues are attributedto small-scale pipe manufacture in home fireplaces(Emerson 1988). In 1997, excavations atJamestown, Virginia may have uncovered evidenceof early pipemaking, supported by the recovery ofpipes made of local clays, a debatable “saggar” (atype of kiln furniture), and documentary research

Figure 1: Volunteer Mac Milhone excavates Feature 7.

3

(Lukketti and Straube 1998). Such hints of to-bacco-pipe production had yet to provide conclu-sive evidence of pipe kiln operations in the colonialChesapeake region.

At the time of his death in 1669, Emanuel Drueowned “one payer of pype moulds, brasse andmaterials belonging to them” (MSA 1670:67)(Figure 2). Archaeological evidence clearlyindicates that Emanuel Drew used this pair of

molds to produce Chesapeake tobacco-pipes as acraft industry on his plantation. Though the moldsthemselves have not been (and are not likely to be)recovered, there are copious quantities (over 200pounds) of highly fired and broken kiln furnitureknown as “muffles,” the pipe manufacturingequivalent of “saggars” used in pottery production.A muffle is a type of kiln furniture, usually a “largerefractory pot set inside the firing chamber formingan inner chamber to contain the pipes … fromdirect contact with the flame” (Peacey 1996).

At Swan Cove, Lost Towns Project archaeolo-gists have discovered the first tangible and distinc-tive archaeological evidence for a full-scale kilnproduction of pipes in the New World at the homesite of Emanuel Drue from the late 1650s until

1669. Even based upon this preliminary assess-ment, the pipe assemblage from Swan Cove is ahighly significant collection. It illustrates, in part,the products and the production of Drue’s industry.Lost Towns Project archaeologists have identifiedtwo distinctive pipe forms at Swan Cove (and twounique forms) that also have been found at nearbycontemporary Providence sites, Burle’s Town Landand Homewood’s Lot.

Historic and EnvironmentalContext

The Drue family is first docu-mented at the Broad Creek parcel onSwan Cove in 1660. However,given the lack of surviving recordsfrom earlier periods, they may havearrived in the region during theinitial 1649 Puritan influx to Provi-dence. Hugh and Emanuel Drue,along with a third male Drue(possibly a brother or a son), appearto have been transported into thecolonies through Virginia. In 1654,Emanuel Drue transports three

people into Maryland, including his “now wife”Elizabeth. Existing evidence reveals that the Druefamily held two indentured servants, a youngwoman and a young man, during their tenure atSwan Cove.

The Drue brothers did not warrant Broad Creekuntil March of 1661. In 1665, Emanuel Drue addsan additional 50 acres, known as Swan Cove, to hisholdings. Upon Emanuel’s death in 1669, his twominor sons, Thomas and Samuel, inherit the BroadCreek/Swan Cove property, to be enjoyed uponreaching their respective ages of majority, and histwo daughters, Elizabeth and Martha are sent tolive with neighbors. Although the sons are scarcein Provincial records, Emanuel Drue’s will indi-cates that they were apprenticed to local artisans,

Figure 2: Portion of Emanuel Drue’s probate inventory.

4

and his daughters were cared for and educatedappropriately by friends of the family (MSA1669:349).

The next reference to Swan Cove is in 1707,revealing that Henry Merriday, a planter, “has longbeen seated” on the tract of land known as SwanCove (MSA 1707: 611). In 1721, Amos Garrettadded this parcel to his larger plantation landhold-ings, purchasing the property from Henry Merridayand his son for £44.2.9 sterling and in considerationof livestock (MSA 1721: 455). Swan Cove ceasesto be occupied as a domestic site after this landtransaction, until William Storck acquired theproperty in the mid-twentieth century.

Archaeological InvestigationsAmateur archaeologist Bob Ogle and property

owner Bill Storck discovered the site duringlandscape grading in the late 1970s and assembleda collection at that time. The site was brought tothe attention of Anne Arundel County Archaeolo-gist, Al Luckenbach, in 1991. Anne ArundelCounty’s Lost Towns Project conducted prelimi-nary investigations in 1992, and returned to con-duct additional testing in 1998. Extensive excava-tions began in the fall of 2000 with excavationunits now numbering nearly 70 5-x-5 ft. units(Figure 3). These excavations have been aug-mented by geophysical investigations, includingcomprehensive magnetometer and ground penetrat-ing radar surveys. Artifacts have been processedand stored at Anne Arundel County’s Lost TownsProject Archaeology Lab in Annapolis, Maryland.In addition, Mr. Ogle’s collection has been madeavailable to the Lost Towns Project for study.

Swan Cove is in close proximity to at leastthree other Providence sites. Located 3000 ft.northwest of Burle’s Town Land, 2500 ft. west ofLeavy Neck, and 6000 ft. west-northwest of theTanyard, the Swan Cove site is a central componentof the hamlet-type settlement of Providence.

