sign in sign up
<<
Home , Danish West Indies, Hill people, Maroons

Journal of Archaeology Copyright 2007 ISSN 1524-4776

THE CHALLENGE IN LOCATING MAROON REFUGE SITES AT MAROON RIDGE, ST. CROIX

Holly K. Norton Department of Anthropology Syracuse University 209 Maxwell Hall Syracuse, 13244-1090 [email protected]

Christopher T. Espenshade New South Associates, Inc. 415-A South Edgeworth Street Greensboro, 27401 [email protected]

The hideouts, lookout points, temporary camps, and concealed communities of runaway slaves may be difficult to locate using traditional methods of archaeological survey. These lo- cations were intentionally made inconspicuous, were likely kept clean of surface refuse, and may have been placed in atypical landscape settings. As well, these sites may be small in size and likely contain only a few durable goods. A typical archaeological survey that combines screened shovel tests on a 20 m interval with surface survey for structural features is not well suited to the discovery of Maroon refuge sites. If these important resources are to be discov- ered, typical methods should be augmented with a GIS-based consideration of locational fac- tors and controlled metal-detector survey. ______The Maroon, or runaway slave, plays a The first holds that the area – known broadly large role in defining the cultural identity of as Maroon Ridge or Maroon – many African Caribbean inhabitants of the served simply as a path or conduit for run- former Danish . The modern in- aways hoping to catch a boat to . habitants of St. Croix, St. John, and St. Tho- The alternative position is that there was a mas proudly point to their unique history of significant Maroon population living perma- resistance to , including marronage. nently in the area from circa 1650, when the They know the landscape of resistance, and island was occupied by the French, until the many island locations are remembered as im- 1760s, by which time the extensive develop- portant in the Maroon experience. ment of the Danish sugar economy on St. Paradoxically, there is little good evidence Croix probably rendered hiding in the bush for the exact nature of Maroon activities in nearly impossible. A similar trajectory was many areas of the islands. For northwestern followed on , where Handler and St. Croix, the geographical focus of this pa- Lange (1978:144) note “Although Barbadian per, there are two general schools of thought. slaves escaped and sought refuge in caves

Journal of Caribbean Archaeology 7, 2007 1 Locating Maroon Refuge Sites Norton and Espenshade and forested areas, by the second half of the recapture and punishment (i.e., Maroon seventeenth century the island was fully oc- refuge sites), in contrast with Maroon cupied, the forests were mostly removed, and communities that were tolerated or condoned there were no opportunities for the establish- by the Euro-. Furthermore, in ment of maroon communities.” considering the nature of Maroon refuges, it is evident that typical archaeological survey How were the using the landscape methods may fail to properly find, recognize, of northwestern St. Croix? This seems like a and interpret such sites. pretty basic question to remain unanswered at this late date, yet it is an important question We offer a preliminary paper at this in understanding slave resistance. As juncture, rather than awaiting field results, discussed more completely below, the because Maroon Ridge is under imminent archival record has little to say to resolve this threat. Resort development has begun to question, and it is time for historical squeeze the area from the south and the east. archaeology to step to the forefront. Is Hoping to counter unchecked development, a archaeology prepared for this task? local consortium has been pushing for the purchase and development of the area as a The present paper is the result of an territorial park to commemorate the Maroon evolving proposal to archaeologically study experience on St. Croix (Maroon landscapes sites of Maroon refuge activity in the Maroon are under threat in many countries. For Ridge area of northwestern St. Croix (Figures example, see Price and Price 2002). Under 1-2). In reviewing the archaeological either scenario, it will become necessary to literature, it became clear that little work has assure that the Maroon sites are discovered, been done on Maroon camp sites, where the recognized, and properly treated. Maroons were under persistent threat of

Figure 1. Portion of the 1754 Map of St. Croix (Beck 1754). Note that Maroon Ridge area is labeled as “Uoptagne Grunde” or Unrecorded (unclaimed) Ground.

Journal of Caribbean Archaeology 7, 2007 2 Locating Maroon Refuge Sites Norton and Espenshade

Figure 2. Portion of a 1942 Navigation Chart (U.S. Department of Commerce 1942) Show- ing Topography and Place Names, “Maroon Ridge” and “Maroon Hole.”

Historic Context individuals banded together, sometimes with Native American groups, to form permanent Marronage existed, in various forms, in settlements. Arguably, the most famous of every slave-holding society in the Western these is , in northwestern , Hemisphere. Although Marronage is which existed for nearly 100 years before classically associated with runaway African finally being crushed by British mercenaries slaves, Indians were often the earliest and Native Americans, after persistent Maroons in the Caribbean (e.g., Yaremko Portuguese attacks (Allen 2001; Orser and 2006). Generally the acts of these self- Funari 2001). and Nannytown, liberating individuals, commonly referred to both located in , were also highly as “Maroons,” is categorized in one of two successful Maroon communities. The ways: petite marronage, also referred to as Maroons of these communities were so truancy or absenteeism, where an individual successful that they were able to negotiate left their or other place of treaties with the British government to ensure enslavement for a short period of time, their continued existence (Agorsah 1993, tending to return on their own; or grand 1994, 2003; Bilby 1997; Kopytoff 1978, marronage, an act of permanent escape (Price 1979; Schuler 1970). Despite these 1979). In the case of grand marronage, examples, most Maroon communities were ethnohistoric evidence suggests that ephemeral and short lived, under constant

