FIVE

Activity Areas and Households in the Late Mississippian Southeast United States

Who Did What Where?

RAMIE A. GOUGEON

INTRODUCTION

The topics of household archaeology, activity area analysis, and gender research are combined here to explore production at the household level. Prehistoric Late Mississippian households in the southeast United States comprised men, women, and children performing activities within and around their domestic structures. Outside of the much-debated realm of specialized production of elite or status items, the majority of activities that occurred at the household level were arguably involved in production for domestic needs and consumption. It has been argued, however, that looking at households as “black boxes” hides the contributions of individuals within them (Wilk 1990). It becomes important then, when discussing household production, to examine the role of the individual. One way of accomplishing this goal is through an examination of activity areas and the division of labor by gender.

The analysis of activity areas is an integral part of household production studies. Households comprise activity groups (Ashmore and Wilk 1988; Carter 1984; Netting, Wilk, and Arnould 1984; Wilk and Netting 1984; Wilk and Rathje 1982); thus, it follows that archaeologists excavate the remains of their activities and the loci of activities. The analysis of activity areas can contribute to studies of household production, consumption, craft specialization, and the gender division of activities and space.

This chapter reanalyzes data from three house floors to discern where specific activities occurred within each structure (Gougeon 2002). By drawing on ethnographic and ethnohistoric sources, gender-based activities are identified. These findings are compared with current models of Late Mississippian households, and some implications of gender-based production models within chiefdom-level societies are discussed.

STUDY SITE

The Little Egypt site (9MU102) was located at the confluence of the Coosawattee River and Talking Rock Creek in Murray County, Georgia, but is now under a reregulation reservoir adjoining Carters Lake that was created in 1976 (Hally 1979, 1980) (Figure 5.1). Three physiographic areas meet here—namely the Piedmont, Blue Ridge Mountains, and Ridge and Valley provinces—creating a nexus of a wide variety of natural resources within relatively easy access. Little Egypt is located where the Coosawattee River leaves the Piedmont, crosses the Cartersville Fault, and enters the Great Valley District of the Ridge and Valley provinces. The site is located in a small cove-like valley that is separated from the Great Valley by a line of small hills to the west.

Little Egypt is likely the location of Coosa, the capital village of a chiefdom and a supposed paramount chiefdom occupied during the Late Mississippian Barnett phase (AD 1500–1625). Coosa was visited by Hernando de Soto late in the summer of 1540, as noted by chroniclers of the expedition (Hally 1994; Hudson 1997; Hudson et al. 1985; Langford and Smith 1990). In the decades following de Soto’s visit (AD 1600–1650) the chiefdom collapsed and people along the upper Coosawattee River began a migration to sites in Alabama and later formed the Upper Creeks (Smith 1998, 2001). The Little Egypt site was later reoccupied by Cherokee groups and is referred to in early maps of the region as Coosawattee Old Town, a Cherokee village. The name “Coosawattee” may be a derivation of a Cherokee word kusawati-yi, meaning “old creek place.” EuroAmerican settlement of the area began after 1830.

EXCAVATIONS AND SAMPLE

David Hally conducted archaeological excavations at Little Egypt from 1969 through 1972, followed by a brief revisit in spring 1974. Excavations revealed the remains of three domestic structures and focused on recovering evidence of daily activities of households.

Figure 5.1. Location of Little Egypt (9MU102) Murray County, Georgia (Oakman, GA, 1971 [photo revised 1985], 7.5-minute topographic quadrangle)

The three structures are winter domestic structures, which were constructed in shallow basins. The exterior walls were constructed using single-set posts. Soil excavated from the basin was piled against the exterior of the wall. A thatched roof with a plastered smoke-hole was supported by four central posts, which also enclosed an area of the floor containing the central hearth. The three structures at Little Egypt vary in size from eighty-nine to forty-one square meters. Structures 2 and 3 are located in the village area, while Structure 1 is located on a low terrace on the east side of Mound A (Figure 5.2). Structures 1 and 3 were destroyed by fire. Structure 2 was abandoned and appears to have stood unoccupied for a time before collapsing.

Figure 5.2. Site map of Little Egypt, Georgia, showing locations of mounds and structures.

