Final Project Report

The final report on the project was submitted to the National Endowment for the Humanities in summer 2006.

A Comparative Archaeological Study of
Colonial Chesapeake Culture
FINAL REPORT

by Julia A. King, Catherine L. Alston, Gregory J. Brown, Edward E. Chaney, John C. Coombs, C. Jane Cox, David Gadsby, Philip Levy, Al Luckenbach, David F. Muraca, Dennis J. Pogue, Benjamin J. Porter, and Shawn Sharpe

Prepared for the National Endowment for the Humanities (RZ-20896-02), June 2006

Acknowledgments

This project began as an exciting idea: what would it be like and what might be accomplished if archaeologists from a number of institutions in Maryland and Virginia came together to collaboratively explore material life in the Chesapeake during the colonial period? A growing number of archaeological sites had been excavated, their collections processed, and the information they contained made publicly accessible  — the pieces were literally in place. What was lacking, however, was an infrastructure — or perhaps, more appropriately, a superstructure  — to provide the resources for cross-institutional collaboration. Fortunately, a major source of support to transform the initial idea into a reality came in the form of a generous grant from the National Endowment for the Humanities (NEH). We are especially grateful to NEH, our program officer, Elizabeth Arndt, and the staff of the Division of Collaborative Research for their interest in and support of our work. We are similarly grateful to the Virginia Department of Historic Resources for its financial support for the project as well as access to collections in its custody and to DHR director Kathleen Kilpatrick for her enthusiasm for the project.

Our respective institutions — Anne Arundel County’s Lost Towns Project, the Colonial Williamsburg Foundation, George Washington’s Fredericksburg Foundation, Historic Mount Vernon, and the Maryland Archaeological Conservation Laboratory — generously allowed us to participate in the project. Our institutions provided meeting space, collections access, administrative assistance, and other forms of support. We anticipate that the discoveries we have made as part of this project will find their way into the research and public programs of each organization.

We are also indebted to our project consultants, including Marley R. Brown III, Cary Carson, John C. Coombs, Andrew Edwards, Keith Egloff, William Kelso, Philip Levy, and Fraser D. Neiman. In addition, many other colleagues contributed their time and knowledge to the effort. David Givens of the Association for the Preservation of Virginia Antiquities’ Jamestown Rediscovery Project deserves special acknowledgment for his assistance with assembling artifact catalogs from the Buck and Sandys sites. Ed Chappell, Willie Graham, and Carl Lounsbury, all from the Colonial Williamsburg Foundation, took time to review architectural evidence from a number of the sites, sometimes resulting in new and exciting interpretations. David Gadsby, formerly of the Lost Towns Project, and Bill Pittman of the Colonial Williamsburg Foundation supplied us with electronic catalogs and maps from collections under their control.

Catherine Alston, who served as Project Archaeologist, did a phenomenal job coordinating the project, assembling and standardizing databases, and keeping project members focused on the task at hand. Our findings are based on databases prepared with meticulous attention to detail and to accuracy. Catherine never blinked an eye through endless requests for more maps and more artifact tables, and she demonstrated extraordinary leadership; this project was completed wholly due to Catherine’s professionalism.

One of the most exciting products of this project is a web page that contains information on the sites we used in this study, including site summaries, photo galleries, distribution maps, downloadable databases, and interpretive papers. In addition, an on-line searchable database allows site visitors to identify artifacts and the sites from which they come. In the beginning, our ambitions for the web page were minimal, but when project member Greg Brown undertook the development of the web page, he clearly had a much bigger vision. Greg and his colleague, Heather Harvey, also of the Colonial Williamsburg Foundation, designed an especially attractive, user-friendly web page, one to be enjoyed and visited often by anyone interested in the region’s archaeology. Please visit the project web page at www.chesapeakearchaeology.org.

Finally, we are grateful to those archaeologists and their crews who, over the last 20 years, labored in the hot Tidewater sun (or damp Tidewater winter) to recover the material evidence of Indian, European, and African life in 17th- and 18th-century Maryland and Virginia. At least one of these sites — Camden — was excavated in the 1960s, long before the genesis of our own project, and yet, the manner in which archaeological evidence was recovered at Camden and at the 17 other sites used in this project continues to pay dividends, through this study and others as this region of the Atlantic continues to attract scholarly attention. Beverly J. Binns, Thomas E. Davidson, Andrew E. Edwards, Christopher P. Egghart, Garrett Fesler, James G. Gibb, David Givens, Charlie Hodges, Mary Ellen N. Hodges, R. Taft Kiser, Christy Leeson, Nicholas Luccketti, Howard MacCord, Dane T. Magoon, Seth Mallios, Martha W. McCartney, Leslie McFaden, Douglas McLearen, Ruth M. Mitchell, L. Daniel Mouer, Fraser D. Neiman, Alain C. Outlaw, Stuart A. Reeve, Christopher Sperling, Beverly A. Straube, Kit Wesler, and Len Winter are a few of the people we know were out there; we know there were probably many more, and all students of the early colonial Chesapeake will benefit from their efforts.

Contributors

  • Julia A. King, Maryland Archaeological Conservation Laboratory
  • Catherine L. Alston, Maryland Archaeological Conservation Laboratory
  • Gregory J. Brown, The Colonial Williamsburg Foundation
  • Edward E. Chaney, Maryland Archaeological Conservation Laboratory
  • John C. Coombs, Florida International University
  • C. Jane Cox, Anne Arundel County’s Lost Towns Project
  • David Gadsby, Anne Arundel County’s Lost Towns Project
  • Philip Levy, University of South Florida
  • Al Luckenbach, Anne Arundel County’s Lost Towns Project
  • David F. Muraca, George Washington’s Fredericksburg Foundation
  • Dennis J. Pogue, Mount Vernon Ladies’ Association
  • Benjamin J. Porter, St. Mary’s College of Maryland
  • Shawn Sharpe, Anne Arundel County’s Lost Towns Project

Project Consultants

  • Marley R. Brown, III, The Colonial Williamsburg Foundation
  • Cary Carson, The Colonial Williamsburg Foundation
  • John C. Coombs, Florida International University
  • Andrew Edwards, The Colonial Williamsburg Foundation
  • Keith Egloff, Virginia Department of Historic Resources
  • William Kelso, Association for the Preservation of Virginia Antiquities
  • Philip Levy, University of South Florida
  • Fraser D. Neiman, Thomas Jefferson Memorial Foundation

Table of Contents

List of Figures

List of Tables

Part I

Introduction

This report describes the work undertaken by a group of researchers exploring the nature of domestic life in the colonial Chesapeake as revealed through the archaeological record. The project, “A Comparative Archaeological Study of Colonial Chesapeake Culture,” represents a collaborative effort on the part of a number of institutions in Maryland and Virginia to examine the materiality of domestic life as it changed over the course of nearly two centuries. With support from the National Endowment for the Humanities, the Maryland Historical Trust, and the Virginia Department of Historic Resources, project participants had as their primary goal the consideration of three broad issues concerning early Chesapeake society, including:

  • • The organization of plantation labor and the rise of slavery;
  • • Intercultural contact and relations;
  • • Living standards and the consumer revolution.