Journal of Caribbean Archaeology 7, 2007 3 Locating Maroon Refuge Sites Norton and Espenshade pressure from the militarily dominant of the fruit is sweet and acidic in taste.) A European societies in which they existed. resident of St. Croix, writing a decade before Oldendorp’s visit, noted that “…planter Marronage within the families were being ruined by the running was similarly varied. Although the historic away of slaves in groups of as many as documentary evidence hints at these semi- twenty to twenty-five in a single night” (Hall permanent internal settlements on St. Croix 1985: 485). and St. Thomas, most of the activity centered around escape from the islands, along what Oldendorp provides details on the Maroon Hall has identified as the “marine Hill people: underground” to Puerto Rico, Vieques and (Hall 1985: 482. See also Chinea For a long time now, a large number of [1997] for an evaluation of the archival these Negroes have established them- record on marronage to Puerto Rico.), islands selves on lofty Maroon Hill in the moun- held by European powers that were often tains toward the west end of the island. hostile to Danish policy. The Maroons that In addition to the lay of the land, they were able to eke out an existence on the small are there protected by impenetrable bush Danish holdings (the islands of St. Croix, St. and by their own wariness. They keep Thomas and St. John fall short of 200 square every approach safe by attempting care- miles combined, the bulk of that retained by fully to conceal small, pointed stakes of St. Croix) did so at a place identified as poisoned wood so that the unwary pur- “Maroon Mountain” or “Maroon Ridge” on suer might wound his foot on them and St. Croix (Pope 1972). therefore be prevented from continuing the chase as a result of the unbearable Maroons of northwest St. Croix were pain. . . . For those foods that they can- discussed by Oldendorp, a Moravian not obtain in the wild, they must search and visitor to the Danish West at sea at night, exposing themselves to Indies in 1767-68. Oldendorp left a detailed life-threatening dangers in the process; account of island culture, including a brief or they can steal them from . description of “Maroon Hill”, which he On St. Croix, they are so bold that they describes as “almost impassable” (Highfield often venture down from their hills dur- and Barac 1987:51). He further notes that the ing the day and go into the Negro mar- Maroons on St. Croix rely on rainwater kets in order to procure the necessities. caught in rock crevices or basins for their It is not at all easy to identify them drinking water (Highfield and Barac among the great numbers of Negroes in 1987:53). Oldendorp (Highfield and Barac the market [Highfield and Barac 1987:106) reports that the fruit of the susack 1987:233]. tree was a major subsistence item of the Maroons, who “often live exclusively on This passage is interesting in suggesting an them” (Susack trees, Annona murciata, grow economic relationship between plantation/ to 25 to 30 feet in height, and produce fruits market Afro-Cruzans and the Maroons. The up to 15 pounds in weight. The fruit has an Maroons must have been offering something inedible skin with stubby spines, but the flesh – possibly wild foodstuffs – in exchange at

Journal of Caribbean Archaeology 7, 2007 4 Locating Maroon Refuge Sites Norton and Espenshade the market. The market contacts also suggest runaways themselves, and in St. Croix that kin ties may have been important in sup- there was a mountain-hideout called plying Maroons. ‘Maroons’ Hole’ just east of Hamm’s Bluff, where hideaways were safely Oldendorp (Highfield and Barac 1987:234) hidden in a cave whose entrance was notes that the Maroon problem was often protected by poles of poisonous wood, addressed through organized hunts for the until they could be transferred to Puerto runaways, yet states “hunts such as these, Rico [Dookhan 1994:164]. however, are not organized to track down those who remain in the high Maroon Hills of An 1828 reference does little to resolve the St. Croix.” Noting “a large number of these nature of the use of Maroon Mountain. The Negroes” on Maroon Ridge, Oldendorp is the Church Missionary Society (1828:619) best champion of the permanent population reports “among them lies the so-called school of thought. Maroon Mountain, where a few run-away Negroes still hide themselves.” Dookhan (1994) offers a somewhat different interpretation of Maroon Mountain: Maroon Ridge remains a historically significant location to St. Cruzans. It has Runaways never comprised a permanent generally remained a rugged, remote place body in the such as the from the seventeenth century through today. maroons in Jamaica, for when the slave- It is mentioned on heritage tours of the island, hunt became too successful, the slaves retains key place names, and is a source of escaped to Puerto Rico. That island had local pride (e.g., Voight 2006). not yet developed a plantation economy and the treatment of slaves there was The archaeological study of Maroon sites is relatively mild. Besides, runaways were important for scholars of the African usually employed on works of Diaspora (Weik 1997, 2004). Not only fortification on the island for one year, would it provide information regarding the after which they were pronounced free internal social structure of maroon and given a plot of land to cultivate. communities themselves, but archaeology Slaves escaping to Puerto Rico became may also shed light on questions concerning lost to the Virgin Islands slave-owners, a identity, agency, creolization, and internal loss which was more strongly felt since economies, to name just a few arenas of only the most robust slaves were scholarly focus. prepared to hazard the dangers of the 40- odd miles of ocean separating the Because the archival record is surprisingly Danish islands from the much larger quiet (and conflicting) on the nature of Spanish island. The numbers of Maroon activity in northwestern St. Croix, runaways were apparently large since for and because paradoxically that activity is 1745 alone it was estimated that about important to the cultural identity of African 300 slaves from St. Thomas and St. Cruzans, it falls to archaeology to address the Croix had escaped to Puerto Rico. The nature of the Maroon existence. traffic became highly organized by the

Journal of Caribbean Archaeology 7, 2007 5 Locating Maroon Refuge Sites Norton and Espenshade

Previous Archaeological Research on groups used the rugged, minimally accessible Maroon Sites landscape of Maroon Ridge for their refuge.

