|Location:||Stewart Castle, Falmouth, Trelawney, Jamaica|
|Occupation Dates:||Last-half 18th through second-quarter 19th century. Phasing and mean ceramic dates can be found on the Chronology Page.|
|Excavator(s):||Jillian Galle, Leslie Cooper, Ivor Conolley and Karen Hutchins, with assistance from Derek Wheeler and archaeological field school students.|
|Dates excavated:||May-June 2007.|
Excavations at the Stewart Castle Main House and Slave Village were initiated by DAACS in May 2007 as part of the DAACS Caribbean Initiative. Located just east of Falmouth Jamaica, in Trelawney Parish, Stewart Castle was patented in 1754. It functioned as a large-scale sugar plantation through the early nineteenth century. Due to a massive decline in sugar prices, the property gradually converted to a cattle pen by the mid-19th century. A 1799 plat by the surveyors Munro, Stevenson, and Innes captures the scope of the sugar plantation in detail, showing the location of the slave village, the fortified main house, sugar works, and slave provision grounds.
Prior to 2007, no archaeological work relating to the historic period occupations at Stewart Castle had been conducted. The goal of the 2007 DAACS excavation was to assess the temporal components at the main house, and to understand both the temporal and spatial occupations at the slave village. With the help of students from the University of the West Indies and the University of Virginia, 176 shovel-test-pits (STPs) were excavated across a 7000 square meter area east of the road that bisects the village. Three 1-x-1 meter units were also excavated, one placed near an early nogged house, one on the interior of a cut limestone foundation, and the third on an earthen terrace. 10,298 artifacts were recovered from these pits and quadrats. Architectural and landscape features, such as stone walls, foundations and limestone nog piles, and landscape terraces, were selectively mapped with a total station at the village.
Stewart Castle, patented in 1754 by James Stewart I, began as a small, 167 acre landholding. By 1799, the sugar plantation had grown to well over 1200 acres, with approximately 500 acres planted in sugar cane. During the second decade of the 19th century, an average of 332 enslaved Africans lived and worked on the property. James Stewart II had difficulty managing the property and was forced to mortgage the plantation in 1797. Although Stewart did not officially sell the property until 1830, it may have been managed by an agent of the London merchants who held the mortgage, hence prompting the creation of the 1799 Munro, Stevenson and Innes plat (Panning 1996a,1996b). Although it has been suggested that the estate became a cattle pen at the time of its sale in 1830, Stewart Castle was clearly converting to cattle well before then, as it had an average of 282 cattle each year during the second decade of the 19th century (Panning 1996a).
The Munro, Stevenson, and Innes plat is currently the best primary source regarding slavery at Stewart Castle. The plat shows at least 43 structures located on approximately 45 acres, which are labeled on the plat “Negro Houses”. Located to the northeast of the village is an area of approximately 170 acres that was designated as “Negro Grounds” and “Rocky Woodlands and Negro Grounds”. A portion of the current road overlaps with the path represented on the plat by the dashed line. The path passes through the village and continues through a marginal limestone forest that was once the provision grounds. Today, the road ends but the path continues over the crest of the ridge, and descends along a steep, rocky face before ending at the wharf. This path is still used by fishermen who make a daily fishing camp at the late 18th-century storehouse located along the wharf. One fisherman, who lives in the Refuge community located near Stewart Castle, has been walking this path daily for at least 60 years. His father walked the same path throughout the early twentieth century and it is likely this path has been in use since slavery.
Stewart II, like most resident Jamaican planters, was clearly concerned about living in close proximity to over 300 enslaved individuals. The almost obsessive focus on fortifications demonstrates a perceived and possibly very real threat to Stewart Castle from the Maroons and those he enslaved. If the Castle came under attack, Stewart had the ability to secure people and animals inside the courtyard, as well as ensure a large supply of safe drinking water stored in a large water cistern (3.1 meters x 5.5 meters by 2.5 meters deep) inside the main house (Panning 1996b:202).