Typical of the other known seventeenth-centuryProvidence sites, Swan Cove is immediatelyadjacent to a spring, near a cove that providesaccess to a navigable waterway (Mill Creek), andon a flat spot of land at an elevation of about 20 ft.above mean sea level.

Emanuel Drue’s Pipe Kiln and Kiln DebrisThe kiln foundation itself has thus far eluded

discovery. Although the search continues, it isquite possible that it failed to survive the ravages oftime, since plowing, erosion, grading, and ditchdigging all occurred at the site over the last threeand a half centuries. However, the vast quantitiesof kiln debris found inside several still-intact trashfeatures at Swan Cove appear to represent periodicrebuilding of the structure, and contain invaluableclues as to its nature.

Pieces of the kiln’s interior structure are

Figure 3: Anne Arundel County’s Historic Sites Planner,Donna Ware, excavates Feature 19.

5

represented by what we have branded “loaves.”These handmade clay objects are vaguely shapedlike bread; they are ovoid, concave on the bottomand convex on the top (Figure 4). The loavesalways exhibit clear evidence of high firing on oneface (either the top or side) and are almost of“salmon brick” consistency on the other face.Drue’s fingerprints frequently provide evidence ofthe homemade nature of theseobjects. Loaves undoubtedlyrepresent either the floor of thekiln itself, or objects that sat onthe floor.

Interestingly, no traditional,rectangular-shaped bricks havebeen found at the site, although itwould have been quite simplefrom Drue to make them himself.Even the imported yellow Dutchbricks, so common at othercontemporaneous Providencesites, are not seen at Swan Cove.Numerous large, river cobbleshave been recovered, which may

be somehow involved in the tobacco-pipe kilnconstruction (Figure 5). This seems to be con-firmed by the discovery of cobbles with one surfacevitrified to a glaze-like consistency by exposure toextremely high temperatures.

The most frequently encountered kiln debrisconsists of nearly 200 pounds of “muffles.”Muffles are large, crude ceramic vessels, roughly

18 in. in diameter, used in the kilnto distribute heat evenly andprotect the contents from heatdamage (Figure 6.) At SwanCove, they are made with stackedrings of clay reinforced by brokentobacco stems. The tobacco-pipestems are placed within thevessel’s fabric at opposing anglesto form a herringbone pattern.This kept the stacked rings lockedtogether. Muffles recovered fromSwan Cove appear to have met acatastrophic end. They showevidence of having undergoneextreme heat and stress before

Figure 5: A vitrified river cobble.

Figure 4: Loaves recovered from Swan Cove.

6

finally disintegrating, at which point they wereremoved from the kiln.

Other kiln furniture has been recovered insmall numbers, including two examples of aninteresting form called a “cross-pipe prop” (Figure7). The only other documented example wasexcavated in Chelmsford, England (Peacey 1996:40,41). Another, enigmatic item is perhaps a “bun”which was used to stack layers of props within themuffle (see Peacey 1996: 46-49) (Figures 8 and 9).

Fragments of what appears to be a low-firedearthenware kiln “dish” used as a kiln spacer (seePeacey 1996: 50-53) also were found. In tradi-tional archaeological parlance, this vessel might betermed “Drue Ware,” its crude, handmade formargues strongly for its expedient nature. Its crudityalso seems to speak volumes concerning Emanuel

Drue’s lack offamiliarity withthe construction oftraditional Englishclay pots.

In studyingthe kiln’s output,the imprints of atleast eight differ-ent pipe-decorat-ing tools havebeen noted. Theseinclude threedecorative stamps,a smaller circularpunch, and fourdistinct roulettingtools (Figure 10).One of the mostinteresting arti-facts recoveredfrom the SwanCove excavationsis one of the

decorative stamps used by Emanuel Drue (Figure11). This is an exceedingly rare object. Even inEngland, only one similar clay stamp has beenfound during the investigations of over 140 kilnsites (Peacey 1996). Comparable examples fromthe New World are unknown.

Kiln-related FeaturesAs stated, several intact features discovered at

Swan Cove promise to reveal significant informa-tion about Drue’s pipemaking operation. The twolarge features primarily discussed below, Features 7and 19, yielded a total of 1,171 pipe fragments andmore than 160 pounds of muffle and loaf frag-ments. Two much smaller pit features, Features 4and 5, also contained small amounts of kiln andpipe debris and appear to be related to the primary

Figure 6: Muffle fragments recovered from Swan Cove.

7

Figure 9: Drue’s fingerprints onthe back of the bun.

Figure 8: Kiln furniture knownas a bun.

Figure 7: A cross-pipe prop, used for supporting pipes within the kiln.

Figure 10: Examples ofdecorations used by Drue. Figure 11: The stamp used by Drue to decorate his pipes.

8

pits. All were sealeddeposits clearly datingfrom the period ofpipemaking at SwanCove, or roughly betweenthe late 1650s and 1669(Figure 12).