Although there has been much progress in It is anticipated that refuge camps in areas the historical, anthropological and such as Maroon Ridge on St. Croix will have archaeological study of Maroon culture over the following traits: the past decade (Agorsah 1993, 1994, 2003, 2006, 2007; Allen 2001; Bilby 1997; Camp 1. Site locations will have been selected with 2002; Forbes 1992; Hall 1985; Kopytoff concealment in mind. Rather than settling 1978, 1979; La Rosa Corzo 2006; Lokken on a broad, open ridge top, the Maroons 2004; Orser and Funari 2001; Price 1979; would have selected an obscure, secluded Weik 1997, 2004) the literature is generally cove. Rather than settling near a major devoid of archaeologically examined Maroon trail, road, or landing, the Maroons would refuge camps (c.f., Vega 1979; Garcia have chosen locations with difficult ac- Arevalo 1986; Maris- 2002, La Rosa cess. In this regard, Maroon camp loca- Corzo 2003, 2005, 2006). Instead, tions might be considered analogous to archaeologists have generally investigated illicit whiskey stills in the southeastern stable communities that were established by . Maroon groups, and that were tolerated by 2. Site locations would have been chosen the dominant Euro-Caribbean culture with defensibility in mind. The site itself (Agorsah 1993, 1994; Allen 2001; Orser and or the access to the site would have been Funari 2001). Such Maroon villages can selected to provide strategic advantage to often be discovered and examined by typical the Maroons. The superior weaponry of methods of historical archaeology because slave-hunters would mean little if the Ma- the villages are typical residential sites. roons were pelting them with rocks from These villages often had a rich material concealed locations 50-100 ft above the culture including permanent structures, refuse trail. Military considerations would have middens, and cleared horticultural plots. played a key role in selecting site loca- tions. In contrast, refuge camps occupied by small 3. Due to 1 and 2, Maroon refuge sites numbers of Maroons living under threat of would not have been located on the land- recapture and punishment are expected to forms targeted by normal archaeological have different characteristics (see La Rosa survey. If such sites are sought by an ar- Corzo 2005 and 2006 for examples of this chaeologist used to searching for planta- type of site in western ). It is reported tions, slave villages, and similar historic that the Maroon Ridge area of northwest St. sites, the survey could easily miss many Croix harbored Maroons from 1670 through or all slave refuge sites. at least 1767. The limited archival record 4. Related to 1 and 2, Maroons would have suggests that these Maroons would have been made a concerted effort to reduce their recaptured and severely punished if their signatures on the landscape. Refuse recapture was easy. To make their capture would have been policed and buried or difficult to impossible, the Maroon Ridge dumped into the sea. A general residen-

Journal of Caribbean Archaeology 7, 2007 6 Locating Maroon Refuge Sites Norton and Espenshade tial midden would not be created. Useful rather coarse, U.S.G.S. topographic maps. items would be cached out of view. Sub- Shovel testing is typically augmented by a stantial houses would not have been con- surface search for structural features (e.g., structed. stone or brick foundations). 5. Depending on the amount of interaction between the refuge Maroons, enslaved The method of shovel testing on 20-meter African Caribbeans, freedmen, and others intervals is premised on the targeted sites be- (e.g., pirates), the Maroons may have had ing greater than 20 meters in diameter. This limited material possessions. Maroons may not be the case for Maroon refuge sites. may have preferred to have limited mate- Five Maroon sites documented in western rial goods, to allow their rapid abandon- Cuba by La Rosa Corzo (2005:166) measured ment of sites. Artifact counts may be 3 x 4 meters, 14 x 5 meters, 1 x 1 meter, 13 x relatively low at refuge sites. 5 meters, and 4 x 3 meters, respectively. Ar- 6. Due to lack of building materials and risk rom and García Arévalo (1986) also com- of loss to slave hunters, the Maroons ment on the small size of Maroon refuge likely utilized indestructible, ready-made sites. rock shelters or caves for many of their sites. Even if the domestic site was dis- The method of shovel testing at 20-meter covered and the Maroons were forced to intervals is also premised on there being a flee, the slave hunters could not burn sufficiently dense midden deposit to assure down or raze a natural overhang. In west- discovery of the site in shovel tests spaced at ern Cuba, La Rosa Corzo (2005:165) re- 20-meter intervals. As argued above, Ma- ports that “within these elevations, all the roon refuge sites probably did not have sheet sites consist of overhangs and caves that middens and may not have had sufficient arti- served as temporary shelters to isolated fact density (Figure 3). Furthermore, the vast groups of cimarrones.” Vega (1979) also majority of artifacts at such sites will be con- addressed the use of rock outcrops, and centrated either in caches or refuse pits. Barnet (1993) provides an account of a Shovel testing is especially ineffective in dis- Cuban Maroon who lived for years in a covering relatively small features. cave. Maroon refuge camps will be small, have Implications for Archaeological Survey low artifact density, and have severe artifact clustering. These attributes represent three The anticipated traits of Maroon refuge strikes against the applicability of shovel test sites have implications for the selection of survey (termed “TPS” or “Test Pit Survey” archaeological survey methods. A typical by Nance and Ball 1986. See also Orser and compliance survey (within the U.S. Section Funari 1992 and Weik 2004 for discussions 106 process) utilizes shovel testing on a 20- of shovel testing on Maroon sites). Follow- meter interval. Shovel tests are generally not ing a thorough review of the mathematical excavated where the landforms are sloped premises of such survey, Nance and Ball more than 20-30 percent, and these slope de- (1986:479, emphasis in original) conclude terminations are often made on the basis of “The net effect is that TPS is biased against

Journal of Caribbean Archaeology 7, 2007 7 Locating Maroon Refuge Sites Norton and Espenshade

Figure 3. Maroon in (from Benoit 1839). Note limited material goods and simple abode. discovery of small, low density sites, espe- “the tactic of falling back before slave hunt- cially when these sites exhibit marked spatial ing militia arrived prevailed throughout the clustering of artifacts. TPS is also unreliable area” (La Rosa Corzo 2003:104) and “it was in such contexts.” a setback when slave hunters discovered a settlement, destroyed their crops, and burned There is little gained at Maroon refuge sites down their huts, but the runaways were able by the surface search for structural features. to recover quickly and easily” (La Rosa The Maroons would have avoided the very Corzo 2003:236). This strategy of dealing types of features sought by the archaeolo- with the omnipresent risk would have in- gists. A rock cistern or a brick hearth would cluded the use of expedient (sensu Binford have drawn attention to the refuge during oc- 1979) structures. In the absence of suitable cupation and would have been a prime target rock overhangs or caves, La Roza Corzo for destruction by the slave-hunters. (2003:245-246) reports that “dwellings of the first type – that is, low, thatched-roof huts Furthermore, it appears that the Maroon with dirt floors – seem to have been more strategy was often to abandon a settlement commonly used and more convenient in whenever there were incursions by slave places where the conditions did not facilitate hunters. La Rosa Corzo reviewed extensive the development of a long-lasting, safe settle- slave-hunter records for Cuba, and concluded ment.” Such structures would have left an