We were especially interested in documenting and interpreting temporal and geographical variability in the context of these topical issues. We approached these topics by assembling collections catalogs from 18 archaeological sites from Maryland and Virginia that had been professionally excavated, many with public funds of one sort or another. This report describes the work, our results, and future directions for the project. Materials found in this report and more can be found on the web page developed for this project at www.chesapeakearchaeology.org.

The Chesapeake region, consisting primarily of Tidewater Maryland and Virginia, was the first area of intensive English settlement in North America. After several ill-fated attempts in what is now North Carolina, English colonists established a permanent settlement at Jamestown, Virginia in 1607. In part because of the early date of its settlement, the colonial Chesapeake has been the focus of sustained scholarly and popular interest for much of the 20th century. Archaeological research began at Jamestown, St. Mary’s City, and Williamsburg in the first half of the 20th century, with programs continuing at all three sites today (Kelso and Straube 2000; Miller 1983; Brown and Samford 1994; Hudgins 1993). In the 1970s, archaeologists expanded their research to include sites outside the colonial capitals, with much of this work undertaken in response to the region’s land development pressures. As a result, an extraordinary amount of material has been unearthed in the ensuing decades, but a lack of institutional coordination had generally kept the full research potential of this material from being realized. The present project provided resources for that coordination and constitutes one of the first collaborative projects of its kind in the region.

Colonial Chesapeake society formed through the ongoing encounters and interactions of a diverse number of groups of varying ethnic and cultural backgrounds. At contact, Native American tribes were competing with one another for autonomy and even dominance, and the arrival of the English complicated and altered these inter-tribal political relationships. As the English struggled to establish colonial hold, they came with their own varying social backgrounds and soon brought men and women of African ancestry to the Chesapeake. In this setting, Old World understandings of social and cultural difference were de-stabilized and transformed, giving rise to a variety of new cultural forms. By taking a broad comparative approach, we have attempted to describe a landscape that represents the diversity of contacts and the implications these contacts have had for shaping new understandings about social and cultural difference.

We chose the plantation as our primary unit of study because the overwhelming majority of immigrants to the Chesapeake lived on tobacco plantations. English and African colonists, regardless of their economic or legal status, usually found themselves members of plantation households throughout the colonial period. Often, Indian neighbors visited these places to trade, work, or socialize with the residents. We also examine two households occupied by local Indians and located on lands reserved for their use. These findings will enhance future interpretations of plantation sites as well as those found in the relatively urban enclaves of Jamestown, St. Mary’s City, Williamsburg, and Annapolis, where long-term programs have been investigating these settlements (Horning 1995; Kelso 1995, 1996; Kelso and Straube 2000; Kelso, Luccketti, and Straube 1997, 1998, 1999; Miller 1983, 1988; Brown and Samford 1994).

This project is the first step in initiating and then sustaining detailed comparative work in this region of the Atlantic world, accomplishing this in large part by producing and making comparable data sets more easily available. We recognize that this project is by no means the first attempt at comparative study of the Chesapeake. One of our co-authors, Dennis Pogue (1993), has previously examined English standards of living in Maryland and Virginia, and he has also compared tobacco pipe assemblages recovered from sites in southern Maryland (Pogue 1991). Al Luckenbach and his colleagues (Luckenbach, Cox, and Kille 2002) at the Lost Towns Project have similarly compared tobacco pipe assemblages from sites in Anne Arundel County. Henry Miller (1988b) has examined the diet of the English colonists in comparative perspective using faunal assemblages. Cary Carson and his colleagues (Carson et al 1981) produced a comprehensive and comparative study of 17th-century earthfast Chesapeake architecture, linking its impermanence to the relentless demands of the tobacco economy. Anne Yentsch (1991) has compared Chesapeake ceramic assemblages, while Susan L. Henry (1979) and Matthew Emerson (1988) have studied red clay pipe assemblages. There are also the collections of articles concerning the archaeology of 17th- and 18th-century Virginia published by the Archaeological Society of Virginia (Reinhart 1996; Reinhart and Pogue 1993), where themes such as fortifications, house patterns, consumer goods and the standard of living, and the life styles of various ethnic groups are discussed. Lastly, the Digital Archaeological Archive of Comparative Slavery (http://www.daacs.org/), or DAACS, is a powerful and useful web-based research tool that pools the artifact and context information from archaeological sites in the Chesapeake once occupied by slaves; while many of the sites found in DAACS were occupied after the colonial period had come to an end, those that were occupied during the 17th and 18th centuries are revealing slave lifeways in the region.

Our project builds on this earlier work, becoming the first multi-institutional collaborative project that attempts to compare more than one or two material types or classes from a number of sites. In so doing, we build on the earlier findings and gain a deeper understanding of the behavioral patterns and relationships of all inhabitants of the Colonial Chesapeake.

Historical Background

The 1607 arrival of the English occurred at a pivotal point in the political development of Native American society in the Chesapeake as the Powhatan chiefdom emerged as an important force. With contact, two very different worlds collided, and conflict, negotiation, and cooperation all came to characterize encounters between the invading colonists and the Indians. Early interactions tended to be more violent than in later years, even as the English were dependent on Indian corn and other supplies for survival. By mid-century, however, the colonists had gained the upper hand, albeit at tremendous cost to both societies (Fausz 1988; Gleach 1997; Rountree 1993). In Maryland, Lord Baltimore avoided some of this cost by adopting a policy of accommodation. Maryland’s Indians accepted Baltimore’s demand of tributary status and used it — and the colonists’ need for allies and corn — to preserve their way of life well into the late 17th century (Merrell 1979; Jennings 1982).

As the English and their Indian neighbors worked out new understandings of one another, the colonists focused their attention on exploiting the region’s resources. During the first half of the 17th century, the settlers participated in the lucrative fur trade as they took stock of the Chesapeake environment. The colonists also began supplying a seemingly insatiable European appetite for tobacco and, by Maryland’s founding in 1634, annual exports of tobacco amounted to nearly three million pounds. The sot-weed emerged as the foundation of the Chesapeake economy, and the plantation system that developed to support its production would persist well into the 19th century. Even today, the legacy of nearly four centuries of tobacco cultivation remains etched in the region’s landscape (Main 1982; Kulikoff 1986).