Excavation history, procedure and methods
In March 2006, DAACS staff undertook a two-day surface survey of the Stewart Castle property. By laying the 1799 Munro, Stevenson, and Innes plat onto a modern topographic map of the area, we were able to pinpoint the exact location of the now densely overgrown Stewart Castle slave village. A thorough surface inspection of the village revealed cut-limestone foundations, piles of limestone cobbles used for the construction of nogged walled houses, as well as a thick surface scatter of late-18th century domestic artifacts spread over approximately 33,457 square meters.
The goal of the 2007 DAACS excavation was to assess the temporal components at the main house, and to understand both the temporal and spatial occupations at the slave village. DAACS began by dividing the Stewart Castle property into three survey areas. Area 1 encompasses the Stewart Castle Main House and Yard, Area 2 is comprised of the area in the slave village west of the road, which bisects the village along a north/south trajectory, and Area 3 is located to the east of the road in the village. Archaeological excavations only occurred in Areas 1 and 3 in 2007. Both areas were covered in heavy bush and required extensive clearing prior to excavation
DAACS staff returned to Stewart Castle in May 2007 with a team of students from the University of Virginia and the University of West Indies, Mona. Excavations began at the main house, where a datum and UTM-grid was established. A total station was used to connect the main house datum to a village datum. As occurred at the main house, both the village datum and total station back sight were excavated as 50 centimeter STPs. These pits were then filled with concrete and rebar was placed into the concrete. A GPS unit was used to confirm the location of the village datum, ensuring that excavations at both the village and main house were accurately placed on the same UTM grid.
A dirt road that overlaps with the path represented on the 1799 plat cuts the village nearly in half along the north/south axis. DAACS used this road to divide the village into two survey areas. The portion of the village to the west of the road is Area 2, and no STPs or excavation units were placed in this area in 2007. Area 3, the section of the village to the east of the road, was the focus of the 2007 excavation season.
One hundred and seventy-four shovel-test-pits were placed on 6-meter centers in a 144-x-56 meter area using a total station. An alphanumeric system was established for naming STPs that combined the Area, the Transect Letter, and the STP number. Transects were labeled alphabetically. Transect K represents that northernmost transect and Transect R is the southernmost transect excavated during 2007. STPs were numbered consecutively within each transect, beginning at the western edge of each transect and moving east. STP labels therefore look like 3-L-01, which translates into Pit 1, on Transect L, in Area 3. The same system was used for the main house.
A total of 176 pits were excavated at Area 3, in addition to the datum and back sight pits. All STPs were 50 centimeters in diameter and all excavated sediment was screened through 1/4 inch mesh. In most cases, the pits were excavated to either subsoil or bedrock. Please note that those using the STP data for analysis will encounter data from 3-DATUM and 3-BKSIGHT, which are the STPs excavated for the datum and back sight and which are not on the 6-meter STP survey grid.
Three 1-x-1 meter units were also excavated in Area 3. The placement of Units 1 and 2 was motivated by the desire to date the occupations of two different housing types, a nogged-walled house with no apparent foundation and a house with a rather substantial, cut limestone foundation. The house with the cut-limestone foundation was located on the eastern edge of a large earthen terrace. Unit 3 was located at the top of the western edge of this terrace, and its purpose was to date and clarify the stratigraphy of this terrace.
One goal of this shovel-test-pit survey is to identify temporal trends in occupation across the village. Unit 1 was placed in the middle of one of the many piles of limestone cobbles that are scattered throughout the village, and which likely represent the remains of nogged-walled houses. This cobble scatter was located just south, and southwest of two STPs (3-M-06, 3-M-07) that produced sherds of White Salt Glazed stoneware, the earliest dated ceramic excavated during the 2007 season. Only three other STPs scattered throughout the village contained WSG.