Part of the challenge of investigating households at Little Egypt stems from the incomplete picture we have of the “household unit” (Netting, Wilk, and Arnould 1984; Polhemus 1987; Sullivan 1987). A household unit in this area of the Late Mississippian Southeast consisted of closely spaced summer and winter structures, outdoor activity areas, and sometimes smaller structures that likely served as storage buildings and shaded work areas. Summer structures varied in size and composition by culture area but are generally marked by a rectangular post pattern that supported a roof (an open-air portico or shed). Winter structures were more substantial and are more easily recognized archaeologically. The household unit has been observed at different sites across the Late Mississippian Southeast and was used during several different phases by several different cultures. For instance, this pattern of paired structures and outdoor activity areas has been found at Dallas, Mouse Creek, and Overhill Cherokee phase sites in eastern Tennessee; various prehistoric Lamar culture sites in northwest Georgia; early historic Cherokee sites in western North Carolina; and historic Creek sites in Alabama.

At Little Egypt only winter domestic structures were excavated. No summer structures, elevated granaries, outdoor activity areas, or any other associated features were excavated near these domestic structures. Identification and analysis of households at Little Egypt is conducted with the understanding that many daily activities and physical features of household units are not available for study. This does not mean, however, that there is not much to learn about Late Mississippian households at Little Egypt. Presumably during the cold winter months many household activities were conducted indoors, in the more substantial and better-insulated winter structures. In a way, winter structures may resemble a microcosm of household activities, with some of the dispersed activities of warmer months brought indoors under one roof. Furthermore, since it is possible that some outdoor activity areas were shared by multiple households (Hally and Kelly 1998; Polhemus 1987), analyzing winter domestic structures almost ensures that individual, independent households performed the activities within them.

METHODS AND TECHNIQUES

Analysis of activity areas within the three Little Egypt domestic structures by David Hally (1980) utilized intuitive pattern recognition techniques. This was accomplished by visually inspecting where clusters of artifacts overlapped. Hally also attempted to identify the gender of the user(s) of some of the activity areas, largely through the use of historic descriptions and ethnographic examples. Visual techniques rely on cognitive abilities to recognize patterns in what are often huge and complex data sets (Kintigh and Ammerman 1982). Archaeologists continue to utilize intuitive visual inspections, though it is now common to apply some data-reduction techniques to simplify the data (Blankholm 1991) or to perform exploratory data analyses (Carr 1991; Tukey 1977). In my research, statistical analyses of artifact classes and the specialized mapping functions of a geographic information system (GIS) were added in an attempt to create a better and more complex picture of household activities in domestic structures (Environmental Systems Research Institute 1998).

The first phase of research was to determine whether there is evidence for discrete, discernable activity areas within the domestic structures. This step was accomplished through a series of analyses, beginning with a review of household activities. The activities carried out in domestic structures for which there is artifactual evidence were discerned by examining classes of artifacts found in association with other classes of artifacts that might have been utilized during a particular activity (Table 5.1). For instance, food-preparation areas are marked by more than just the presence of preserved plant and animal remains. There are also the tools used to process foods, including cutting blades, grinding stones, and percussion tools. Other evidence includes vessels used to store and cook foodstuffs. Studies of specific artifact functions were also considered during this stage of activity area analysis (Conner 1985; Hally 1980, 1983a, 1983b; Pennington 1977).

Table 5.1. Classes of artifacts analyzed from house floors

Second, a Pearson’s r test was used to explore the data for relationships among classes of artifacts. The classes of artifacts from each structure were tested independently. In this way, relationships among classes of artifacts that exist in one or two of the structures would not be blurred or hidden, which might occur if the data sets are examined as a single sample. The Pearson’s r test returns a correlation coefficient, a measure of the strength of an association between two variables (Burt and Barber 1996). I then calculated the coefficient of determination (r²) for artifact classes with strongly correlated distributions (generally r > 0.6). The coefficient of determination is the proportion of the sum of the squares of deviations of the y values about their mean that can be attributed to a linear relationship between x and y (McClave and Dietrich 1985). The coefficient of determination can be thought of as the percentage of variation in y that can be explained by x. With the results of the Pearson’s r test, the distributions of strongly related artifact classes could then be visually examined within each structure using artifact distribution maps generated in ArcView.