For most if not all of the 17th century, English population growth along the tobacco coast was sustained through immigration. Men and women from different parts of England and from different stations in life came to the Chesapeake in search of greater economic opportunity. The overwhelming majority of immigrants were bound laborers, put to work either in the fields growing tobacco or in plantation households as domestic servants. Men outnumbered women and morbidity and mortality were high: if a newly arrived male survived his initial period of “seasoning,” he had a life expectancy of 40 to 45 years. Women (especially pregnant women) and children were similarly at risk. These factors — the predominance of servitude, unbalanced sex ratios, and short life expectancy — imposed limitations on the formation of traditional households in the Chesapeake. Only in the very late 17th and early 18th centuries did a predominantly native born European population begin to emerge in the region (Earle 1979; Horn 1979, 1994; Tate and Ammerman 1979; Carr et al 1991; Rutman and Rutman 1984).

By the late 17th and early 18th centuries, the native population was declining and English immigration was slowing. Beginning in the late 1670s, the price of tobacco fell so steeply that the region’s economy was thrown into a severe and prolonged depression. Tobacco production continued to demand land and labor, however, and during this period, planters made the crucial transition from white servitude to black slavery. Inhabitants of African ancestry, a small minority through most of the 17th century, dramatically increased in number and, by the mid-18th century, exceeded the English population in many localities. The reasons for this conversion to slavery and its economic and social consequences remain a subject of lively debate among colonial historians (Jordan 1968; Sobel 1987; Vaughan 1989; Davis 1997; Breen and Innes 1980; Berlin 1998).

To accommodate the varied backgrounds of the newly arrived inhabitants, the continued presence of Indians, and the harsh realities of frontier life in Maryland and Virginia, old rules were continuously modified and new ones developed. These rules attempted to establish boundaries and define appropriate behavior between planters and laborers, between men and women, and between adults and children. English efforts at replicating their culture were complicated, however, by ongoing interaction with local Indians and with the growing number of people of African origin, interactions requiring a whole new set of rules. Continued encounters between such disparate cultural groups ultimately transformed Old World understandings of social and cultural difference and, in turn, gave rise to a variety of new cultural forms. All of these social negotiations undoubtedly occurred in any number of places, but for many people, the plantation and its household became the most important setting for such encounters (Hudgins 1984; Mouer 1993).

Until recently, the images produced by most historical and archaeological scholarship have tended to represent 17th-century Chesapeake landscapes as populated predominantly if not exclusively by English men and women. These representations suggest little interaction among English settlers, Indians, and Africans (cf. Horn 1994; Carr et al 1991). When researchers have investigated 17th-century Native American culture, more often than not they have located Indians in spaces mostly unoccupied by the English (cf. Fausz 1988; McCartney 1985; Merrell 1979; Potter 1993; Rountree and Davidson 1997). While it is true that both groups used territory and geography to define their place in the emerging colonial world, Indians and Europeans often encountered one another in plantation contexts (Mouer 1993).

Africans were present in the region almost from the beginning of English settlement, but because they were so few in number, little is known about the circumstances in which they lived. Contemporary accounts suggest that Englishmen found Africans strange and unsettling and, though the latter may have constituted a minority throughout most of the 17th century, both the Virginia and Maryland assemblies passed laws defining their place in colonial society (Jordan 1968; Walsh 1997). To understand the rise of race-based slavery in the Chesapeake by the early 18th century, the experiences and encounters of Africans and Europeans in the 17th century must be explored. Indeed, there is a great need for a study of colonial Chesapeake society that returns English, Indians, and Africans to a shared landscape.

The Archaeological Study of Domestic Life

The archaeological study of colonial households has a relatively long history in the Chesapeake, beginning at St. Mary’s City in the late 1970s. Archaeologists examining spatial distributions of architecture and artifacts at several sites in Maryland’s first capital found that, for much of the 17th century, planters, their families, and their (usually European) laborers worked and lived together in shared spaces on plantations. A similar use of space was documented at the Clifts plantation site in Westmoreland County, Virginia. By the end of the 17th century, however, as the switch from servitude to slavery was underway, the growing population of strange new laborers was increasingly housed in separate quarters, often at some remove from the planter’s dwelling. As planters spatially segregated themselves from their laborers, they were in effect reifying the growing social distance they perceived between themselves and their bondsmen. At the same time, rooms within planters’ dwellings were becoming more formalized and specialized in use, relegating slaves to spaces where they were rendered invisible — at least to English eyes. The slaves in turn were able to manipulate space in ways that significantly enhanced their mobility within the plantation landscape, carving out spaces they could control (Keeler 1978; King and Miller 1987; King 1988, 1990; Neiman 1986, 1993; Pogue 1997; Upton 1985).

More recently, archaeologist Matthew Emerson (1988, 1999) has attempted to locate Africans in the archaeological record by identifying specific artifacts believed to have been made by these men and women. Emerson suggests Africans were responsible for the production of many if not all of the local tobacco pipes often found at 17th-century archaeological sites. Similarly, Anne Yentsch (1994) has concluded that glass beads recovered from the Calvert site in Annapolis indicate that slaves working there in the 18th century were able to retain traditional African patterns of adornment. On the other hand, Emerson and Yentsch, as well as other archaeologists, acknowledge that linking ethnicity to artifacts is a complicated and tricky business (Mouer et al 1999; Perry and Paynter 1999).

Fraser Neiman argues that, at least for local pipes, perhaps a more important question is not who made these objects, but rather who used them. Neiman suggests that local pipes were smoked predominantly by servants, and the analysis of their spatial distribution may suggest how labor was organized on early plantations (Neiman and King 1999). Other studies, however, complicate this interpretation. For example, locally-produced pipes are found at nearly every 17th-century site excavated in the Patuxent River drainage, comprising one to four percent of each site’s total pipe assemblage. At one site, Patuxent Point, however, local pipes account for nearly 14 percent of the pipe total and are concentrated in an area that may not have been a locus of servant activity. Bound laborers are known to have resided at all of the Patuxent River sites, so the smaller proportions of local pipes at each site do not necessarily reflect the absence of servants. The strikingly higher percentage of local pipes at Patuxent Point demands explanation, suggesting that the residents of that site had differential access to these objects, and likely a form of social interaction different than what was common elsewhere (King and Chaney 2004).

The distribution of colono-ware bowls at the same Patuxent River sites presents additional problems for consideration. Fragments of these hand-built, unglazed ceramics are usually found in association with food preparation contexts, such as milk houses or in fireplace refuse. Occasionally, colono-ware fragments are also found in association with servant’s quarters. Planters’ wives and their servants were the people who used these vessels in their household work, and their distributions do not coincide with those of local tobacco pipes. These subtleties indicate that variation linked to class and gender, and probably other categories, is recoverable from the archaeological record (Henry 1980; Noël Hume 1962; Veech 1997).