Journal of Caribbean Archaeology 7, 2007 8 Locating Maroon Refuge Sites Norton and Espenshade extremely light archaeological signature, should be completed by archaeologists with probably with no surface features (but possi- extensive experience in metal-detecting. bly with sub-surface post holes). Survival in the Caribbean forests and shore- It is difficult to know what percentage of line depended heavily on metal objects. La Maroon refuge sites would be missed by a Rosa Corzo (2005) found machetes, a hoe, a standard archaeological survey. However, if shackle, and buttons in his test excavations in our expectations about such sites are correct, Cuba. Slave-hunters from Cuba (La Rosa the potential loss of information is signifi- Corzo 2003:104) reported Maroons “armed cant. with machetes, knives, and five “nearly use- less” shotguns.” An account from An Improved Approach in 1667 (Baptiste Dutertre 2004 [1667]; Eng- lish translation by Espenshade) emphasizes In addressing Civil War and other military the importance of metal items: sites, Espenshade et al. (2002) noted that tra- ditional survey methods were poorly suited Los segundos al estar más acostumbra- for finding camps, picket posts, and skirmish dos al país, no se convierten jamás en locations, for many of the same reasons de- cimarrones más que cuando han puesto lineated above (see also Connor and Scott orden en sus asuntos; es para ello que se 1998; Fox 1993; Scott et al. 1989; Sterling proveen de herrajes, tales como sierras, and Slaughter 2001). Espenshade et al. hachas, y cuchillos; se llevan sus hara- (2002) argued that survey methodology pos, hacen provision de mijo y se retiran should be changed in areas likely to contain a los lugares más elevadoes de las mon- military resources, and that controlled metal tañas que son casi inaccesibles donde detector survey should be an important ele- abaten árboles, hacen un huerto y plan- ment of such research. Controlled metal- tan mandioca y ñames, y a la espera de detector survey is especially effective for que lleguen a madurar, vienen por la no- military sites because a large portion of the che a los límites de bosque donde los surviving material culture is metallic, and be- otros negros no faltan de llevarles de cause artifacts are likely to be sparsely dis- comer lo que tengan. Cuando no están tributed, except for a few refuse features. seguros, van osadamante en el noche a robar a las haciendas y toman todo lo Clearly, the same applies to Maroon refuge que encuentran; ha ocurrido que han lle- sites. Survey in areas of known or suspected gado a robar hasta la espada y el fusil de Maroon refuge sites should include intensive su amos [Baptiste Dutertre 2004 (1667)]. metal-detector survey. These areas, at least on St. Croix, generally lack deep soils and (The second group, being more accus- have not undergone extensive disturbance by tomed to the country, never become run- plowing, logging, or relic-hunting. This aways before they have put their affairs means that metallic artifacts, even if buried in in order; they gathered implements such refuse pits, should be within the operating as saws, axes, and machetes; they clean depth range of a metal-detector. Intensive their clothes, make provision of metal-detecting of all suitable landforms and retire to the most elevated places in

Journal of Caribbean Archaeology 7, 2007 9 Locating Maroon Refuge Sites Norton and Espenshade the that are almost inaccessi- tary Probability, although without using that ble, where they fold over the trees, make term. La Rosa Corzo (2003:223) reports: an orchard, and plant manioc and yams, and while they await the ripening, they In fieldwork, both archaeological and come at night to the limits of the forest ethnographic aspects were helpful, but in where the other negroes never fail to this phase of the work, obtaining on-the- give them whatever food they have. spot knowledge of geographic condi- When they are not safe, they go daringly tions that favored the founding of run- in the night to rob the haciendas and away slave settlements as a form of ac- they take everything they encounter; it tive slave resistance was most important, has happened that they have stolen even since, in order to understand and explain the sword and the gun of their masters.) this phenomenon, it was necessary to make direct contact with the environ- In their call for better approaches to finding ment in which the incidents took place military sites, Espenshade et al. (2002) also [La Rosa Corzo 2003:223]. called for archaeologists to apply the concept of Inherent Military Probability (as defined The Other Half of the Equation by Hans Delbrück and Alfred Burne. See Keegan 1976). This basically requires the Although it would be great to intensively archaeologist to think like a soldier. Is this metal detect every square meter of Maroon location defensible? How high up the ridge Mountain, this would be cost and time pro- should the rifle trenches be placed? What are hibitive (Figure 4). Luckily, Geographic In- the best observation points? formation Systems (GIS) provide a means of defining those areas with the highest prob- Similarly, the archaeologist searching for ability of containing Maroon sites, while also ephemeral Maroon sites must put themselves defining areas that will not require survey. in the mindset of the Maroon, to the degree The challenge is to translate expected pa- possible. The archaeologist must let go of the rameters of Maroon site selection into vari- typical parameters of site location – level ables that can be derived from available geo- ground, access to a good water source, expo- graphic data sets. sure to cooling breezes, proximity to trans- portation corridors, nice view -- and think There are consistencies in descriptions of like a Maroon. The archaeologist must keep Maroon refuge camps. Price (1996:5-6) in mind concealment, defensibility, and es- notes “to be viable, Maroon communities had cape routes. The archaeologist must abandon to be almost inaccessible, and villages were their concept of what a house will look like, typically located in inhospitable, out-of-the- and which support features (an oven, a cis- way areas. . . . Successful maroon commu- tern) ‘must’ be present at a residence. The nities learned quickly to turn the harshness of archaeologist must abandon the typical ex- their immediate surroundings to their advan- pectation for midden. tage for purposes of concealment and de- fense.” La Rosa Corzo (2003:225) offers this In his work in Cuba, La Rosa Corzo (2003, in agreement with Price: 2005) embraced the concept of Inherent Mili-

Journal of Caribbean Archaeology 7, 2007 10 Locating Maroon Refuge Sites Norton and Espenshade

Figure 4. Image of Maroon Ridge from Google Earth, Showing Ruggedness of the Terrain.