Unit 2 was excavated inside the only clearly identifiable cut-limestone foundation in Area 3. It is likely that the construction of the earthen terrace to the west of this foundation relates to the construction of this house. Sediment layers within Unit 2 were filled with limestone rubble. Although an intact plaster floor was not discovered, concentrations of mortar and plaster were found in 002A (60 mortar fragments) and 002B (114 mortar fragments). This debris is likely evidence of nogged walls or possible plaster floors.
Unit 3 was placed on the western edge of the earthen terrace, whose construction may have been linked to the construction of the house with the cut limestone foundation. Unit 3 contained multiple layers of fill, all of which have mean-ceramic dates that are between 1795 and 1805.
Artifacts were washed and cataloged to DAACS standards on Jamaica and in the DAACS lab at Monticello. Faunal remains from the main house and village excavations are currently being analyzed at Colonial Williamsburg’s Zooarchaeological Laboratory and will be available through this website in early 2008.
Architectural and landscape features, including stone walls, terraces, limestone cobble piles, and a grave, were selectively mapped with a total station. The dense overgrowth in the village made it difficult to map landscape features in a systematic fashion; those represented on the site map are features that could be easily mapped with the total station and limited clearing. Future seasons will focus on clearing large swaths of the foliage to ensure more complete coverage.
Excavation plans for future seasons include extending the STP transects into Area 2, and to the north and south in Area 2. We will continue to use 1 x 1 units to clarify stratigraphy, explore architectural features, and date buried deposits. All data collected in future excavation seasons will be cataloged to DAACS standards and provided through the DAACS website.
Summary of research and analysis
The most complete historical research on Stewart Castle to date was conducted by Steven Panning in the mid-1990s. In addition to conducting research in Jamaican archives, Panning produced an accurate measured drawing of the Stewart Castle Main House complex (Panning 1996a, 1996b). He visited the main components of the property, including the main house, wharf, village, overseer’s house, and works. Panning recommended that archaeological excavations be conducted at the village (Panning 1996b: 200).
In her honors thesis titled “Surveillance and Production on Stewart Castle Estate: A GIS-based analysis of models of plantation spatial organization” (2007), Lynsey Bates used viewshed and cost-surface analysis to analyze plantation organization at Stewart Castle. Through the combination of modern topographic maps and the 1799 plat, Bates concluded that surveillance of the enslaved population at Stewart Castle was secondary to minimizing travel time between the village, the sugar works, and certain fields. By placing the works and village in locations that minimized travel between the two, Bates demonstrated that Stewart sacrificed visibility, since neither the provision grounds nor village were observable from the main house or the overseer’s house (Bates 2007). Bates’ thesis is available for download at: http://www.daacs.org/research/BatesThesis.pdf.
A number of research papers and posters related to DAACS’s 2007 research at Stewart Castle will be presented in 2008. These will be made available through the DAACS website after they are presented.
The Digital Archaeological Archive of Comparative Slavery
Things you need to know about the Stewart Castle Village before you use the data:
- Field measurements are in meters and centimeters.
- Excavation units and shovel-test-pits at both the Main House and Village are on the UTM grid system.
- One hundred seventy-six shovel-test-pits and three 1-x-1 meter units were excavated at the Stewart Castle village in 2007. More extensive area excavations are planned for subsequent seasons.
- Architectural and landscape features including stone walls, terraces, limestone cobble piles, and a grave, were selectively mapped with a total station. The dense overgrowth in the village made it difficult to map landscape features in a systematic fashion; those represented on the site map are features that could be easily mapped with the total station and limited clearing. Future seasons will focus on clearing large swaths of foliage to ensure more complete coverage.
DAACS is extremely grateful for the generous support of The Reed Foundation, which provided scholarships and internships for students from the University of the West Indies, Mona.