All of these archaeological findings come from a very limited number of sites which, most researchers acknowledge, are skewed toward the wealthy. Such limitations are due in large part to the labor and time intensive nature of archaeological data recovery. A single site often requires months if not years of professional study. As a result, one or two sites have come to ‘stand for’ an entire region. With the investigation of more and more sites, however, archaeologists are recognizing that considerable variability characterized Chesapeake plantation settlements and the archaeological assemblages these settlements left behind. Most of the sites upon which the findings from St. Mary’s City are based, for example, represent village rather than plantation households (Miller 1988), and it is likely that social interactions and relationships in village or town settings differed from those found on plantation sites. There is also growing evidence that intra-regional variability linked to the production of sweet-scented or oronoco tobacco can be found between plantation sites in Maryland and Virginia (cf. Walsh 1999).

After nearly three decades of archaeological research, collections representing archaeological sites from all time periods and localities throughout the region are now available for comparative study. These include materials recovered from sites occupied by servants and tenants as well as by middling and elite households, and sites occupied by English, Africans, and Indians. Curated by a range of organizations and institutions in Maryland and Virginia, many of these collections are publicly accessible, but most organizations lack the resources or mission to undertake large scale, integrated, comparative study. For this project, NEH support provided an overarching framework and support staff by which we examined these collections to test current interpretations of material life in the Chesapeake and develop new ones. Through comparative study, we consider how the men and women who made up this emerging new society — English, African, and Indian — used material culture to define and shape their relationships with one another.

Methods

A primary purpose of this project was to begin a detailed comparison of the material conditions of life in the colonial period, especially the use of domestic space, as revealed by the archaeological record. We began by choosing domestic sites associated with plantations that together represented the span of the colonial period, from about 1620 until about 1750, and that provided geographical coverage, from the James River in Virginia to the Severn River in Maryland. In addition, we selected sites that had been fairly extensively tested, including not just features but plowed midden deposits as well. Eighteen sites were identified that met these criteria and that were publicly available for study; the locations of these sites are shown in Figure 1 and their temporal relationships are depicted in Figure 2.

Figure 1. Location of archaeological sites used in the study.
Figure 2. Temporal distribution of archaeological sites used in this study.

We then began assembling the artifact catalogs for these 18 sites. Many of these catalogs existed in electronic form, while a few were available only on paper.1 Further, while the 6 assemblages from these sites met certain criteria, the excavation programs used at each of the 18 sites varied, sometimes dramatically. In some cases, projects had to be completed within several months, while others represent longer-term programs. Further, once excavated, the assemblages included here were processed and catalogued at a variety of highly respected institutions, each with its own rules and procedures for organizing archaeological data. For the most part, we have accepted these original catalogs and performed no additional checking of the artifacts themselves.

Such an approach is not without problems. For example, we found that the level of detail captured in the electronic and paper catalogs varied among institutions. Terminology also varies. Some catalogs contained obvious — and therefore correctable — errors. Errors less evident, such as typographical ones (keying “0” when “9” was intended, for example), may not have been caught in the process of assembling these databases. The structure of the electronic databases, when they were available, also differed from institution to institution. We made some effort to standardize structure in order to enhance comparability, but the final versions of the individual site databases are not fully equivalent. To have made them so — by forcing them into a uniform template — would have risked the loss of potentially valuable information from certain collections. These databases have been made publicly available as part of the project, and our colleagues are strongly encouraged to peruse downloaded individual databases before subjecting them to additional analyses.

One of the important objectives of the project was to address the substantive issues posed at the beginning of this report through a spatial lens. We were interested not only in the types and quantities of artifacts from each site, but where at the site the materials were used and eventually discarded. We wanted to know which artifacts were closely associated, which were not, and how these materials related to architecture and other yard features at these sites. From these associations, we hoped to infer various types of activities taking place on these plantation homelots, where they were occurring, who was involved, and any changes through time and across geographical space that might indicate changes in the economic or social interactions of plantation residents and their visitors. In search of this information, we selected, for the most part, sites for which a large and reliable sample of the plow zone had been collected. By “large and reliable,” we mean plow zone units ranging in size from no less than 2.5-by-2.5-feet to 5-by-5-feet (or 1-by-1-meter to 2-by-2-meters) with the soil screened through ¼-inch mesh.2 Ideally, plow zone units were relatively closely spaced, with no more than 25 or 30 feet separating the center points of nearby units. These conditions were met for most of the sites used in this study; we elected to use some assemblages for which the plow zone sample was less than ideal in order to achieve geographical and/or chronological coverage.

Artifact distribution maps are useful for presenting visual representations of selected types of materials recovered from these archaeological sites, including their quantities and spatial locations. For this project, most of the distribution maps we produced represent materials recovered from plow zone contexts. In a few cases, spatial distributions of materials recovered from feature contexts were also generated to augment the plow zone evidence. The maps we produced were generated using the SURFER mapping program available from Golden Software, Inc.; however, the methods we used to construct each map are not unique to SURFER. First, it is important to remember that these distribution maps represent density projections based on a set of sampled evidence. The reliability of each map is only as good as the underlying data on which it is based. The SURFER distribution maps we generated represent information contained in specific data points (or excavation units, see below) — the greater the spacing between data points (or excavation units), the less reliable the projected density representation. In addition, studies have shown that systematically-collected data rather than data collected from ‘critical points’ (i.e., areas of high or low concentration) or from randomly sampled points yield more reliable results.

In our maps, data points represent the center point of excavation units. For example, a 5-by-5-foot excavation unit with grid coordinates of N100-N105/E200-205 would have its data point represented as N102.5/E202.5 (center point of the unit). All materials recovered from that particular 5-by-5-foot unit would be assigned to that data point. Not all projects used 5-by-5-foot units, however; some projects used 2.5-by-2.5-foot or 1-by-1-meter units. In a very few cases, excavation units used at the same site varied in size, so we standardized units by using multipliers. For example, at the Sandys site (44JC802), excavators used both 2.5-by-2.5-foot and 5-by-5-foot units. To standardize counts, then, we multiplied quantities of material recovered from 2.5-by-2.5-foot units by 4 (alternatively, we could have divided materials from 5-by-5-foot units by 4). While not a perfect measure, this method allowed data point values to be standardized and therefore more reliably comparable. Finally, at one site, Carter’s Grove 8 (44JC647), many of the materials recovered from the plow zone had been piece-plotted. These materials were handled in our project by creating a grid for the site and aggregating materials by squares superimposed on the site plan.

