In 1987, excavations at Seville Plantation, an 18th-century sugar plantation situated on the north coast of Jamaica at St. Ann’s Bay, were initiated by the Seville African Jamaican Archaeological Project. Led by Douglas Armstrong of Syracuse University, these excavations concentrated on two slave villages, one dating from the early-to-mid 18th century and the other dating from the late 18th-through mid-19th century. Houses 15 and 16 are located in the first-period African settlement of the plantation. These structures are located to the southwest of the planter's residence.

Excavations occurred at House 15 between 1987 and 1993. Archaeological features such as structural postholes (F01-F11) and probable stone foundation walls confirm that this dwelling was constructed of wattle and daub. Unlike other house areas in the early village, House 15 did not have an associated human burial. Both European and locally-made artifacts were recovered, and they represent a wide range of activities from food production and consumption to spiritual practices.

The early African village at Seville, where House 15 and 16 are located, is represented on a 1721 map of St. Ann’s Bay. This map included a number of accurate details, such as navigational vectors used to guide mariners entering St. Ann’s Bay. In addition, the early map locates the village behind and upslope from the main house. Two linear rows of tightly spaced houses are arrayed along a road or path. The road heads south from the planter’s residence. This is the only identified map showing the location of the early village and the approximate locations of Houses 15 and 16 (Armstrong 1990, 1991).

A 1792 estate plan shows the location of a second slave village that dates from the last quarter of the 18th-century. The later village on this map indicates the houses were clustered together, with each house oriented independently, and with significantly more yard space between each house (Armstrong and Kelly 1992).

These maps are the main textual sources for Seville. In 1982, Armstrong conducted several oral histories with Mr. Carpy Rose, who was born in one of the last houses to remain standing in the later village. These interviews are discussed in a number of presented papers and publications (Armstrong 1990, 1991).

Archaeological and historical research associated with the early slave village at Seville was initiated in 1987. It built on an initial survey of the area conducted by Armstrong in 1981 (Armstrong 1991). This early survey indicated that there were two distinctive village areas at Seville. Prior to that time only the more recent slave village was recognized. The later village is clearly shown on a map dating to 1792. It was assumed that an earlier 1721 map of St. Ann’s Bay (which at the time was misattributed by the National Library as being made in 1691), was thought to simply show the same village. In examining this map, Armstrong was able to determine that it was made in 1721. Based on this map, he decided to survey not only the area to the west of the planter’s residence, where ruins of a village were known, but also the area uphill and behind the planter’s house, as indicated on the map. Slipware, delftware, and local earthenware were present on the surface and Armstrong hypothesized that this may have been the location of an early village site on the property.

When the Seville African Jamaican archaeological project was initiated in 1987, the idea was to first define the boundaries of both slave settlements and then compare data relating to early and late residences on the property. An intensive walking survey in 1987 confirmed the findings from the 1981 survey: there were two distinct loci of slave occupation on the property. The area uphill and behind (south and southwest) of the planter’s residence was included in the preliminary walking survey, even though it was owned by the St. Ann’s water company rather than the National Trust. At the time, the Jamaica National Heritage Trust owned all of the other areas associated with industrial production and residential life on Seville sugar estate. When the study began in 1987, local farmers were in the process of expanding their banana fields into the area of the early village site. The project convinced them to expand in another direction and ultimately got support from the land owner for the protection of the property. This information was used by the JNHT to later acquire this area and include it in the heritage park.

The house sites were identified via a walking survey traversing the study area. House 15 was located in the same line of houses as House 16, with another house, House 20, located between the two structures (DAACS did not analyze House 20). Located along the western row of houses, it is believed to have been constructed near the middle of the village.

The objective was to define both the house structure and any associated features. Excavation was initiated with a single test unit then expanded for aerial coverage. A grid of 1-x-1 meter quadrats was laid across the site. Grid north was 35 degrees east of magnetic north. Excavation units were labeled using an alphanumeric system with each one-meter unit designated using a letter and number. In most cases, the letters increase to the north and numbers increase to the west (i.e. A1, A2, B1, B2, and so forth). The exceptions at House 15 are transects Y and Z. These transects were the southern most transects on the site, so the north/south transect labels proceed as follows: Y, Z, A, B, C, etc. The southeasterly most quadrat at House 15 was Y11.

In order to generate spatial distribution data that was compatible with DAACS, DAACS analysts assigned a new coordinate system, expressed as northings and eastings, to Armstrong's existing alphanumeric grid. DAACS established a datum (0/0 as opposed to A1) that was located to the south and west of the excavated area. Quadrat boundary data in DAACS for House 15 are therefore represented in the number of meters away from the DAACS datum, i.e. E9N1.