The DAACS field work at Stewart Castle would not have occurred without the support and colloboration of Louis Nelson, Associate Professor of Architectural History of UVA, and director of The Falmouth Field School in Architectural History. He introduced DAACS to the Stewart Castle site in 2005 and has been a strong supporter of the DAACS program.
Many thanks to the wonderful students, interns, and volunteers who put such incredible energy into the field and laboratory work for Anthropology 382: The Falmouth Field School in Historical Archaeology (University of Virginia). They include: Vanessa Bonner (UWI), Lauren Burn (UWI), John Chenoweth (UCB), Vanessa Clark (UWI), Shemika Crawford (UWI), Anna Doctor (UWI), Clive Grey (UWI), Shailean Hardy (UWI), Alexandra Jones (UCB), Sarah Kidder (COC), Brian McCray (UVA), Andrew Mullan (UVA), Susan Sherwood (UMW), and Julene Wright (UWI). Students from The Falmouth Field School in Historic Preservation also particpated throughout the field season.
The project is indebted to Rim Patterson-Gooden, John Reynolds, Peter Maxwell, and Kemar Walters for their incredible machete skills and valued guidance. The site would have been inpenetrable without them. James Parrent and his fabulous staff at Falmouth Heritage Renewal, in Falmouth Jamaica, ensured that the field school ran smoothly each day. To find out more about the good work that Falmouth Heritage Renewal is doing to preserve the historic landscape of Falmouth, while providing training in the building trades, please go to: http://www.falmouthjamaica.org/.
Edward Chappell (Director of Architectural Research, Colonial Williamsburg) and Matthew Webster (Director Preservation, Drayton Hall) lent their amazing skills to the recordation of the Overseer’s House at Stewart Castle and the outbuildings, including the identification of the fortified three-seater privy, at the Main House complex.
We are grateful for the support of our colleagues at the Jamaica National Heritage Trust, epscially Dorrick Gray and Anne-Marie Howard Brown, and the University of the West Indies, Mona, including Dr. Sabrina Rampersad, Dr. Swithin Wilmot, and Dr. James Robertson.
A number of DAACS and Monticello staff provided their invaluable advice, skills, good humor, and great effort to ensure the success of the 2007 field season. They include Ivor Conolley, Leslie Cooper, Karen Hutchins, Fraser Neiman, and Derek Wheeler. Jesse Sawyer and Brian McCray contributed to the artifact analysis upon our return to Monticello.
The Digital Archaeological Archive of Comparative Slavery
There were no archaeological features identified or excavated at the Stewart Castle Village.
DAACS Seriation Method
DAACS has developed an uniform set of methods to infer intra-site chronologies for all of the sites included in the Archive. These methods, which include frequency-seriation and correspondence analysis, were developed by DAACS (see Neiman, Galle, and Wheeler 2003 for technical details). The use of common methods for all sites in the Archive is designed to increase comparability among temporal phases at different sites. The methods and the phase assignments they produced are summarized below. For some sites, the original excavators developed intra-site chronologies and where these exist, they are included on the Background page for the site. Archive users may also use the Mean Ceramic Date queries provided on the DAACS query page (http://www.daacs.org/queryDatabase/MCDQueries.html) to calculate MCDs for individual contexts or features.
Stewart Castle Village Chronology
This section summarizes the frequency-seriation based chronology for at Stewart Castle. DAACS seriated ceramic assemblages from the slave village that contained more than 5 sherds from individual excavated contexts. Please note that at the Village, ware types, not mean-ceramic-date types, were used in the frequency seriation, correspondence analysis, and in developing the dates for each occupational phase. Please go to http://www.daacs.org/aboutDatabase/MCDTypes.html for more information on the differences between ware types and mean-ceramic-date types.
DAACS computed the frequency of ceramic ware types in each individual context. The seriation chronology is derived from a correspondence analysis of the ware-type frequencies. Seriated contexts were assigned to four phases. Phases are groups of assemblages that have similar correspondence-analysis scores and are therefore inferred to be broadly contemporary. Phases assigned by DAACS have a P-prefix that precedes the phase number (e.g. P01 equals Phase 1).