Third, the original field maps were digitized into ArcView (ESRI 2000). Each structure was digitized as a separate view, and a separate theme was created for each artifact class within each view. Piece-plotted artifact and feature locations were digitized from original copies of field maps. Artifacts recovered in systematic flotation sampling of each structure were entered in a database. These data were linked to a theme of points marking the location of each sampled excavation unit. Isopleth distribution maps of each class of artifacts recovered through flotation, as well as those that were piece-plotted, were then generated in ArcView (Figure 5.3). By first analyzing the data for strongly correlated artifact classes and then visually examining the distribution of these related artifacts, I defined areas within each structure that were the locations of specific repeated domestic activities.

Figure 5.3. Isopleth distribution maps of classes of artifacts at Little Egypt, Georgia.

Table 5.2. Household activities by age and gender

The second phase of research sought to identify the gender of the individuals working in the discrete activity areas. This required knowledge of not only the tasks commonly performed in domestic structures but also whether each gender was responsible for different household activities. Fortunately for the sake of this study, Southeastern Indian cultures, like many cultures around the world, practiced a somewhat strict division of labor by gender.

The gender division of labor in prehistoric Southeastern societies has been examined in several studies (Polhemus 1998; Smith 1978, Thomas 2001), primarily through the use of data collected by Swanton (1946) and Hudson (1976). Notable among these is Bruce Smith’s (1978) use of data compiled by Murdock and Provost (1973) on gender division of labor from ethnographic studies of 185 societies. For the purposes of my research I have refined Smith’s work to also include butchering and leather working, in addition to tool production and maintenance activities carried out by women (Table 5.2).

THE “WHAT” OF HOUSEHOLD ACTIVITIES AT LITTLE EGYPT

Food preparation is the most common activity performed in domestic contexts at Little Egypt. These activities comprise several stages, from procurement through consumption. There is ample evidence to suggest that some initial processing of plant materials occurred in the houses. Evidence of these types of activities takes many forms, including botanical remains found in conjunction with percussion tools. Some wastes (e.g., large fragments of nutshell and corncobs) may have been reserved for use as fuel (Hally 1981). Tools for removing kernels from cobs include deer mandibles and corncobs. Vessels exhibiting wear from leaching corn with lye and different types of storage vessels are also indicators of plant-processing activities that occurred within domestic structures.

Preparation of game likely included such activities that may have been performed at the kill site, particularly gutting activities, but possibly also including some initial butchering and skinning (Reitz and Wing 1999:204). These activities alter the “completeness” of the faunal assemblage recovered from domestic contexts, introducing another level of complexity to archaeological analysis. However, some evidence for butchering activities can be found in domestic structures at Little Egypt. These include faunal remains found in association with flaked-stone tools (scrapers, blades, projectile points/knives [pp/k]) and some percussion tools.

Cooking food was accomplished in several ways. Stews, soups, gruels, and other forms of boiling plant and animal foods were common in the prehistoric Southeast. Other techniques included roasting, frying, and, to a lesser degree, baking. Evidence for each of these types of cooking techniques takes different forms. For instance, stews and similarly boiled foods required a large-mouthed vessel capable of being placed in or near the direct heat of a fire (Hally 1984). Soot deposits on the sides and shoulders of bowls and jars with wide orifices are strong evidence for this type of cooking. Large vessel fragments were used as griddles as evidenced by distinct sooting and oxidation patterns left on the side of a large pinched-rim jar fragment recovered from Structure 1 (Hally 1983b).

In addition to activities related to food preparation are those activities associated with production and maintenance of flaked-stone tools. Evidence for these types of activities includes concentrations of retouch or tertiary flakes, formal tools in various stages of production (preforms through reworked tools), generalized percussion tools, and specialized stone-working kits. Flint-knapping kits often include specialized pressure flakers of antler tines, round hammerstones, abrading discs for preparing edges prior to flake removal, cores, and preforms.