Next, interval levels were calculated for each map. In other words, what levels of artifact concentration should the maps show? First, the total number of the particular artifact type or class being mapped was calculated. For example, for a map of the distribution of tin-glazed earthenware, we calculated the total number of fragments in the map’s database, along with the mean and standard deviation. These are common statistics and can be easily reproduced by researchers evaluating our data or generating new maps. We believe that these statistics are better than simply “eyeballing” interval levels. Indeed, we have found that using an eyeball method can be difficult to reproduce for the same researcher, let alone others. Interval levels were then calculated by distance from the mean using ½ standard deviation.

For categories of artifacts with small total numbers, we generated maps that showed the location of each object or object fragment. While the small total number may preclude further statistical analysis, the spatial distributions of these materials are still relevant to the questions we posed in this project and are valuable for comparing distributions between sites.

Dozens of maps were generated for each archaeological site. These maps generally included total domestic artifacts, total architectural artifacts, total ceramics, ceramics by type, ceramics by form, total tobacco pipes, white pipes by stem bore diameter, terra cotta pipes, bottle glass, table glass, shell, bone, nails, brick, window glass, and many more artifact types and classes. Of course, if a particular type or class of artifact was not recovered from a site, no map of that type was prepared. Further, while we produced dozens of maps, many more could be generated if a researcher desired using the data sets available on the project’s web page (www.chesapeakearchaeology.org).

Midden Analysis

At most domestic sites in the colonial Chesapeake, refuse was often deposited in yard areas adjacent to doors and windows, and activities conducted in the nearby spaces can often be inferred from materials recovered from the associated middens (cf. King and Miller 1987). This pattern of refuse disposal is clearly evident at the 18 sites used in this study, and we developed a standard method for identifying and then analyzing middens at these sites. This analysis took into account that most sites in the Chesapeake have been subjected to post-occupational agricultural plowing at one time or another, undeniably damaging the archaeological integrity of midden deposits. Nonetheless, the horizontal relationships of materials found in midden deposits are only minimally disturbed, suggesting that information on the use of space within these plantation household sites is recoverable from plow-disturbed midden deposits.

After artifact distribution maps were prepared for each site, we visually compared maps from the same site to identify midden areas. Overlapping concentrations of ceramics, tobacco pipes, glass, and animal bone indicated a potential refuse midden or area where refuse was consistently deposited. Once midden areas were visually identified, plow zone excavation units from within those areas were pulled and aggregated. In some cases, especially at sites occupied for long periods of time, midden areas were difficult to identify. In those cases, sections of yards (i.e., north yard, south yard) have been isolated and analyzed. Although the types and groups of artifacts selected for comparison varied from site to site, in general, the following categories of artifacts were first compared:

Total Domestic Artifacts: This category generally includes white clay tobacco pipes, red clay tobacco pipes, ceramics, case and wine bottle glass, and animal bone, and can provide evidence about the location of consistent refuse disposal at the site. In some cases, Total Domestic Artifacts was calculated with and without animal bone, especially if a site’s plow zone sample was less reliable and most of the evidence for that site came from better preserved feature deposits (such as at Jordan’s Point). Since bone preserves better in feature contexts than in the plow zone, this adjustment allowed us to provide a check when comparing sites with and without plow zone evidence.

Ceramics: Both temporal and functional information are available from ceramics. Ceramics have been organized by type and, when possible, by vessel form. Vessel form distributions represent sherd counts of a particular vessel shape and not individual vessels. For example, two colander sherds would be counted as two colander sherds, even if those sherds represent one vessel. Again, this is not a perfect measure, but reconstructing vessels from plow zone artifacts is difficult, given the damage done to sherds in the plowing process. It may also be difficult if not impossible to determine if two colander sherds are or are not from the same vessel. If a midden yields mostly colander, milk pan, and butter pot sherds, however, this information can be used to argue that a particular deposit derived from the activities of dairying and food processing. Individual vessel data was not available from all sites.

Pipe Stem Bore Diameters: Stem bore diameters of white clay tobacco pipes provide useful chronological information, especially with regard to changing patterns of refuse disposal at a site. The method is not perfect, as it tends to be less reliable for pipe assemblages pre-dating 1660 or those less than 100 stems in number. Nonetheless, the method is important for establishing basic chronological relationships in many cases.

For comparative purposes, raw counts of various material types were converted to percentages; a chi-square test was used to measure strengths of variation and association.

In some cases, such as at Jordan’s Point, the analysis of plowed midden deposits was not attempted because the available data are not sufficiently robust to produce reliable interpretations. In that case, and in a few other instances, materials recovered from spatially distinct feature contexts were compared in an effort to explore intrasite spatial structure.

Only a portion of the maps and charts generated for this project are reproduced in this report. All of the maps and charts are, however, available by site on our project web page (www.chesapeakearchaeology.org).

An Archaeological Narrative

The questions posed in the original grant proposal and at the beginning of this report are broad in scope, but the information for addressing these questions comes from archaeological sites representing plantation households occupied by individual men and women — not social categories. Within these households, individuals made choices or decisions shaped by class, ethnicity, gender, legal status, and religious confession in the context of a colonial situation. The individuals who occupied these sites, many of whom remain unidentified and unnamed, are critically important for documenting the actions, practices, behaviors, and relationships within these households, and, from these individual actions, we may begin to abstract the larger social and historical processes relevant to our questions. In this section, we present an archaeological narrative that provides a brief introduction and biography of each site and how those findings serve to build the context for interpreting the other sites in this study. This narrative is presented chronologically, and should serve to illustrate the diverse nature of the plantation settlements found over the course of nearly two centuries in Maryland and Virginia.