The stratigraphic relationships among contexts are summarized in the Harris Matrix for the Stewart Castle Village. Phase assignments from the seriation are shown on the Harris Matrix in color, facilitating comparison of the seriation chronology and the stratigraphic chronology of the site.
Stewart Castle Village Phases
Based on the correspondence analysis, DAACS divided the Village occupation into three phases. Phase 3 is comprised of three STP-contexts, which are distinguished by the inclusion of whiteware. These STPs may be capturing evidence of a post-emancipation occupation at the village. Ware-type mean ceramic dates with start dates of 1700 were used to compute MCDs for each phase. The MCDs for the three phases are given in the table below.
The table also includes three estimates of the ceramic TPQ for each phase. The first TPQ estimate is the usual one – the maximum beginning manufacturing date among all the ware-types in the assemblage. The second estimate — TPQp90 — is the 90th percentile of the beginning manufacturing dates among all the sherds in the assemblage, based on their ware-types. The TPQp95 provides a robust estimate of the site’s TPQ based on the 95th percentile of the beginning manufacturing dates for all the artifacts comprising it. These last two TPQ estimates are more robust against excavation errors and taphonomic processes that might have introduced a few anomalously late sherds into an assemblage.
|3-L-06, 3-L-20, 003CLEANUP, 3-K-08, 3-Q-21, 3-K-07, 3-O-02, 3-R-05, 3-M-05, 3-R-18, 3-K-12, 3-K-20, 3-O-12, 002C, 3-R-08, 003C, 003D, 3-M-06, 3-Q-06, 3-O-07, 3-P-13, 3-P-05, 3-M-08, 3-M-03, 3-P-18, 3-Q-02, 3-O-11, 3-O-18, 3-L-09, 3-O-06, 3-O-05|
|3-M-09, 3-K-13, 3-N-21, 3-M-07, 3-P-04, 3-L-12, 3-M-11, 3-N-14, 3-P-09, 3-R-13, 3-R-20, 3-O-13, 002B, 3-P-06, 003A, 003B, 3-R-06, 3-N-08, 001B, 001A, 3-R-07, 3-N-10|
|3-N-06, 3-Q-07, 3-Q-08|
|3-O-09, 3-Q-05, 3-N-07, 3-Q-04, 3-O-21, 3-P-02, 3-P-03, 002CLEANUP, 3-P-07, 3-P-08, 3-P-10, 3-P-17, 3-P-20, 3-P-21, 3-P-22, 3-P-23, 3-Q-01, 3-Q-03, 002D, 002A, 003F, 3-O-03, 3-O-04, 3-P-15, 3-L-18, 3-L-13, 3-N-19, 3-N-11, 3-M-16, 3-M-19, 3-M-20, 3-N-02, 3-N-18, 3-O-19, 3-O-20, 3-P-11, 001C, 003E, 3-Q-09, 3-Q-10, 3-R-15, 3-R-16, 3-R-17, 3-R-19, 3-R-21, 3-R-22, 3-R-23, 3-R-24, 3-P-19, 3-M-02, 3-M-14, 3-M-18, 3-N-15, 3-O-01, 3-O-15, 3-P-12, 3-Q-23, 3-R-14, 3-R-12, 3-R-10, 3-Q-11, 3-Q-13, 3-Q-12, 3-Q-14, 3-Q-15, 3-Q-16, 3-Q-17, 3-Q-18, 001CLEANUP, 3-Q-19, 3-Q-20, 3-Q-22, 3-R-01, 3-R-03, 3-R-02, 3-R-04, 3-R-09, 3-L-15, 3-M-04, 3-N-09, 3-L-10, 3-O-17, 3-L-08, 3-L-07, 3-M-13, 3-L-03, 3-K-15, 3-N-03, 3-N-16, 3-K-19, 3-N-17, 3-L-14, 3-P-14, 3-L-16, 3-L-11, 3-O-22, 3-M-10, 3-N-05, 3-M-17, 3-M-15, 3-N-20, 3-O-16, 3-L-17, 3-L-05, 3-N-12, 3-P-16, 3-L-02, 3-K-18, 3-M-12, 3-K-11, 3-K-06, 3-K-03, 3-L-04, 3-O-08, 3-N-13, 3-K-01, 3-BKSIGHT, 3-L-19, 3-DATUM, 3-K-21, 3-K-14, 3-K-09, 3-O-14, 3-P-01, 3-K-05, 3-O-10, 3-N-01, 3-M-01, 3-K-04, 3-K-02, 3-K-17, 3-R-11, 3-K-10, 3-K-16, 3-N-04, 3-L-01|
Stewart Castle Village Harris Matrix