These features arecertainly indicative of anintensive kiln operation,although their precisefunction has not yet beendetermined. In theabsence of a kiln struc-ture, these featurescurrently provide the onlyindication of the exactnature of Drue’s opera-tion.

The two large pits detailed below were adjacentto one another, with Feature 7 just to the north ofFeature 19. Feature 7 is a large rectangular pit,approximately 6-x-3 ft., averaging 1½ ft. in depth(Figure 13). Much of the kiln debris was concen-

trated in the northwestern portion of the feature,and included raw clays in numerous colors.Lightly interspersed throughout were domesticartifacts, oyster shells, bones, etc. The ceramicsrecovered consisted of at least twenty individuallead-glaze redware vessels (80 percent of the

assemblage) and only two delftwarevessels (8 percent), which wereconcentrated toward the eastern halfof the feature. The remaining 12percent of the ceramic assemblage isrepresented by one vessel each ofRhenish brown, North Devon graveltemper, and Midlands purple. Fea-ture 7 also produced numerousburned fragments of imported Dutchroofing material (pantile), inter-spersed with the kiln debris. Thebase of this feature was sloping, witha notably deeper section occurring inthe eastern half.

Feature 19 measured about 5 ft.across and 1 ft deep to a level bottom

Figure 12: Feature 7, excavations in progress.

Figure 13: Site plan of Swan Cove (2001.)

APPLE TREE

f4f5

f19

f7

Swan Cove (18AN934)

20 Ft. North

9

(Figure 14.) Concentrations of kiln debris, includ-ing large pockets of multicolored raw clays, werefound throughout this feature. Along with otherdomestic debris, the primary ceramics recoveredinclude seven delftware vessels (39 percent) andthree lead-glaze redware vessels (17 percent). Theremaining ceramic assemblage consisted of onevessel each of Rhenish blue and gray, Rhenishbrown, Borderware, Midlands purple, North Devongravel-tempered, North Devon sgraffito, andBorderware. Importantly, there were no cross-mends between the Feature 19 and Feature 7ceramics.

The relative frequency of utilitarian redwaresand concentrations of pantile in Feature 7, incomparison to the few lead-glaze redwares andpantile found in Feature 19, provides the basis forspeculating about the kiln operation. English pipeexpert Allan Peacey has noted that inexpensiveredwares are often used as props when placingitems in a kiln. He has also commented thatpantiles, with their curved shape, would make

excellent covers for vent or heat trenches (AllanPeacey, personal communication, 2001).

The stratigraphic distributions of variousartifacts within these features provide hints as tothe possible location and structure of the kiln. Itappears, based upon the density of kiln-relatedartifacts recovered from the area surroundingFeatures 7 and 19, that the kiln production area waslocated to the west of these features, while moredomestic activities may have been concentrated tothe east. The intermingling of kiln and occasionaldomestic debris does suggest that the work andhome areas were not necessarily well-segregated,lending support for the notion of a cottage industry,although on a relatively massive scale.

Based upon the assemblage from each of thesepits, it appears that the artifacts are representativeof an active pipe kiln and do not represent the kiln’sdemise. These features are invaluable as theyprovide snapshots of a tobacco-pipe kiln at differ-ent points of operation. Additional insights into thedifferent uses for each of these features can be

Figure 14: Feature 19, excavations in progress.

10

provided by considering the physical characteristicsof Drue’s pipes, as well as the raw clays, kilnfurniture and vessels Drue used to produce thepipes.

Raw Clays fromFeatures 7 and 19

The actual raw,unfired clays Drue usedto produce the myriadcolors of pipes at SwanCove have been recov-ered from Features 7and 19, providing aunique opportunity toexplore Drue’s manu-facturing process(Figure 15). Notsurprisingly, like thepipes they produced, the raw clays vary greatly incolor. Drue clearly went to a great deal of effort toobtain a spectrum of different colored clay. Inter-estingly, they do not originate at the site itself,

which only produces a yellow variety common inthe region. Although not yet confirmed by chemi-cal analyses, Drue’s clays seem to have originatedapproximately 13 miles up the Severn River wherebanks of clay outcroppings were discovered in

December of 2001 (Figure 16). Pockets of rawclay were distributed throughout the deposits withpipes and muffles. In Feature 7, the raw clays

Figure 16: Potential source of Drue’s raw clays discovered on the Severn River.

Figure 15: Raw clays recovered from intact features at Swan Cove.

11

exhibit five different colors--from reddish pink, toyellow, to dark and light gray to a nearly pure whiteor kaolin hue. In Feature 19, the color of claysrecovered was much more limited, restricted to agray hue and a greenish-gray hue.