Figure 5.4. Model of Barnett phase household activity area structuring, Little Egypt

Other activities for which there is evidence within domestic structures at Little Egypt include hide-working and stone-pipe production. Evidence of possible hide-working activities includes specialized tools and particular faunal elements found in Structure 1. Stone-pipe production in Structure 3 is evidenced by the presence of unworked phyllite and broken fragments of stone-pipe bowls/stems in mid-production.

THE “WHO” AND “WHERE”: GENDER AND SPACE IN HOUSEHOLD ACTIVITY ANALYSIS

The distribution of artifacts across areas of the house floors suggests that different activities were carried out in different areas of the structures (Figure 5.4). The pattern seen in the three structures excavated at Little Egypt suggests there was a cultural template or norm for where certain people worked. By extension, this also influenced where particular activities were performed, given the division of labor by gender that existed in prehistoric Southeastern Indian societies (for discussions of gender and the household in other cultures, see chapters in this volume by Beaule, Douglass and Heckman, Gonlin, Henderson, Neff, Snow, and Wiewall). In this way it is particularly difficult to make the argument for discrete activity areas without also discussing the issue of division of labor by gender.

The central hearth area was the focus of many household activities, primarily those requiring heat but also any activities requiring light. This area was kept clean of debris, and we can assume that any large vessel fragments and tools recovered in this area were in use when the structures were abandoned. This being the case, the central hearth area was not a discrete activity area per se but a temporary extension of peripheral work areas located at the openings of the compartments. One can imagine the bulk of traffic within the structure passed through this central area, thereby making permanent work areas a hindrance or hazard and incongruent with the movement of people.

Evidence for division of household space can be found in the construction of the structures themselves (see discussion of the function of architecture in chapters by Ciolek-Torrello and Snow in this volume). Domestic winter structures were physically divided into compartments through the use of partition walls. Partition walls at Little Egypt are represented by posthole alignments extending from the exterior walls toward the center, sometimes with adjacent concentrations of fired daub. Partition walls for which there is no direct evidence can be inferred through analysis of artifact distribution. Walls would have prohibited the even distribution of artifacts across the floor of the structure and would also be areas where tools, vessels, and refuse would have likely been deliberately placed or eventually come to rest. Linear clusters of artifacts in areas that lack direct evidence of structural elements (e.g., postholes or daub) are probable indicators of partition walls. Some of the partition walls proposed for the structures at Little Egypt were inferred in this manner.

Further support for the argument of discrete activity areas by gender can be found within single large compartments. There is evidence to suggest that when two or more individuals of different genders shared a compartment, the division of activity areas was still practiced. In both Structures 1 and 2 a compartment to the right of the entrance appears to have been utilized by a female and a male, perhaps simultaneously. Food waste, raw materials, broken ceramic vessels, tools in various stages of completeness, and other items that might have interfered with the activity being performed were deposited between the two activity areas. This demarcation of a “no-man’s-land” with refuse demonstrates that the two activity areas were viewed as separate from each other. If activity areas were not separate, we would likely observe refuse deposits along walls and partitions exclusively and not in a pattern resembling a partition wall of trash in the center of a large compartment. Where compartments were the locus of single activities or multiple activities performed by a single person, no deposits of refuse are observed in potentially usable space. Here wastes are pushed to the edges of activity areas, along walls, and under benches.

While partition walls served to delineate some male and female areas within Late Mississippian domestic structures, other gender-specific areas were known to be men’s or women’s areas by the individuals who commonly used them (e.g., women preparing food by the central hearth). Upon entering a domestic structure at Little Egypt, visitors would almost assuredly know which areas were used by males and which areas were used by females.

In her study of late prehistoric Siouan communities in the western piedmont of North Carolina, Jane Eastman (2001:58) proposes that women “experienced more profound changes in their gender roles and identities as they aged than did men.” She argues that these changes were marked in many ways, including expectations in behavior, specific dress, and division of labor. These changes may be evident in burials through the presence of tools associated with gender-specific tasks and items of dress or decoration. I suggest that these changes in gender role and identity through a life cycle would also be seen in the spaces they occupy in domestic structures. Children might occupy one common area or sleeping bench while adults would occupy another. However, as girls and boys matured and their responsibilities changed, the areas where they worked and lived would change. For a female this might culminate in the establishment of her own household.