The Harris Matrix summarizes stratigraphic relationships among excavated contexts and groups of contexts that DAACS staff has identified as part of the same stratigraphic group. Stratigraphic groups and contexts are represented as boxes, while lines connecting them represent temporal relationships implied by the site’s stratification, as recorded by the site’s excavators (Harris 1979).
Stratigraphic groups, which represent multiple contexts, are identified on the diagram by their numeric designations (e.g. SG10) followed by the original excavator’s descriptions of them (e.g. “occupation zone”). There are no stratigraphic groups within the Stewart Castle Village, therefore the Harris Matrix contains only the relationships between individual contexts, which are identified by their individual context numbers (e.g. 96).
Boxes with color fill represent contexts and stratigraphic groups with ceramic assemblages large enough to be included in the DAACS seriation of the site (see Chronology). Their seriation-based phase assignments are denoted by different colors to facilitate evaluation of the agreement between the stratigraphic and seriation chronologies. Grey boxes represent contexts that were not included in the seriation because of small ceramic samples.
See the Stewart Castle Village Chronology for stratigraphic and phase information.
This Harris Matrix is based on data on stratigraphic relationships recorded among contexts in the DAACS database. It was drawn with the ArchEd application. See http://www.ads.tuwien.ac.at/arched/index.html.
For a printable version, download the Harris Matrix [528.67 KB PDF].
PDF of main house complex and slave village excavation areas, compiled by DAACS using Panning’s measured drawings and DAACS Total Station.
PDF of excavator’s plan showing closeup of test units with surrounding shovel test pits and landscape features.
PDF of excavator’s plan of slave village showing shovel test pits, test units, and landscape features.
2007 Surveillance and Production on Stewart Castle Estate: A GIS-based analysis of models of plantation spatial organization. B.A. Honors Thesis, Department of Anthropology, University Press of Virginia, Charlottesville, Virginia.
Harris, Edward C.
1979 Principles of Archaeological Stratigraphy. Academic Press, London, England.
Neiman, Fraser D., Karen Y. Smith , Sara Bon-Harper , and Derek Wheeler
2008 Measuring Settlement Pattern Change on the Monticello Plantation Home Farm. Paper presented at the Annual Meeting of the Society for Historical Archaeology. Albequerque, New Mexico, January 2008.
1996a Exploring Stewart Castle Estate. Jamaican Historical Society Bulletin, 10 (14), pp 172-180.
1996b Exploring Stewart Castle Estate. Jamaican Historical Society Bulletin, pp 200-205.
Ramenofsky, Ann , Fraser D. Neiman , and Christopher Pierce
2009 Measuring Time, Population, And Residential Mobility From The Surface at San Marcos Pueblo, North Central New Mexico. American Antiquity, 74(3): 505-530.
1999 Spatial and Multivariate Analysis, Random Sampling Error, and Analytical Noise: Empirical Bayesian Methods at Teotihuacan, Mexico. American Antiquity, 64(1):137-152.