As already stated, the places in which For an example of these consistencies in the runaway slaves chose to settle had to US Virgin Islands, the reader can compare meet the most basic requirements for the Oldendorp description of Maroon Moun- living under attack: distance (as far as tain (see above) with an early eighteenth cen- possible from colonial population cen- tury description from St. Thomas: ters and from means of communication), inaccessibility (that is, they had to be in J.L. Carstens, who was born in St. Tho- locations that were difficult to reach by mas in 1705 and died in in passersby, farmers, and cowhands and 1747, noted in his memoirs that in those that had few probabilities of being stum- early years runaways occupied the is- bled upon), and natural concealment (a land’s coastal cliffs, where they shel- place whose topography and vegetation tered in almost inaccessible caves. offered its protection). These three con- Those first maroons chose well, with a ditions, which often overlapped, corre- keen strategic eye, for the cliffs could sponded to three different – though re- not be scaled from the seaward side and lated – spatial levels [La Rosa Corzo vegetation obstructed the landward ap- 2003:225, emphasis in original]. proaches. Such refugees went naked and subsisted on fish, fruit, small game such

Journal of Caribbean Archaeology 7, 2007 11 Locating Maroon Refuge Sites Norton and Espenshade as land turtles, or stolen provender. ing the use of a predictive GIS model places Slave hunts, organized three times a us squarely within a particular camp (i.e., em- year, could neither loosen their grip on bracing some variety of environmental deter- freedom nor dislodge them from the minism), it is not the intention of the authors cliffs [Hall 1992:127]. to argue the theoretical implications within this paper. A substantial literature exists How does one model inaccessibility? How highlighting the controversy (Brandt et al. does one model a condition described in 1992; Ebert 2004; Llobera 1996; Wheatley Cuba in 1828 as “there was no way for beasts and Gillings 2002; Gaffney and Van Leusen to get through. We went down to it along a 1995; Kvamme, 1997; Kuna 2000, Wheatley stony stream. The descent was very difficult 1998). Instead, our intention is to suggest the and we had to hang on vines” (La Rosa Corzo use of GIS predictive model as an atheoreti- (2003:124)? Perhaps by scoring the land- cal tool, a methodology that can hopefully be scape by travel labor, derived from distance applied to a number of theoretically derived and slope relief. How does one account for a research questions. desire to remain hidden? The tools of viewshed analysis, as widely applied in cell Viewshed analysis requires establishing tower studies, can be used to map which loca- which points on the physical landscape are tions are visible from key points (Bruce Lar- visible from a particular viewpoint on that son [personal communication 2007] reports landscape (Conolly and Lake 2006; Wheatley that suspected Maroon sites on Vieques were and Gillings 2002). It has been used primar- located in the rear portions of caves that were ily to identify the visual landscape of monu- hidden at least 100 m back from the beach. mental structures (Bevan and Conolly 2002- See Sanders et al. 2001, 2003). Where are 2004; Wheatley 1995). However, as dis- potential rock outcrops on Maroon Moun- cussed by Llobera (2006), all landscapes are tain? Local ecologist Olasee Davis knows of inherently visually structured. This is proba- one cave with two compartments on the high bly most true for societies such as Maroon slope and additional caves close to the beach. communities or refuge sites where conceal- Oral history interviews, the soil survey, and ment is central to their survival. Not only Digital Elevation Models can create a signa- would Maroons have been concerned with ture for potential shelters and caves. Upon their own visibility to outsiders, but the abil- survey, potential shelters and caves may re- ity to actively scrutinize activities around veal signs of past use, such as sooting of the them, to anticipate their detection by authori- ceiling. ties would have been a requirement for site placement as well. The use of cumulative The benefits of incorporating a GIS compo- viewshed analysis for the entire area of Ma- nent into an archaeological survey of Maroon roon Ridge and surrounding locales may Ridge would be two-fold. First, GIS can be therefore be an effective tool in establishing used to create a predictive model, allowing areas of priority; locations on Maroon Moun- the researcher to narrow the focus of inten- tain that are masked from below, but that al- sive metal-detector and other remote-sensing low views from the ridge-top should be given survey. GIS allows for a targeted landscape highest priority, those sites with no intervisi- survey. Although we recognize that suggest- bility second priority, and those with full in-