While the clay colors found in both features arereflective of the pipe assemblage across the site, itis interesting to note the lack of greenish clay inFeature 7 and the lack of white, pink and light grayclays in Feature 19. These facts, together with thelack of ceramic cross-mends, and a different ratioof pipe forms (discussed later) strongly suggest thatthe two features are not perfectly contemporaneous.This conclusion is important because during theinitial excavations of Feature 7, consideration wasgiven to whether the contents derived from thekiln’s destruction. Since generally similar muffle,loaf, and kiln furniture fragments came from bothpits, the materials clearly had to be attributed tosome kind of period refurbishment of the kiln,rather than its demise.

Kiln Debris from FeaturesInterestingly, the muffles recovered from these

two major pits demonstrate variations in construc-tion technique that might be attributed to differentuse patterns. Typical of the 69 pounds of muffle inFeature 7 are signs of reduction firing and in manysamples, thermal glazing can be found on theexterior of the vessels. The muffles themselves aresurprisingly thin, often constructed in layers, withreinforcements of waster pipe stems within thewalls, usually in a herringbone pattern. Suchconstruction is seen in the home-counties aroundLondon, England (Allan Peacey, personal commu-nication 2000). Thus, English-derived muffleconstruction provides another indicator as to Drue’sEuropean origins. This layering constructionindicates that the muffles were used repeatedlyover a relatively long period of time. The mufflesare a lighter terra-cotta shade with extensive

whitewashing or luting. According to Peacey, theapplication of luting (a whitewash applied to theinterior of the muffle) is a gauge of the number offirings for which a muffle may have been used(Allan Peacey, personal communication 2001).

The 93 pounds of muffle recovered fromFeature 19 exhibited different physical characteris-tics. The muffles had a thicker-walled constructionand virtually no pipe reinforcements. Also, thefragments were a more consistent dark gray toterra-cotta hue. Relatively few pieces showed signsof high firing and the resulting thermal glazing.The muffle recovered in Feature 19 tended towardlarger fragments than those found in Feature 7.

The meaning of these construction variantsremains within the realm of speculation. Perhapsthe thinner-walled vessels were used in a morecontrolled firing environment, whereas the thickervessels of Feature 19 would have been used in alower firing environment. Thin-walled vesselswith pipe temper may have allowed for more aircirculation, thus more heat control would be neededfor positive results. Perhaps the thinner vesselswere seen as more disposable in nature, thus the re-application of clays on the interior and the layeringof luting were intended to prolong the life of thesevessels. The thicker, more robust vessels found inFeature 19, although cruder, may have had a longerlife expectancy. Until more experimentation can beconducted to recreate firing environments, ques-tions remain as to the functional realities of mufflevessels.

The two kiln-related features discussed abovewere clearly crucial to the production and firing oftobacco-pipes at Swan Cove. The pits appear to benearly, but not quite, contemporaneous, and neitherappears to represent the destruction of the kiln.Archaeological evidence strongly suggests aphysical division between the two pits, perhaps awall, which would also signal a division of activity

12

areas. The northern pit (Feature 7), witha more refined muffle construction, abroader variation of pipe colors, andheavier concentrations of both pantileand kiln-related utilitarian ceramics,indicates an activity area that is closelyrelated to kiln use and operation. Thepantile and lead-glazed redware appearto have been integral to the firing processand were likely used as kiln props tosupport the pipes during firing. Thesouthern pit (Feature 19), with muffleconstructed in a less sophisticatedfashion and a higher percentage ofrefined ceramics, appears to be some-what more multi-purpose and less centralto the kiln operation, even though itshows evidence of use in conjunctionwith kiln operations.

Emanuel Drue’s Tobacco-PipesDrue’s Principal Types

The pipes Emanuel Drue made at Swan Coveare very likely the products of the two moldsmentioned in his 1669 inventory. Two main typesinclude an angular elbow type, which is a classic“Chesapeake” pipe form (Drue Type A) and atraditional mid seventeenth-century English “bellybowl” form (Drue Type B).

The straight-sided Drue Type A bowl (LT Type15.01) is an angular elbow form that is quitefamiliar in the universe of Chesapeake and NativeAmerican pipe forms (Figure 17). The pipe isdistinguished by the relatively sharp angle betweenthe stem and bowl. The bowls average 1-¾ in. inheight. There is a smooth transition from stem tothe joint and through the elbow, with no heel. Thisbowl form has a small volumetric capacity, averag-ing <4.25 ml, with a notably narrow interior.Several of these bowls have compressed edges,likely a result of the kiln loading and firing pro-

cesses.Distinctive to the Drue Type A is the propensity

for decoration, including extensive rouletting onstem and bowl rim, along with the application ofvarious stamps(Figure 18). Several examplesexhibit extensive decoration. Thus, Drue Type Abowls seem to be following a decorative grammarderived from the Chesapeake, not England.