The Earliest Settlements, 1620-1660

The earliest site included in this study was the settlement located at Jordan’s Point, approximately 35 miles north of Jamestown on the south side of the James River at its confluence with the Appomattox River. The property was previously occupied by Weyanoke Indians, one of the groups that formed the Powhatan chiefdom, but the site’s excavators believe that these Indians had abandoned the site by the time Samuel Jordan, his family, and their servants arrived about 1620. Jordan and his family were definitely at Jordan’s Point when Opechancanough led his devastating attack against the English settlers in March 1622. The Jordans survived — the fortified settlement at Jordan’s Point either was not attacked or proved sufficiently protective — but Jordan (who had come to Virginia in 1610) was dead of other causes by 1623. His wife, Cicely, remained at Jordan’s Point, eventually marrying her neighbor, William Farrar, who had relocated to Jordan’s Point after his own property had been attacked in 1622 (Mouer et al 1992; McCartney 1988).3

The muster of inhabitants living at Jordan’s Point in early 1625 suggests that the Jordan-Farrar household was one of several scattered throughout the area, a pattern of settlement also reflected in the number and distribution of early 17th-century sites found in the vicinity. The site complex excavated by Virginia Commonwealth University and the one used in this project, 44PG302, was probably the Jordan-Farrar household. A fortified compound was uncovered at 44PG302, enclosing at least five buildings that appear to have served as dwellings (Figure 3). Five “houses” were listed for the Jordan-Farrar household in the 1625 muster, and these houses would have sheltered William Farrar, Cicely Jordan, Jordan’s two daughters, a third female child, and ten male servants ranging in age from 16 to 35. Some of these buildings were relatively large by early 17th-century standards, with as many as three rooms in up to 880 square feet per structure, but their construction was what architectural historians would describe as impermanent, with principal posts and other framing supports placed directly in the ground at the mercy of wood-devouring insects and wood-damaging cycles of freezing and thawing. Other buildings include domestic or agricultural service structures. The compound was fortified early in the site’s occupation if not at the beginning,4 with access in and out in two places: on the east side through what the excavators have described as a gatehouse and on the west side by a doorway in the palisade’s west wall.

Figure 3. Plan view of structures and other features at Jordan’s Point.

The fortification found enclosing the Jordan-Farrar household complex is not unusual for this time. Wooden fortifications have been found at a number of other contemporary sites, including Jamestown, Wolstenholme Town at Martin’s Hundred, and Flowerdew Hundred. A number of researchers, including the Jordan’s Point excavators, have pointed out the similarities between these Virginia fortifications and similar enclosures erected in Ulster, a 17th-century English settlement in Ireland (Hodges 1992; Mouer et al 1992:50; Noël Hume 1982). Mouer et al (1992:51-52) note how historically documented households living nearby but outside the enclosed Jordan-Farrar compound (and for which archaeological evidence exists) replicates the Irish pattern, and Jordan and later William Farrar might have viewed themselves as “minor barons,” providing a safe, fortified haven in times of attack or other uncertainty.

When the 1625 muster was taken, the Jordan-Farrar household appears to have been well-provisioned with 200 bushels of corn, 200 pounds of fish, 16 cows, and 20 chickens. The household was also well-armed with powder, shot, “fixt pieces,” and “coats of male.” The Jordan-Farrar household was better supplied than most neighboring households, although this is not especially surprising given Cicely Jordan’s and William Farrar’s social and economic standing in the colony. William Farrar was respected in early Virginia for his leadership qualities, having been appointed in March 1626 to the Governor’s Council. He also served as a commissioner “for the Upper Parts kept above Persie’s Hundred,” where he made decisions to hold local court sessions at Jordan’s Point. As a wealthy and free widow, Cicely Jordan enjoyed her own power, attracting at least two suitors (Farrar and a local Anglican minister) who could offer her additional economic and social resources. Indeed, the minister sued Jordan to force her to honor what he thought was her promise to marry him. Jordan instead married Farrar, and only after Farrar and she had been sharing the same settlement for a year or two. That Jordan and Farrar were able to push the boundaries of early 17th-century English senses of propriety suggests the standing the couple enjoyed in frontier Virginia, as well as the reality of establishing English customs in early Virginia (McCartney 1988).

The artifact assemblage recovered from Jordan’s Point is exceptional. Unfortunately, however, only a rapidly collected sample of the site’s surface was made before the plow zone was mechanically stripped, so it is difficult to make statements about the structure and use of space at the settlement.5 Indeed, the nature of these surface materials provides a striking contrast with the materials recovered from feature contexts. Table 1, which presents a summary list of the artifacts from the field’s surface, reveals that the overwhelming majority of materials appear to be associated with the earlier Weyanoke Indian occupation, including stone tools and flakes and ceramics (fire-cracked rock was also recovered in large quantity). Because the site had probably been abandoned by the time Jordan arrived, the majority of the Native American-manufactured materials are not believed to be directly related to the English occupation. Surface materials associated with the English occupation include early 17th-century ceramics, white and red clay tobacco pipes, and colonial bottle glass. These materials, however, do not match the diversity and richness of material recovered from the feature contexts.

ARTIFACT TYPE COUNT ARTIFACT TYPE COUNT
Lithics 2,679 Bottle Glass 25
Indian Ceramics 338 Beads 2
Animal Bone 20 English Flint 28
European Ceramics 59 Brick Fragments 188
White Pipe Fragments 37 Window Glass 4
Red Pipe Fragments 7 Wrought Nail 14
Table 1. Materials recovered from surface collection units at Jordan’s Point.

The artifacts recovered from the feature deposits at Jordan’s Point suggest that, although residents lived in relatively large but flimsy and short‑lived wooden structures, they graced their tables with Late Ming Chinese porcelain bowls and wine cups and German Westerwald sprig­molded stoneware jugs (Mouer et al 1992:117). They consumed tobacco from beautifully crafted Indian tobacco pipes while fortifying their compound from attacks by these same Indians. Thousands of armor fragments, including plate armor, brigandine fragments, and chain mail, as well as gun fragments, gunflints, lead shot, sword parts, and related artifacts suggest the fear that must have existed after Opechancanough’s 1622 attack, although, in the end, most of these materials ended up being discarded. Indeed, some of this armament and weaponry may have been as much about making a statement to other colonists as it was about providing real protection from Indian attack. The artifacts also suggest that in the midst of all this colonial posturing, the everyday domestic activities necessary to the plantation’s success were ongoing, from livestock tending to food processing and preparation.

Exactly which structure(s) the Jordans and later Farrar and the ten male servants occupied is unclear. Structure 4, however, is an especially interesting building (cf. Figure 3). This earthfast building, which served as a dwelling, measured 51‑by‑16‑feet and was constructed early in the site’s occupation. The site’s excavators inferred a passage in the fortified enclosure directly across from Structure 4. Just outside the palisade’s entry was the settlement’s cemetery, consisting of a hodgepodge of burial arrangements scattered along the palisade’s exterior west and north walls. Presumably, this is a portion of the yard the occupants at Jordan’s Point would have wanted some level of regular access to. Individuals entering the enclosure by the cemetery and through this doorway would see two entrances into Structure 4. The first, directly across from the break in the palisade, led into the unheated north room of Structure 4. This room with its cellar was interpreted by Mouer et al (1992) as a buttery. To the right of this service entrance was what appears to be the building’s formal entrance, leading into a small lobby for the reception of visitors before proceeding into rooms on either side of the lobby entrance.