Journal of Caribbean Archaeology 7, 2007 12 Locating Maroon Refuge Sites Norton and Espenshade tervisibility lowest priority. Other factors One of the greatest advantages of GIS to should be considered along with viewshed constructing predictive models for Maroon data, both environmental and cultural, such as refuge sites would be in the ability to con- proximities or paths to markets, to sea and stantly asses and modify the model through- land escape routes, availability of natural out the duration of field work as GIS soft- shelters such as rock overhangs, presence of ware “…enables the visualization of data pat- economic trees and bushes, etc. Although not terns at or soon after their collection… utilizing GIS software and mapping capabili- facilitating a ‘reflexive’ approach to data col- ties, Gabino La Rosa Corzo had great success lection…” (Conolly and Lake 2006: 37). Be- in identifying Maroon refuge sites on Cuba cause we are unsure of the exact nature of looking at a variety of environmental condi- Maroon use of Maroon Ridge, it will be espe- tions as predictors that he wrought from his- cially important to adjust and refine the GIS toric documentary sources (La Rosa Corzo model as initial survey data become avail- 2003, 2006). One the most compelling envi- able. By creating a GIS model, the scope of ronmental consistencies that La Rosa identi- the metal detector survey can be reduced to a fied was the near absence of water sources at feasible level, and the key locations can be refuge sites; only 3 of 30 had springs/water examined intensively. sources, and of those two were seasonal sources, which fits with accounts of refuge The second benefit to a GIS based approach sites on other islands, such as St. Croix (La would come after the field data are in, when Rosa Corzo 2006: 9). Maroons may well we can begin analyzing and explaining the have avoided settling close to springs because relationships of Maroon sites across the water sources were the first place a slave physical and social landscape of St. Croix. hunter might look. Testing this hypothesis For instance, was there a change through may fit well with an agent-based model of the time, say between when the French held the type created by Lake and Woodman to test island versus when the Danish occupied, in the importance of hazel-nut collecting to an- the character of maroon sites? Did multiple cient foragers (Conolly and Lake 2006; Lake sites exist simultaneously? The creation and 2000a, 2000b; Lake and Woodman 2000). testing of the model provides an interpretive This model provided agents with “…a set of tool for better understanding Maroon strate- goals, decision-making abilities and risk- gies. Once a GIS model has been developed taking parameters…” along with information and tested, the key variables can be used to on the environment and known archaeologi- assess the potential for Maroon refuge sites cal site and artifact patterns (Conolly and on St. John and St. Thomas, as well as ques- Lake 2006: 49). A compelling hypothesis, tions concerning the spatial relationships be- and one that may be incorporated into an tween Maroon sites and plantation or other agent-based model, is that self-liberating institutional sites across the island, and the slaves may have purposely avoided known spatial and temporal relationship between water sources as these would have had the Maroon sites themselves. potential to be high traffic areas of the gen- eral populace of the island. A number of GIS layers already exist for St. Croix. The United State Geological Sur-

Journal of Caribbean Archaeology 7, 2007 13 Locating Maroon Refuge Sites Norton and Espenshade vey offers digital elevation models (DEMs), anonymous peers reviewed the article for the Journal the Environmental Protection Agency of lay- of Caribbean Archaeology. ers for hydrology and land cover, and the Vir- References Cited gin Islands Conservation Data Center can provide access to soil survey, hypsography, Agorsah, K. LANDSAT enhanced thematic mapper mosa- 1993 Archaeology and Resistance History in the ics, and NWI wetlands. Caribbean. The African Archaeological Review 11:175-195. 1994 Maroon Heritage: Archaeological, Ethno- Conclusion graphic, and Historical Perspectives. University of West Indies, Barbados. Little is known about the lifeways of Ma- 1999 Ethnoarchaeological Consideration of Social roons in refuge areas such as Maroon Ridge Relationship and Settlement Patterning among Africans in the Caribbean Diaspora. In African in St. Croix. At least in the former Danish Sites Archaeology in the Caribbean, edited by J.B. Virgin Islands, there is not a good archival Haviser, pp. 38-64, Ian Randle Publishers, King- record to guide reconstructions, and archae- ston, Jamaica. ology is the most promising avenue to inter- 2003 Tracking Down the Maroons: Archaeo- preting the Maroon refuge existence. How- geography of Marronage in the Caribbean. Paper presented at the International Congress for Carib- ever, archaeology may fail to properly locate, Archaeology, Santo Domingo, Dominican evaluate, and interpret Maroon refuge sites if Republic. only traditional survey methods are used. It 2005 New Trends in Maroon Archaeology in Suri- is argued here that we should apply a two- name – Transformation Processes. Paper pre- st step approach that begins with GIS modeling sented at the 21 meeting, International Associa- tion of Caribbean Archaeologists, St. Augustine, of desirable (in a Maroon mindset) locations and Tobago. and ends with the intensive metal detector 2006The Other Side of Freedom: The Maroon Trail in and landscape survey of those locations. By Suriname. In African Re-Genesis: Confronting using this different approach, archaeology Social Issues in the Diaspora, edited by J.B. can find key sites that ultimately will provide Haviser and K.C. MacDonald, pp. 191-203. UCL 2007 Predicting and Reconstructing the Formation valuable information on Maroon lifeways in and Nature of Maroon Settlements in Suriname. the former Danish Virgin Islands. Paper presented at 22nd Conference of the Interna- tional Association for Caribbean Archaeology, Acknowledgements. The writing of this paper was Kingston, Jamaica. supported by New South Associates, Inc. The paper’s Allen, S. genesis was a series of conversations with individuals 2001 " Nunca Vai Morrer": History, The concerned with preserving Maroon Mountain, includ- Practice of Archaeology and Race Politics in ing Mr. Onaje Jackson and Dr. George Tyson. Dr. Brazil. Unpublished PhD Dissertation, Depart- J.W. Joseph at New South Associates and Dr. Theresa ment of History, Brown University. Singleton at the Department of Anthropology, Univer- Armstrong, D.V. sity of Syracuse provided in-house peer review. Dr. 2003 Creole Transformation from Slavery to Free- Lynne Guitar, Dr. Gabino La Rosa Corzo, Ms. Sue dom: Historical Archaeology of the East End Sanders, and Bruce Larson assisted us by providing Community, St. John, Virgin Islands. University source material, and Mr. Olasee Davis provided infor- Press of Florida, Gainesville. mation about caves on Maroon Ridge. An abbreviated Arrom, J.J., and M.A. García Arévalo version of this paper was presented at the 2007 meet- 1986 Cimarrón. Fundación García Arévalo, Santo ings of the International Association for Caribbean Domingo. Archaeology in Jamaica. Dr. Chris Clement and two Baptiste Dutertre, Padre Jean 2004 (1667) De Los Esclavos de Martinica, Guada- Journal of Caribbean Archaeology 7, 2007 14 Locating Maroon Refuge Sites Norton and Espenshade