The Drue Type B belly bowl form from SwanCove (LT Type 11.1) is short, averaging 1½ in. inheight, with a bulbous bowl. This bowl form has adistinctive broad, flat heel, approximately ½ in. indiameter. So similar is this pipe form to English

Figure 18: Typical decorations on Drue Type A.

Figure 17: Drue Type A (LT Type 15.01).

13

forms of the mid-seventeenth century that it seemsapparent that this pipe was made from a moldbrought from England and in an English traditionof pipemaking. This pipe form, produced in a rangeof white and terra-cotta clays, was recovered atneighboring Providence sites in Anne ArundelCounty, including seven from Homewood’s Lot(n=9) and Burle’s Town Land (n=3) (Figure 19).At first glance, two of the belly bowl fromHomewood’s Lot would seem to be Europeanimports, yet measurements and close inspection ofthe clay indicates local manufacture.

Although Drue used a range of color variations(decidedly not an English characteristic), he limitedthe decoration of belly bowl pipes to very simplerouletting around the rim, like most contemporaryEnglish examples. Exceptions include a recoveredheel featuring a wheel stamp, much like a tradi-tional English maker’s mark, and a white bellybowl fragment with the same stamped design on itsside. These two fragments may belong to the samepipe. Minimalism and conformity to Old Worldvocabularies appear to have been the predominatedecorative traits of Drue Type B.

It is important to note that when Drue Type Bpipes are produced in a hard, snow-white variety,they are virtually indistinguishable from Europeanforms. Before the investigations at Swan Covethese pipes would have been considered exportvarieties without question.

Since the two primary pipe forms were recov-ered from generally contemporaneous contexts, it isnot possible to determine if one pipe form predatesthe other. Clearly, there is a higher level of produc-tion for the Drue Type A form on site, suggestingthat this form was more in demand in the market-place.

The belly bowl forms recovered challengeassumptions held about European preferences andtrade networks. The preference for Drue Type Apipes suggests a transition from European tastes toa new native form of pipe--one that is being locallyproduced, marketed, and sold by colonists. Thisraises the questions of when and why Drue devel-oped this new, form which resembles a trade pipe.Was the antecedent for the form and decoration onDrue Type A pipes based on his observation ofNative American pipes? Did this pipe’s forminfluence the American Export boom that domi-nated the pipe trade through the eighteenth cen-tury?

Drue Specialty PipesIn addition to the two standard types of pipes,

Emanuel Drue apparently felt the occasional needfor greater artistic expression. This is evidenced bythe recovery of at least two unique, handmade

Figure 20: Drue Type C (LT Type 20.01) with a uniquestamp on the heel recovered from Burle’s Town Land.

Figure 19: Drue Type B (LT Type11.1).

14

pipes that appear to be presentation pieces, bearingwitness to Drue’s considerable abilities as a decora-tor. This claim is supported by the recovery of aunique example of what has been labeled DrueType C from the home of Drue’s neighbor, RobertBurle (Figure 20). This bowl form features a largeheel with an unusual dot, circle, and diamonddesign. The exact clay stamp tool used in makingthis design was recovered from the Swan Cove site,

making it possible to identify this handmade pipebowl as a Drue Type C (LT Type 20.01).

Drue’s most impressive example of specializedproduction is seen in the Drue Type D pipe (Figure21). The Crumn horn pipe is fashioned to resemblea horn—perhaps that of various archaic Europeancattle breeds. Detailed comparison of the Crumnhorn length and dimensions with a complete DrueType A reveals a very high statistical possibility (96percent) that the horn pipe originated in the DrueType A mold and while still green and soft, and washandcrafted into an elaborately shaped form. Acombined moldmade/handfinished technique is notuncommon according to Allan Peacey, yet often thehandmade characteristics overshadow the

moldmade features (Allan Peacey, personal com-munication, 2001).

Drue decorated this piece with multiple stampsand rouletting, and it clearly represents a decora-tive extreme. The Drue Crumn horn shows theresult of at least 94 individual hand actions involv-ing six different tools. In addition to the individualstamps, three motifs are created by the combineduse of these tools. This is clearly not an economic

method with which to approach themanufacture of what was, after all,an inexpensive and highly fragileitem. This fact is particularlynoteworthy at a time when thevalue of labor in the Chesapeakewas so high that its expense wasconsidered a driving force in theeconomy, influencing aspects asdiverse as housing construction andthe advancement of the slave trade(Carson et al. 1988). This ex-tremely unusual pipe has only oneknown analogy recovered inHolland, which now resides in thePijpenkabinet museum inAmsterdam. The Dutch example--

nearly identical in shape, size, and decorative style--has been assigned a general date of ca. 1650.

It seems more than plausible that Drue mod-eled his Crumn horn after a very similar pipe,perhaps one he had seen before emigrating to theNew World. Until more is learned of Drue’sactivities before he came to Providence, how Druecame to make a Crumn horn pipe could be thesource of endless speculation. As an expression offolk art, Emanuel Drue’s Crumn horn pipe muststand near the pinnacle of his body of work.