Although Mouer et al (1992) do not consider Structure 4 to be the “manor house” at Jordan’s Point (they give that designation to Structure 1), the materials found in the fill of the northern room’s cellar forced us to reconsider just what function Structure 4 served. The cellar in Structure 4, identified as Feature 320, extended approximately 2.8 feet below subsoil and was completely excavated. Mouer et al (1992) suggest that the cellar stood open for some time after the dismantling of Structure 4, and that eventually what remained of Structure 4 was thrown into the cellar and burned. Strata V‑VIII were deposited prior to the burning episode, and Strata I‑IV during and after. For this analysis, we are including materials recovered from the pre‑burning levels because we believe that these materials were used in or around Structure 4 and represent the material evidence of activities taking place there when the building was occupied (Table 2).6

  SITE TOTAL FEATURE 320 FEATURE 435 FEATURE 409 FEATURE 430
White Pipe 18.6 53.3 20.9 19.8 31.1
Red Pipe 0.5 4.3 0.6 0 0.2
Case Bottle 5.0 4.0 10.4 1.4 19.9
Ceramics 7.2 16.4 12.4 13.2 11.6
Bone 68.7 22.1 55.7 65.7 37.3
COUNTS N=21,457 N=935 N=2,518 N=2,296 N=2,184
           
Armor 2,209 573 80 35 78
Gun artifacts 756 346 49 88 173
Jettons 178 54 14 9 26
Porcelain 23 2 0 2 1
Table 2. Materials from selected features, Jordan’s Point. Top as percentages; bottom as counts.

We acknowledge problems with using cellar fill to make statements about a building’s function, particularly the fact that the cellar was likely filled after the building had been abandoned and was no longer in use.7 With this in mind, we nonetheless observed significant differences between the distributions of artifacts recovered from the cellar fill (Feature 320) in Structure 4 and from the fill of other feature deposits (cf. Table 2).8 Artifacts recovered from the cellar in Structure 4 yielded percentages of white and red clay tobacco pipes and ceramic fragments significantly larger than for the site as a whole. Indeed, the Structure 4 cellar contained more than a third of all the red clay pipe fragments recovered from the site. The cellar also yielded a larger percentage of ceramics than other features, including pieces from a Late Ming porcelain wine cup. And, far greater quantities (in counts) of armor, gun parts, and other weaponry, and of copper alloy jettons were recovered from this cellar than from anywhere else on the site. A silver sequin and a cord of woven fiber wrapped in gold and silver foil were also recovered from the cellar’s fill. Conversely, animal bone was recovered from the cellar in a significantly smaller proportion than that observed for the other features or for the site as a whole.

Mouer et al (1992:54) interpret nearby Structure 20 as a “servant quarter and gate house or trading post” because its west wall appears to have been incorporated into the palisade fortification. Under this scenario, “those living within the fort — specifically, Samuel Jordan and William Farrar — and those living outside the fort servants, other members of the community, and Indians‑came together for common purposes.” For Structure 20 it is difficult to prove or to challenge this interpretation because there are no artifact‑rich features directly associated with Structure 20 and midden deposits were mechanically removed. However, the archaeological evidence does suggest that Structure 4 also served a public purpose, where participants engaged in a greater amount of tobacco smoking with less food consumption surrounded by weapons and armor. Perhaps Structure 4 operated as the space for the court sessions held from time to time at Jordan’s Point. It may have very well been the place where members of outlying households and friendly Indians were received during their visits to the Jordan‑Farrar compound. It may also have served as the residence of Samuel Jordan or William Farrar because there is evidence of a stair behind the lobby entry that would have led to a second story or loft. Without the evidence, it is impossible to rule out (or rule in) Structure 20 as a more public structure, but the archaeological evidence from Structure 4 suggests that this latter building served an important public purpose at Jordan’s Point.

The fortified settlement at Jordan’s Point appears to have been abandoned no earlier than 1630 and possibly as late as 1635, with materials from at least one structure salvaged for reuse, probably elsewhere in the sparsely populated English neighborhood. As Jordan’s Point was abandoned, occupation was underway at Sandys, Buck, and Carter’s Grove 8 (CG 8), small plantation households located farther down the river within six or seven miles of Jamestown. Not one of these three sites was fortified; by the early 1630s, the English at these sites were apparently feeling a bit more secure, perhaps because the settlers were closer to Jamestown or because they had somewhere else to go in case of attack. These three sites were also located east of the Middle Plantation palisade, a structure first conceived in 1626 and finally erected in 1634 in an effort to clear the land south of Jamestown of Indians. The palisade traversed the peninsula from the James to the York rivers and was designed to keep uninvited Indians out of what was now considered by the English to be English territory. While many historians have minimized the significance of the palisade, Philip Levy (2004:265‑287) has argued that the wall’s physical enactment of a policy of racial (or Indian) exclusion was more than a psychological comfort to the Virginia English, and resulted in a very real and dramatic displacement of the region’s native population. Sandys, Buck, and CG‑8 may have been occupied shortly before the palisade’s eventual construction in 1634, with the process of native displacement represented by English settlement already well underway.

Sandys, Buck, and CG 8 appear to represent households drawn from different economic levels in early Virginia society. Sandys, just north of Martin’s Hundred and situated on the James River, was probably occupied by John Wareham, a merchant who also served as a burgess in the Virginia assembly in 1632 and 1633. Wareham was dead by 1638; the site may have been occupied for several years after his death by John Browning (Mallios 2000). The Buck site, located just outside Jamestown approximately a half mile from Back River, was occupied either by descendants of the Reverend Richard Buck (known for officiating at the marriage of John Rolfe and Pocahontas) or by servants of the Buck heirs; the excavators believe the latter although the evidence is not clear (Mallios 1999). In contrast, there is no documentary evidence identifying the residents of CG 8, which was located nearly a mile inland on the Martin’s Hundred tract. The site’s excavators believe CG 8 was occupied by a poor tenant household. For this reason, CG 8 is considered one of the more important sites in this project, given that so little is known archaeologically about the lower economic levels of early Virginia society (Edwards 2004).