Large compartments that were utilized by one gender usually have evidence of several activities within them, although most of these activities relate to the completion of a larger task. For example, a large activity area for females in Structure 1 contains percussion and grinding stone tools, flaked-stone tools, botanical remains in the form of nutshell fragments and corncobs, and pigment mineral fragments and palettes. Most of these artifacts are related to various stages of food production. The presence of pigments and palettes, however, suggests that other activities may have also occurred in this same area. All of these activities were performed by a woman in a space that was considered to be a female activity area.

The presence of multiple types of activities within a single gender-specific activity area is likely an indication of the age of the person utilizing the space. Adults undertook more types of activities than subadults or children (Eastman 2001). Subadults were taught specific skills by adults, likely within or near adult activity areas. Thus, we might expect to see activity areas with evidence of fewer types of activities within them located apart from more complex adult activity areas. These can likely be interpreted as activity areas of older subadults performing the additional gender-appropriate duties that came with changes in age.

HOUSEHOLD PRODUCTION AT LITTLE EGYPT

What does the exploration of activity areas and gender contribute to the examination of Late Mississippian household production? First, because of the size and location of Structure 1, I assumed there would be some differences in the activities of this presumed elite or higher-status household when compared with the two structures from the village area. Instead, my findings show remarkable similarities among all three households with all performing the same basic household production tasks. Additionally, in spite of the greater size of Structure 1, it appears as though the same cultural template was used to order activity areas and living spaces. David Hally (1981) found no significant differences in the botanical samples from the three structures, suggesting the diets of elite and commoner households were similar. It is possible that status markers at Little Egypt may have been expressed in ways other than those that left evidence in the archaeological record.

Gero (1991:170) states that women are portrayed as the most visible in household contexts, perhaps even “disproportionately represented” in household middens. This appears to be the case at Little Egypt. In all three structures women’s activities dominate the assemblages and occupy the most space. This stands to reason, because upon examination of the lists of activities commonly performed by women in Southeastern Indian societies, one can see that the majority of them take place within or near domestic structures. Men’s activities commonly occurred outside of structures or away from village settings entirely (Spain 1992). A brief discussion of female and male activities follows.

Evidence of food-preparation activities is the most common in domestic contexts at Little Egypt. These activities occur exclusively in female activity areas in all three structures. Whole vessels that functioned as cooking pots and storage containers are found in female activity areas. Partial vessels are also located in these areas and indicate that they functioned as tools (e.g., lids, scoops, and griddles) (Hally 1983a). Other tools, plant parts, and faunal remains are also found primarily in female activity areas. Plant and animal remains found outside of female activity areas arguably represent the consumption of food, as they are not found in association with storage and cooking vessels or processing tools (Hally 1981). It can be argued that activities related to the production of food for domestic consumption occupied not only the bulk of female activity areas but also the majority of women’s time.

Perhaps the most surprising result of this reanalysis of Little Egypt households is the discovery of lithic production areas within female areas of the structures. I propose that women produced and maintained some of the flaked-stone tools found in domestic structures. The evidence for this includes the presence of chert and quartz debris in female work areas, often in association with percussion tools. The notion of female production of flaked-stone tools has been addressed in recent decades, most notably by Gero. She states that it is “inconceivable that they [women] sat and waited for a flake to be produced or that they set out each time to borrow one” (Gero 1991:170). Flaked-stone tools recovered from female work areas at Little Egypt include formal scrapers in several forms and projectile points that have been reworked into specialized scraping and cutting tools.

The user of a tool is the best judge of the adequacy of the tool for a particular task (Gero 1991:170). It stands to reason that if females were making a variety of vessel forms to suit particular needs, so too would female knappers produce points that suited tasks not performed by male knappers. Whether females produced the original tools they later altered through use and resharpening cannot be stated with much certainty. All of the Mississippian point forms are found in male and female areas of the domestic structures at Little Egypt, particularly in the shared production areas. Male knappers sharing this area may have produced generalized cutting tools along with the more finely flaked projectile points used for projectiles, knives, and even exchange. It is possible, though, that women also produced some of the tool forms while working in these heavy processing areas adjacent to male knapping areas. Males and females may have flaked stone with a general understanding or template of the shape the tool would eventually take. This would make distinguishing the points of male and female knappers as difficult as identifying the works of different female potters (Gougeon 2000).