A Non-Drue FormA final terra-cotta form, represented by at least

four heel fragments, has been recovered from

Figure 21: Drue Type D, the Crumn horn pipe.

15

Swan Cove, however it does not appear to havebeen one of Drue’s products. This form has aweak, oval-shaped heel and is made of distinctivegrainy, brownish clay. No complete bowls withthis characteristic have been recovered from SwanCove, but were abundant at the Broadneck site,which is narrowly dated to the 1650s. Along withthe four recovered from Swan Cove, the compa-rable pipes from the Burle’s Town Land site, and anearly (1660-1665) sealed feature fromHomewood’s Lot, this form has been classified asLost Towns Type 11.2, and is often referred to as a“Broadneck” pipe.

Pipes with this form have been recovered fromPope’s Fort, a Puritan-occupied site in St. Mary’sCounty that dates to the 1640s (Miller 1991). Thistype of pipe has also been recovered in abundancefrom a southern Virginia site. It appears that thisform originates near the Chesopean site (1630-1650), south of the James River in Virginia (TaftKiser, personal communication 2001).

Considering the broad regional distribution andthe Puritan associations with each of these sites, theBroadneck pipe appears as a possible marker forthe Puritan Diaspora, tracing the migration ofPuritans who originate in Southside, Virginia,travel through St. Mary’s County and settle in AnneArundel County’s Providence. The discovery ofthis form at Swan Cove supports the assertion thatthe Drue family occupied the site during the early1650s despite the scarcity of archival evidence. Asfar as can be determined, Drue’s bowl form is notfound in deposits dating after ca. 1660.

Pipe Forms and DecorationsAcross the site, the frequency of Drue Types A

and B pipes are skewed decidedly toward the DrueType A form. Of 1,171 pipe fragments recoveredfrom Features 7 and 19, 113 bowls and jointfragments are Drue Type A, while only 18 bowlscan be attributed to Drue Type B. The two primary

pit features present variations on these ratios ofoccurrence, with 84 Drue Type A pipes to only 5Drue Type B pipes recovered from Feature 7 (94percent Type As). Feature 19 has a notably higherfrequency of Drue Type B forms (n=13, or 69percent), though the English-derived belly bowl isstill in the minority at Swan Cove. As noted above,the Drue Type A form exhibits more decoration,with frequent stamping and rouletting as theprimary decorative motifs. The Drue Type B formis by comparison, sparsely decorated. With only afew exceptions, this type B form has only simplerouletting, usually on the rim, as decoration.

The remainder of the pipe forms recoveredfrom excavations immediately adjacent to Features7 and 19 include 1 LT Type 1.00 (heeled bellybowl), 1 LT Type 2.00 (straight-sided heeled), 2 LTType 5.00s (trade pipes) and 4 LT Type 11.2s (aterra-cotta “Broadneck” style).

Stem Bore DatingTerra-cotta pipe stem bores were analyzed

primarily to discern if European pipemakingtechniques and tools were being used in Drue’sproduction. The possession of brass pipe moldswould suggest that Drue learned his pipemakingcraft in England. It is very likely that Drue’slocally produced pipes might produce evidence forthe use of European-made tools, such as wires forboring. If this was the case, a clear correlationbetween the pipe stem bores for comparable whitepipes and for the terra-cotta pipes should beevident. The Binford date for both features was1683, at least 14 years too late for the Drue occupa-tion. Interestingly, the Hanson dates for eachfeature provide a better correlation to the expectedoccupation date based on archival data. TheHanson date for Feature 7 was 1670 (+/- 29 years)and 1682 (+/- 24 years) for Feature 19. In bothcases, using the lower extremes of the deviationprovides the more accurate date.

16

Such analysis may be useful in assess-ing the “relative” bore stem of Chesapeakepipes in relation to trends seen in Euro-pean-manufactured pipes. Should consis-tent patterns be found, it may be possible todevelop a bore stem dating formula morespecific to locally-made or Chesapeakepipes. Seth Mallios and others (2001) haveexplored such speculation when consider-ing Chesapeake or “Colono” tobacco in theJames River Valley of Virginia.

Color Variations in Drue PipesAll Drue products have been found in

an extraordinary array of colors. Theserange from black, slate gray and dark brown to tan,orange, pink, off-white, and pure white (Figure 22).It was first assumed that this diversity of color wasindicative of a lack of temperature and oxygencontrols in Drue’s kiln. These variables can makeidentical clays fire to different shades and hues.However, as intact trash deposits were excavated,and numerous lumps of discarded, unfired clayrecovered, it became apparent that Drue wasdeliberately experimenting with the production ofpipes from different colored clays. Slate gray,green, white, yellow, pink, and variegated pink/white varieties were encountered, and soon theexpression “gourmet clay” was prevalent amongthe excavators. More about Drue’s production andpossible market can be gleaned when considering

the variations between pipe colors found in Fea-tures 7 and 19 (Table 1).