Archaeologically, not only were these slightly later sites not fortified, far fewer buildings were uncovered at each than were found at Jordan’s Journey. At Sandys, John Wareham erected two and possibly three buildings, and only one appears to have been heated and used as a dwelling (Figure 4). This building, known as Structure 2, measured 38‑by‑18‑feet and was probably divided into two rooms with the larger room heated. The other two buildings may have been used as storage or agricultural buildings. At CG 8, the unidentified resident(s) occupied a smaller earthfast dwelling measuring 24.6‑by‑16.4‑feet with a lean‑to addition of 6.9‑by­ 10.4 feet (Figure 5). While merchant John Wareham had approximately 684 square feet of enclosed domestic space, the tenants at CG 8 had only about 475 square feet (including the lean­ to addition). This contrasts with the enclosed domestic space apparently available at Jordan’s Point, where Cicely Jordan, William Farrar, Jordan’s children, and ten male servants occupied five houses enclosing 3,828 square feet of domestic space. This measurement does not include attached lean‑to additions, sheds, second stories, or other nearby service structures in the Jordan’s Sandys Site (44JC802)  compound; divided by the 12 adults living in the Jordan‑Farrar household in 1625, about 320 square feet of enclosed living space appears to have been allotted for each adult. Unfortunately, we do not know the sizes of the households at Sandys or CG 8, including whether the householders had families or servants living with them. Based on surveys of early Chesapeake households undertaken by historians, it is likely that at least one other adult did live in these households (Main 1982). Assuming two adults in each household, 342 square feet of living space was allotted per individual at Sandys and 238 square feet at CG 8. More adults are likely, though, suggesting that, after little more than a generation, Virginia planters were building and occupying a diversity of architectural forms.

Figure 4. Plan view of structures, other features, and middens at Sandys.
Figure 5. Plan view of structures and other features at CG 8

Although the occupants of CG 8 may have lived in a smaller structure, the recovery of some brick, albeit in small numbers, suggests that the dwelling had few amenities, at best some brick incorporated into the hearth or chimney. A paling fence also enclosed a small yard adjacent to the dwelling’s south side. Interestingly, the dwelling at Sandys, though bigger than the one at CG 8, also had only a few fragments of window glass and a relatively large number of brick fragments. No intact brick hearths were found, and the brick may have come from Utopia, a later site not far from Sandys.9

Although a large area of ground was exposed at the Buck site, revealing a number of post holes and molds, if a dwelling was situated there, its precise configuration is difficult to identify (Figure 6; Mallios 1999). Four one-cell structures described as “sheds” were inferred from some of the post mold patterns, but these structures — if in fact that’s what they were — only measured 5‑by‑5‑feet in plan. Nonetheless, artifact distributions and three wells provide strong evidence that a dwelling was located in this area. Areas with few artifacts and features indicate that the dwelling may be located in the eastern portion of the excavation, but this is conjectural. The building may have also left no sub-­plow zone trace. Interestingly, not a single fragment of window glass or lead came was recovered from the Buck site, indicating that any building in this vicinity probably did not have glazed windows.

Figure 6. Plan view of features at the Buck site.

When the site assemblages from Sandys, Buck, and CG 8 are compared, the low economic standing of the household at CG 8 is revealed in the raw counts of materials comprising the category Total Domestic Artifacts (Table 3). Just over 2,600 artifacts comprise the Total Domestic Artifacts from CG 8. In terms of counts, the Sandys assemblage has almost four times the number of Total Domestic Artifacts, not surprising for a man who was both a merchant and a burgess. The Total Domestic Artifact count for Buck is mid‑way between Sandys and CG 8. The percentage of ceramics from each site varies, but not as dramatically as the tobacco pipe and case bottle categories. Sandys yielded the largest percentage of tobacco pipe fragments‑36 percent with Buck and CG 8 yielding 16.1 and 18 percent, respectively. Relatively few red clay tobacco pipes were recovered from Sandys, while red clay pipes account for about half the tobacco pipe fragments recovered at CG 8 and the overwhelming majority of pipes recovered from Buck. Case bottle glass was recovered from all three sites, accounting for just over one‑quarter of the total domestic assemblage from Sandys, and nearly half the total domestic artifacts from CG 8 and Buck.

  JORDAN’S POINT SANDYS CG 8 BUCK OLD CHAPEL FIELD
White Pipe 59.8 32.9 8.7 4.8 13.4
Red Pipe 1.7 1.4 9.5 19.2 25.7
Case Bottle 16.0 27.1 48.5 47.2 1.4
Ceramics 23.1 38.6 33.4 28.8 59.5
COUNTS N=6,695 N=10,505 N=2,621 N=4,879  
           
Armor 2,209 342 0 2 0
Weapon artifacts 756 582 22 145 61
Porcelain 23 6 0 1 14
Table 3. Materials recovered from Jordan’s Point, Sandys, CG 8, Buck, and Old Chapel Field. Top set of artifacts presented as percentages; bottom set as counts. Animal bone totals are not included because of differential recovery methods between Jordan’s Point and the other four site assemblages.

The residents at all three sites appear to be armed; weaponry was recovered from each but in varying amounts. Sandys, however, yielded an impressive 582 items, while only 145 and 22 items were recovered from Buck and CG 8, respectively. The majority of these artifacts from all three sites consisted of lead shot. At Sandys, a sword, sword belt artifacts, a dagger, a cannon ball, and eleven gun fragments were also recovered. A snaphaunce fragment was recovered from Buck. Although 19 pieces of shot and two gunflints were recovered from CG 8, no gun artifacts were found, suggesting that any gun in the household was taken care of and did not enter the archaeological record. A sword fragment was, however, recovered. At Sandys, brigandine fragments, chain mail, and other armor artifacts were recovered in large quantity. Not a single fragment of armor was recovered from CG 8, while two pieces were recovered from Buck.

Distributions of artifacts at these three sites are valuable for suggesting how the respective residents were now, post‑1622, organizing their homelots on the English frontier. We’ve already noted that none of these sites was fortified like Jordan’s Point. And, while weapon artifacts were recovered at Sandys, CG 8, and Buck, only Sandys approaches the counts from Jordan’s Point. Similarly, only Sandys and Jordan’s Point have significant quantities of armor artifacts. Wareham may have kept himself armored given his proximity to Martin’s Hundred, which suffered terrible losses in 1622. Still, Jordan’s Point has more than six times the amount of armor as Sandys. With regard to domestic artifacts, the later sites have fewer tobacco pipes and more ceramics and case bottle glass.

At the Sandys site, four refuse middens were identified, and their analysis reveals differential uses of space at this homelot. Middens include a large area of domestic refuse on the western edge of the site (A), two middens on either side of Structure 2 (B and C), and an area behind a large but apparently unheated structure (D) (cf. Figure 4). A possible fifth midden area (E) was identified at the north edge of the site, but given the low artifact counts recovered from this area and the limited testing, we have not used its assemblage in this analysis. All of the refuse middens at Sandys contained types of domestic materials expected at all of the sites, including ceramics, tobacco pipes, bottle glass, and animal bone. Significant variation, however, in midden composition at Sandys reveals how the homelot was differentially used by its occupants.

The compositions of Middens B and C suggest the kinds of activities taking place in Structure 2, the building the excavators identified as the dwelling (Table 4). Midden B is characterized by large proportions of case