lupe y Antillas Francesas; Del Marcado de Negros Maritime Maroons into Puerto Rico, 1656-1800. en Africa; de lo que hacen; de su Familia; de sus The Journal of Caribbean History 31(1-2):51-87. Comidas y sus Casas; de sus Vestidos y Adornos; Church Missionary Society de sus Fiestas y Danzas; de los Castigos y de los 1828 The Missionary Register for MDCCC XXVIII. Cimarrones; de otras Curiosas Costumbres. In L.B. Seeley & Sons, London. Crónicas de las Poblaciones Negras en el Caribe Conolly, J. and M. Lake Francés, Second Edition, edited and translated 2006 Geographical Information Systems in Archae- into Spanish by Eugenio Fernández Méndez, pp. ology. Cambridge University Press, Cambridge. 101-130. Centro de Estudios Avanzados de Connor, M., and D.D. Scott Puerto Rico y El Caribe, San Juan, Puerto Rico. 1998 Metal Detector Use in Archaeology: An Intro- Barnet, M. duction. Historical Archaeology 32(4):73-82. 1993 The Autobiography of a Runaway Slave. Eng- Dookhan, I. lish Translation by J. Innes. Macmillan Carib- 1994 A History of the Virgin Islands of the United bean, Oxford, England. States. Canoe Press, Kingston, Jamaica. Beck, J.M. Ebert, D. 1754 (Map) Kort Over Eylande St. Croix udi Amer- 2004 Applications of Archaeological GIS. Cana- ica. Image available on-line at http:// dian Journal of Archaeology 28(2):319-342. cas.umkc.edu/StCroix/mapping/beck.htm . Ac- Espenshade, C.T., R.L. Jolley, and J.B. Legg cessed February 27, 2007. 2002 The Value and Treatment of Civil War Mili- Benoit, P.J. tary Sites. North American Archaeologist 23 1839 Voyage a Surinam. Societe des Beaux-Arts, (1):39-68. Brussels. Fernández Méndez, E. Bevan, A. and J. Conolly 2004 Crónicas de las Poblaciones Negras en la 2002-2004 GIS, Archaeological Survey and Land- Caribe Francés. Centro de Estudios Avazados de scape Archaeology in the Island of Kythera, Puerto Rico y El Caribe, San Juan, Puerto Rico. Greece. Journal of Field Archaeology 29 Forbes, E. (1/2):123-138. 1992 African Resistance to Enslavement: The Na- Bilby, K. ture and the Evidentiary Record. Journal of Black 1997 Swearing by the Past, Swearing by the Future: Studies 23(1):39-59. Sacred Oaths, Alliances and Treaties Among the Fox, R.A., Jr. Guianese and . Ethnohistory 44 Archaeology, History, and Custer’s Last Stand. (4):655-689. University of Oklahoma Press, Norman. Binford, L.R. Garcia Arevalo, M.A. 1979 Organization and Formation Processes: Look- 1986 El Maniel de Jose Leta: Evidencias Ar- ing at Curated Technologies. Journal of Anthro- queólogicas de un Possible Asentimiento Cimar- pological Research 35(3):255-273. rón en El Región Sud-Oriental de la Isla de Santo Brandt, R., B. Groenewoudt and K. Kvamme Domingo. Cimmarón Serie Monografica No. 18, 1992 An Experiment in Archaeological Site Loca- Fundación Garcia-Arevalo, Santo Domingo. tion: Modeling in the Netherlands using GIS Hall, N.A.T. Techniques. World Archaeology 24(2, Analytical 1985 Maritime Maroons: "Grand Marronage" from Field Survey):268-282. the Danish West Indies. The William and Mary Camp, S. Quarterly, 3rd Series 42(4):476-498. 2002 'I Could Not Stay There': Enslaved Women, 1992 Slave Society in the Danish West Indies: St. Truancy and the Geography of Everyday Forms of Thomas, St. John, and St. Croix. The Johns Hop- Resistance in the Antebellum Plantation South. kins University Press, Baltimore. Slavery and Abolition 23(3):1-20. Handler, J.S., and F.W. Lange Campbell, M.C. 1978 Plantation Slavery in Barbados: An Archaeo- 1988 The Maroons of Jamaica 1655-1796: A His- logical and Historical Investigation. Harvard tory of Resistance, Collaboration & Betrayal. University Press, Cambridge, . Bergin & Garvey, Granby, Massachusetts. Highfield, A.R., and V. Barac Chinea, J.L. 1987 A Caribbean Mission. English Translation of 1997 A Quest for Freedom: The Immigration of A History of the Mission of the Evangelical Breth- Journal of Caribbean Archaeology 7, 2007 15 Locating Maroon Refuge Sites Norton and Espenshade