The most common color found in both featureswere pipes in traditional terra-cotta colors, orangeto brown in hue. The next most common colorpresent in both features was buff, representingnearly 19 percent of the combined feature totals.Interestingly, Feature 7, with a predominance ofDrue Type A forms, also had more buff-coloredpipes. Next, in frequency overall, are agated pipes,followed closely by white-clay pipes in Drueforms, which are often indistinguishable from theclay used in European counterparts.

White pipes accounted for almost 16 percent ofthe overall color selection. Next in color frequencywere dark gray pipes with extremely hard bodies.

Figure 22: Range of pipe colors from Swan Cove.

17

These pipes appear to have been overfired, whichcould account for their extremely dark appearance.The original clay color is difficult to determine.One might speculate that the dark gray pipes werewasters—thus,not desirable in the Providencemarketplace. The final, and least frequent, color ofpipe found at Swan Cove is pink—some could bedescribed as “bubble-gum” pink. This pipe colorhas no precedence that we can identify and obvi-ously represents a unique aesthetic.

Agated PipesDrue used the various clay colors available to

him in surprisingly artistic ways. Drue did morethan simply experiment with different color clays,he also demonstrated a proclivity to mix them inorder to achieve an agatized effect, slip them with athin wash of different colored clay, and decoratethem with trailed clay slips. Whether this repre-sents the mind of a scientist or the sensibilities ofan artist remains a matter of perspective.

Beyond what might be considered incidentalswirling or agating of pipes, it appears that Drueused agating as a conscious design motif—devel-oped for an aesthetic, not simply as a by-product ofclay availability. Nearly 22 percent of the pipesrecovered from Feature 7 and 8 percent of the pipesrecovered from Feature 19 are agated. The produc-tion of pipes in mixed colors is a deliberate attemptto use the color variations in the local clays fordesign purposes. Several pipes have been recov-ered which have a distinctive striped effect, pro-duced by layering clays in a spiral around the stem(Figure 23). This design might be best described asa “barber pole” design. As far as is known, there isonly one other tobacco-pipemaker on either side ofthe Atlantic who even attempted anything similar.

Currently the subject of investigation by TaftKaiser, in cooperation with the Lost Towns Project,this producer has been called the “SouthsideMaker” or “Bookbinder” in reference to his elabo-

rate decorating tools (Figure 23). Apparentlylocated somewhere near the Chesopean site inVirginia Beach, Virginia, the products of this makerhave been recovered over a large area of theChesapeake. Lacking raw materials, the Southsidemaker did not produce pipes in pinks and greens,but he frequently did agatize his pipes by mixingclays, and is the only pipemaker--other than Drue--to sliptrail his pipes with barber pole decoration.

The broad and distinctive variations of pipecolor may represent production in response to awell-developed and sophisticated aesthetic byDrue’s patrons.He is producingvariations in hisproduct toanswer the needsof a sophisti-cated market-place whileexploring anartistic endeavor.Either due toclay availabilityor local prefer-ence, the tradi-tional orangehue of terra-cotta is the most desirable productcolor, followed closely by buff and white pipes.The dark gray, agatized, and pink pipes are some-what less in demand or address more specializedProvidence pipe preferences. When consideringthe market Drue was addressing and the decorativevariations found across the site, he is clearly one ofAnne Arundel County’s first artists. The scarcity ofDrue’s products on nearby sites in comparison tothe extensive pipe-producing capabilities found atSwan Cove raise the question of where Drue’spipes are being sold. Obviously there are seven-teenth-century markets in the mid-Chesapeake Baythat have not yet been identified archaeologically.

Figure 23: The trail-slip“barberpole” decoration.

18

ConclusionThe first, and perhaps most important, aspect of

the Swan Cove kiln site, is simply its discovery.The local production of tobacco-pipes in the NewWorld has been the subject of intense interest tohistorical archaeologists in those rare locales whereit occurs. Nowhere is this more true than in theChesapeake Bay region of Maryland and Virginia,where local pipes have been the focus of numeroushistorical theories, and where such a productionsite has been sought for decades. It had long beenassumed that when such a site was found it wouldrepresent little more than the waste products of aminor cottage industry. Unlike the more formalenterprises in England or Holland, Chesapeakeproduction was considered to be more likely thesimple output of individuals, often lacking even amold to produce the pipes, and probably firingthem in their home fireplaces without the benefit ofa kiln. Occasionally, evidence supporting suchactivity has even been recovered.

Swan Cove proved startlingly different thanthis conception. Emanuel Drue clearly possessed akiln the equal of its European counterparts, capableof reaching extremely high temperatures. Hismuffles and kiln furniture derived from contempo-rary English examples, and indicate his nationalorigin and/or training. He possessed English-stylebrass molds, and decorated his pipes with a largevariety of homemade tools. Given a source of purewhite clay, he was capable of producing pipes thatwere virtually indistinguishable from exportvarieties produced in London and elsewhere inEngland.