It is also interesting to note that curated formal flaked tools from earlier Archaic and Woodland periods are found primarily in female and shared activity areas. In some cases these tools appear to have been reworked, perhaps by the last Mississippian period users of the tools. The fact that curated tools do not often occur in male activity areas may be a reflection of different cultural attitudes toward flaked-tool production by each gender. Males used projectile points not only for hunting and warfare but also as a medium of exchange with other males (Matthiesen 1994:90, 92). Like males, females used formal tools for specific cutting and scraping activities but do not appear to have exchanged them in the same ways that males did. That is to say, if women were exchanging tools, they were then used as tools by the women receiving them. Finely flaked Mississippian projectile points traded among men are found in male burials, suggesting that they were exchanged for social and not functional reasons (Matthiesen 1994). Women may have viewed Archaic, Woodland, and Mississippian tools as tools and not as objects that could potentially be identified as their handiwork. Certainly males and females recognized some finely flaked points and blades as culturally loaded items that symbolized relationships or alliances among men. The exchange of tools may have been the act that separated these projectile points from nearly indistinguishable copies found in fragments or retouched into new forms in domestic activity areas.

Evidence for male activity areas in domestic structures at Little Egypt is highly geographically limited. A flint-working area containing debris from the production of flaked-stone tools is found in Structures 1 and 2, but this area is not as clear in the final stage of Structure 3. In both Structures 1 and 2, males apparently shared a compartment to the right of the entrance with females. As previously discussed, a line of debris divides the large compartment into two work areas. Flint-knapping kits, or tools commonly associated with them, are found in the male half of the compartment. These smaller activity areas are located away from a larger shared adult bench and might reflect the users’ desire to keep sharp and hazardous flakes out of sleeping areas. This behavior has been noted in an ethnoarchaeological study of refuse disposal among the Lacandon Maya of Chiapas, Mexico (Clark 1991). There knappers worked into a cloth to prevent debris from scattering across other living spaces, usually the kitchen. The debris was collected and removed to out-of-the-way places. All surveyed knappers cited the importance of keeping sharp flakes away from barefeet. The close proximity of a stone-working area and an initial food-processing area in winter structures at Little Egypt is somewhat unexpected. If, however, the foodstuffs coarsely processed in the shared activity area were taken across the structure to a female activity area for cooking and consumption, the hazards of flake debris becoming incorporated into food may have been somewhat mitigated.

MODELS OF LATE MISSISSIPPIAN HOUSEHOLD ACTIVITY AREAS

Several studies have provided models of Late Mississippian household activities and are reviewed briefly here. In Hally’s (1980) analysis of house floors at Little Egypt he suggested that areas of the structures were used for specific activities, including storage, flaked tool production, and food preparation. Gender assignments for particular activities were made based on ethnohistoric accounts. No attempt was made, however, to present a formal model of households for the site, region, or time period.

Through his work at the Toqua site Richard Polhemus (1987, 1990) devised a model of Dallas phase household activity structure. In this model the domestic structure was divided into public and private areas. A central hearth demarcated a public area where a number of activities took place, including preparation of food and activities requiring light from the fire. Private areas consisted of beds and storage areas. Beds were located along the walls. Corners were used for storage, with foodstuffs commonly found in the southeast corner, “non-food” in the northwest corner, and “general” storage in the northeast and southwest corners. Ethnographic accounts report individuals were buried near the bed they used in life. Polhemus used burial placements and associations between genders and specific activities to support claims for engendered areas of the structures. He suggested females were most often associated with the north and south walls, and males were associated with the west wall.

Polhemus (1998) revised this model in his doctoral dissertation, based on analysis of the Loy site. In the new model adult males are associated with the wall opposite the entrance. Adult females utilized the bed and area to the left of the entrance, and subadults were associated with the bed opposite the adult females. Storage areas are assigned to each gender in this refined model. Males utilized the storage area in the right rear corner of the structure, and females used the left rear corner. Food was stored in the front left corner. The right front corner was used for lithic reduction, plant food processing, and other “heavy” or initial coarse processing.