ren on the Caribbean Islands of St. Thomas, St. 1996 Exploring the Topography of Mind: GIS, so- Croix, and St. John, written in 1777 by C.G.A. cial space and archaeology. Antiquity 70:612-622. Oldendorp. Karoma Publishers, Ann Arbor, 2006 What You See is What You Get?: Visu- Michigan. alscapes, visual genesis and hierarchy. In Digital Keegan, J. Archaeology: Bridging method and theory, edited 1976 The Face of Battle. Viking Press, New York. by T. Evans and P. Daly, pp. 148-167. Routledge, Kopytoff, B. K. London. 1978 The Early Political Development of Jamaican Lokken, P. Maroon Societies. The William and Mary Quar- 2004 A Maroon Moment: Rebel Slaves in Early terly 3rd Series; 35(2):287-307. Seventeenth-Century . Slavery and 1979 Colonial Treaty as Sacred Charter of the Ja- Abolition 25(3):44-58. maican Maroons. Ethnohistory 26(1):45-64. Maris-Wolf, E. La Rosa Corzo, G. 2002 Between Slavery and Freedom: African 2003 Runaway Slave Settlements in Cuba: Resis- Americans in the Great Dismal Swamp 1763- tance and Repression, translated by Mary Todd. 1863. Unpublished Masters thesis, Department of English translation of 1998 Spanish edition. The Anthropology, College of William and Mary, Wil- University of North Carolina Press, Chapel Hill. liamsburg, . 2005 Subsistence of Cimarrones: An Archaeologi- Nance, J.D., and B.F. Ball cal Study. In Dialogues in Cuban Archaeology, 1986 No Surprises? The Reliability and Validity of edited by L.A. Curet, S.L. Dawdy, and G. La Rosa Test Pit Sampling. American Antiquity 51(3):457- Corzo, pp. 163-180. University of Alabama Press, 483. Tuscaloosa. Orser, C. and P. Funari 2006 Arqueologia del Cimmaronaje: Utiles Para la 1992 Pesquia Arqueológica Initial en Palmares. Resistencia. Paper presented at the 71st Annual Estudos Ibero-Americanos 18:53-69. Meeting, Society for American Archaeology, San 2001 Archaeology and Slave Resistance and Rebel- Juan, Puerto Rico. lion. World Archaeology 33(1):61-72. Lake, M. Pope, P. 2000a MAGICAL Computer Simulation of Meso- 1972 A Maroon Settlement on St. Croix. Negro lithic Foraging on Islay. In Hunter-Gatherer History Bulletin 35(7):153-154. Landscape Archaeology: The Southern Hebrides Price, R. Mesolithic Project 1988-98 edited by S. J. Mithen, 1996 Maroon Societies: Rebel Slave Communities pp. 465-495. Volume 2, Archaeological Fieldwork in the (third edition). The Johns Hop- on Colonsay, Computer Modeling, Experimental kins University Press, Baltimore. Archaeology, and Final Interpretations. The Price, R., and S. McDonald Institute for Archaeological Research, 2002 Maroons Under Assault in Suriname and Cambridge. . Cultural Survival Quarterly 25 2000b MAGICAL Computer Simulation of Meso- (4). lithic Foraging. In Dynamics in Human and Pri- Sanders, S.L., M.A. Simons, C. Goodwin, D. Davis, F. mate Societies: Agent-Based Modelling of Social Vento, and K. Anderson Cordova and Spatial Processes, edited by T. A. Kohler and 2001 Patterns & Transformations in the Prehistory G. J. Gummerman, pp. 107-143. Oxford Univer- and History of Vieques, Volume I: Archeological sity Press, New York. Survey and Inventory of the VNR. R. Christopher Lake, M. and P. E. Woodman Goodwin Associates, Frederick, . 2000 Viewshed Analysis of Site Location on Islay. Sanders, S.L., J. Brown, R.C. Goodwin, D. Davis, F. In Hunter-Gatherer Landscape Archaeology: The Vento, and K. Anderson Cordova Southern Hebrides Mesolithic Project 1988-98, 2003 Patterns and transformations in the Prehis- edited by S. J. Mithen, pp. 497-503. Volume 2, tory and History of Vieques, Volume II; Archeo- Archaeological Fieldwork on Colonsay, Computer logical Survey and Evaluation of Historic Period Modeling, Experimental Archaeology, and Final Sites. R. Christopher Goodwin Associates, Freder- Interpretations. The McDonald Institute for Ar- ick, Maryland. chaeological Research, Cambridge. Sayers. D.O. Llobera, M. 2006 Diaspora Exiles in the Great Dismal Swamp, Journal of Caribbean Archaeology 7, 2007 16 Locating Maroon Refuge Sites Norton and Espenshade

1630-1860. Transforming Anthropology 14 White, C. (1):10-20. 2005 Report on the 2004 Archaeological Investiga- Scott, D.D., R.A. Fox, Jr., M. Connor, and D. Harmon tion of a Maroon Settlement in Suriname, South 1989 Archaeological Perspectives on the Battle of America. Paper presented at the 21st meeting, In- the Little Bighorn. University of Oklahoma Press, ternational Association of Caribbean Archaeolo- Norman. gists, St. Augustine, . Schuler, M. Yaremko, J.M. 1970 Ethnic Slave Rebellions in the Caribbean and “Gente Bárbara”: Indigenous Rebellion, Resistance . Journal of Social History 3(4):374- and Persistence in Colonial Cuba, c. 1500-1800. 385. KACIKE: The Journal of Caribbean Amerindian Schwartz, S.B. History and Anthropology [On-line Journal]. 1996 Slaves, Peasants, and Rebels: Reconsidering Available at: http://www.kacike.org/ Brazilian Slavery. University of Illinois Press, Yaremko.html . Accessed February 15, 2007. Chicago. Sterling, B.B., and B.W. Slaughter 2001 Surveying the Civil War: Methodological Approaches to the Antietam Battlefield. In Ar- chaeological Perspectives on the American Civil War, edited by C.R. Geier and S.R. Potter, pp. 305-322. University of Florida Press, Gainesville. United States Department of Commerce 1942 (Map) Chart 3242, West Indies, Virgin Is- lands, St. Croix. Image available on-line at the Historic Map & Chart Project web page of the National Oceanic and Atmospheric Administra- tion at http://historicals.ncd.noaa.gov/historicals/ histmap.asp . Accessed February 27, 2007. Vega, B. 1979 Arqueólogia de los Cimarrones del Maniel del Bahoruco. Boletin, Museo del Hombre Domini- cano 12:11-48. Voight, C. 2006 Northwest St. Croix: Our Spiritual Legacy. St. Croix This Week, edition of July/August 2006. Published in Christiansted, St. Croix, U.S.V.I. Weik, T. 1997 The Archaeology of Maroon Societies in the Americas: Resistance, Cultural Continuity, and Transformation in the . Histori- cal Archaeology 31(2):81-92. 2004 Archaeology of the African Diaspora in Latin America. Historical Archaeology 38(1):32-49. Wheatley, D. 1995 Cumulative Viewshed Analysis: a GIS-based method for investigating intervisibility, and in archaeological application. In Archaeology and Geographic Information Systems: A European Perspective, edited by G. Lock and Z. Stancic, pp. 171-185. Taylor & Francis, London. Wheatley, D. and M. Gillings 2002 Spatial Technology and Archaeology. Taylor and Francis, London. Journal of Caribbean Archaeology 7, 2007 17