But Drue was not producing tobacco-pipes thatwere simply intended to copy the varieties madeback home. Something more seems to have been atwork here. In addition to using the terra-cotta claysavailable beneath his feet, Drue made a concertedeffort to gather an incredible range of “exotic” local

clays. A mix of scientist and artist, Emanuel Drueexperimented not only with different colors, butalso blended them, applied slips to them, anddecorated them with trailed clays and homemadetools.

Particularly interesting is the tendency toagatize and use trail-slip “barber pole” decorationseen in the pipes found in abundance at theChesopean site in Virginia. The fact that Southside,Virginia is the region from which most of thePuritans migrated to the Providence settlement inMaryland is highly intriguing. Since EmanuelDrue is presumed to be part of this migration, thepossibility of a connection between Drue and theSouthside maker seems easy to imagine. Unfortu-nately, no historical documentation yet ties Drue tothis region.

Another important factor that has come to lightin the investigations of Swan Cove has to do withDrue’s products. The vast majority is either theDrue Type A angular elbow/Chesapeake form orthe Drue Type B European-style belly bowl. Thefact that the Drue Type B pipe receives simpleEuropean-style decoration, while Drue Type A pipeis more likely to have elaborate Chesapeake-styledecoration, clearly implies two different marketsare involved. Intuitively, one would assume thatsome individuals were more interested in smokingpipes that mimic the English exports, however, thiscould not be the explanation. Both types of Druepipes occur in pure white, but only rarely. Sincemost of the belly bowls are terra-cotta colors, theysimply would not be mistaken for English products,despite their shape.

An outgrowth of the Drue Type A/B dichotomydiscovered in Emanuel Drue’s output has been atheory that similar dichotomies existed in theproducts of Virginia pipemakers. Taft Kaiser’sVirginia data seems to support this theory, includ-ing the propensity to decorate belly bowls in amore conservative, traditional fashion. However,

19

until the discovery of Swan Cove, it was assumedthat the many angular elbow and belly bowl groupsfrom Virginia were from different producers.

Finally, the Swan Cove results do extensivedamage to a number of fondly held theories con-cerning the Chesapeake pipe. The early debateconcerning European versus Native Americanorigins for these pipes was eventually replaced withseductive suggestions that African American orCreole populations produced them. Since givinghistorical voice to unheard populations like slaves

is an understandably important aspect of colonialarchaeology, researchers welcomed a readilyrecognizable marker in the artifactual record.

All available evidence, however, would indi-cate that Emanuel Drue, an English Puritan, wasmaking the terra-cotta Chesapeake pipes at SwanCove. Furthermore, rather than representing apotential anomaly, Drue clearly fits into a fairlywidespread Virginia pattern of peers. They notonly share dichotomous bowl forms, but also aclear decorative grammar involving alternatingrouletted panels and stamping.

Even the theory that masters were smokingwhite pipes while servants and slaves smoked terra-cotta ones is done a certain degree of harm by the

Swan Cove results. First, the Drue Type A/Bdichotomy discussed earlier, if anything, seems toargue that color is less significant than shape.Second, the two most elaborate and labor-intensivetobacco-pipes produced--perhaps even meant as“presentation pieces”--by Emanuel Drue were bothdone in terra-cotta.

One critical missing link is the lack of Drueproducts recovered elsewhere. So far, Drue’s pipeshave only been recovered from the sites of Drue’simmediate neighbors in Providence. The level of

pipe production points toward a substantialoperation and corresponding output (Figure24). Where, then, is Drue selling his wares? Aclear possibility is that he marketed his pipes inregions that have received little or no archaeo-logical investigation. Notable possibilities lieto the east, on Kent Island or the Eastern Shoreof Maryland. If Drue pipes can be identified inlocations more remote from the kiln, it will besignificant to the evaluation of local tradenetworks in early colonial Maryland. Further-more, finding Drue’s pipes on other sites couldelucidate questions about the pipes’ productionand demand in the marketplace. Does Druemake each form of pipe in response to the

needs and requirements of his contemporaries? Ishe an artist developing this form with little interestfor or from the market place? Is having a pipe thatis similar in so many ways to a Native Americanpipe form a statement by the colonists that they arein a new and different land? Do the belly bowlforms meet the needs of new European immi-grants?

One thing that is clear is that Emanuel Drue isadapting Old World manufacturing techniques withNew World influences as an outlet for distinctexpression. If the clay tobacco-pipe is the highestmanifestation of folk art found in the seventeenth-century Chesapeake, as maintained by Dan Mouer(1993), then Drue must stand among its masters.

Figure 24: Pipe asssemblage from Feature 7.