A total of 81 1-x-1 meter units were excavated at House 15. All units were excavated by hand and all sediment recovered was passed through screens with 1/8 inch mesh. Every unit was excavated following a combination of natural and arbitrary stratigraphic designations. Each level was dug to a maximum of 10 centimeters unless a cultural feature was encountered or until there was a natural change in soil color and/or texture. In most cases the upper 10 centimeters included a mix of materials that dated from the 17th century through the 20th centuries. The second level, 10-20 centimeters in depth, contained primarily early material culture. The third level was of variable depth and it was in this level that the foundations and features of the house were fully exposed. Given the slope of the hill, generally the northern units, or down-slope portions of the house site, had structural features that were exposed on or near the surface. However, features in the southern units, or upslope portions of the house, were covered with 20-30 centimeters of sediment.

Levels are represented numerically. For example, C11.1 stands for level 1 in unit C11. C11.2 stands for level 2 in unit C11, and so forth. In the original field records, 0 (zero) was used to represent surface levels. DAACS replaced the 0 with a "S", therefore C11.S in DAACS represents the surface level in unit C11.

One area of the early village was later reused in the mid-19th century when a house thought to be associated with East Indian laborers was constructed. This later occupation, identified as a distinct later assemblage, overlays part of the early village. This East Indian household, defined as House Area 14, provides an important comparative base for studies of ethnicity and social cultural attributes associated with laborers on the property (see Armstrong and Hauser 2004). Fortunately, even though the site has relatively shallow stratigraphy, only two of the early African context house sites (Houses 12 and 13) were directly disturbed and even these two retained stratified contexts and sequential layers of house construction materials. Although only Houses 12 and 13 were directly impacted by this later 19th century occupation, the presence of these later contexts, along with other more casual discarding of materials in the area, accounts for more recent materials found in the upper layers of all house sites in the early village.

Site plans and detailed photographic documentation was done for all excavated house sites at the early village. Dates were compiled in the field using field recording forms and transferred to a dBase 4 database in the field. The data were downloaded into AUTOCAD and SURFER in order to carry out spatial analyses. Much of the data was then compiled and analyzed using Lotus spreadsheet and graphs. Artifacts recovered from House 15 are curated by the Jamaica National Heritage Trust.

House 15 was initially identified by a row of rocks that defined the northern foundation wall. The structure appears to have originally have been two rooms of the same size, similar to House 16, and was most likely occupied around the same time. At some point during its use, a third room was added. The house was of wattle-and daub construction using post construction on a stone foundation. All features identified at House 15 are postholes (F1-F11), the majority belonging to the structure. None of the identified House 15 features were excavated. As a result, there are no context records or artifacts associated with the postholes. They are only represented on the final site plan.

Like all of the other houses in the early village, the floor of House 15 was leveled using limestone to build up the lower (northern) end of the house. Hence, the more formally constructed north side of the house was most easily recognized and the less formally constructed post holes on the south side were less easily defined.

This house site had a hearth in the yard and a distribution of artifacts that suggested informal yard boundaries between houses. Evidence suggests that this house, like all houses along the path, were aligned in a similar fashion and that they utilized relatively narrow spaces between the houses as a means of entry into their yards. Unlike House 16, House 20, and several other house-yards, House 15 did not have an associated burial. Remote sensing was done on the site in 1993 confirming the absence of the type of anomaly represented by burials dug into the bedrock of the site

Analyses of materials from this site are included in the DAACS database. Analysis of them has shown distinct patterns of material use including reuse of items such as gun flint and raw flint for strike-a-light fire starters, reworking of local and imported ceramics for gaming pieces. The distributions of artifacts around the edge of the yard support the importance of the yard as a place in which activities of the household and community took place. Artifacts like ground cowry shells, local earthenware, and locally made tobacco pipes provide clues as to the importance of both continuities of African traditions and the generation of goods and trades by persons of African descent in Jamaica. Patterns of material use, including imported and local wares reflect a household with limited financial means particularly when compared with materials from managerial contexts at Seville. There is greater reuse of items such as glass bottles. This house and its neighboring house sites provide a solid picture of early life within a laborer household under conditions of slavery.

The study of Seville has been published in a series of articles that have addressed thematic issues relating to the plantation. These include an overview of the cultural landscape of Seville Plantation which focuses on the African Jamaican settlements and their interpretation through time (Armstrong 1999), a comparative analysis of the processes of internally defined transformations within society (Armstrong 1998), landscapes and settlement patterns in relationship to social relations (Armstrong and Kelly 2000), house-yard burials (Armstrong 2000, Armstrong and Fleischman 2003), and comparative studies related to the material record of race, ethnicity and labor conditions (Armstrong 1998 and Armstrong and Hauser 2004). Comparative analysis of African and East Indian laborer contexts are presented in Armstrong and Hauser 2001 and 2004. In addition, details related to a refined analysis of temporal contexts using both mean and variable measures of variance from the mean using whisker plots is described in a methodological paper (Armstrong 2005). Now the DAACS database is presenting detailed analysis of selected sites on the property so that data can be compared and shared. Several recent conference papers that use Seville data are available through the DAACS website's Current Research page (Galle 2007a, 2007b; Nelson, Neiman, and Galle 2007). With this renewed interest in the project related to the DAACS project, a synthetic analysis of the project is being prepared by Armstrong and Mark Hauser.