The model of Barnett phase household activity structure suggested by my reanalysis of Little Egypt households is similar to Polhemus’s model. In the diagram of the Barnett phase household model presented here I have utilized some of the terminology presented by Polhemus to facilitate comparisons. As in the model of Dallas phase households, the Barnett phase domestic structure is divided into public and private areas. The area enclosed by the four central roof support posts and containing the central hearth (Area VI) demarcates a public area where a number of activities took place, including preparation of food and activities requiring light or heat from the fire. Private areas consist of those areas along the outer walls containing benches (Areas I, II, III, and IV) and storage areas (Area V).

In the Barnett phase model the compartment immediately to the right of the entrance (Area I) contains both male and female activity areas. In Structures 1 and 2, evidence suggests males used the area adjacent to the entrance, while females and males used the far end of the compartment. In Structure 3 this same compartment appears to have been cleaned prior to the fire that destroyed it, and evidence for these separate areas is sparse. The compartment across from the shared compartment is a female activity area (Area II), as seen in all three structures. It is possible that this second female activity area was utilized by older subadult females, perhaps an older daughter of the female head of the household.

The compartment to the rear and right of the entrance is a shared bench area, likely utilized by the adult male and female heads of the household (Area III). While this area was likely the loci of some activities (e.g., phyllite pipe production in Structure 3), it appears as though the primary activities were eating and presumably sleeping. In the model the compartment to the immediate left of the entrance is associated with subadults (Area IV). Storage areas (Area V) are found in the corners to either side of Area II. Both of these corners appear to have been used to store food items, vessels, and vessel fragments (potential tools) and were also areas where trash accumulated.

Slight differences between Little Egypt households and the model are likely reflections of different household compositions. In Structure 1 the compartment to the immediate left of the entrance was divided into two smaller rooms by a partition wall. These areas might have been utilized by subadults of different genders or by subadults of substantially different ages. No activities related to household production appear to have occurred in either of these smaller rooms. In Structures 2 and 3 this compartment is open but marked by evidence of female activities, suggesting that older female subadults utilized this area. If changes in age and status were also marked by changes in location within domestic structures, it is likely that the compartments to the left and rear left of the entrance (Areas IV and II, respectively) were flexible in their function. Occupants of these areas changed throughout the use-life of the structure (e.g., older daughter moving to occupy the compartment to the rear left of entrance as other children are born and utilize area to the left of the entrance), or the activities that occurred within them changed with the added responsibilities of the occupants.

A comparison of Polhemus’s and my models suggests substantial similarities between Dallas and Barnett cultures. For example, Polhemus (1998:300) identifies a large area just inside the entrance as a heavy processing area used by both males and females. This type of activity area is also seen in Structures 1 and 2 at Little Egypt. Minor differences include where adult female, adult male, and subadult activity areas are located. This appears to be influenced by the placement of the entrance (i.e., mid-wall at Toqua and Loy and at the corner at Little Egypt). The designation of separate areas for household members based on gender, however, is more significant than their specific locations within the structures.

In spite of the small differences in the physical layout of Dallas and Barnett phase structures, the pattern of female activity areas and male activity areas is similar in both models. This similarity is likely a reflection of exogamous marriage structuring, matrilineal and matrilocal principles, the division of labor by gender, and the ubiquity and importance of female activities in households in the Late Mississippian Southeast.

In the Late Mississippian Southeast, rules of exogamy dictated that individuals married outside of their lineage (Hudson 1976). Matrilineal practices likely included a matrilocal postmarital residence pattern, dictating that the husband move into his wife’s household. Apart from the young unmarried sons of the households, adult males were “outsiders” to the lineage. Females attached to the household, including unmarried daughters, grandmothers, and the female head, were all related and part of the matrilineage that gave the household its identity. Females did not marry out of the household. Rather, they formed new households attached to their mother’s, resulting in a pattern of household clusters centered on a common patio-like area (Hally and Kelly 1998; Kelly 1988; Polhemus 1987). As discussed above, females may have occupied several areas of the domestic structure as they matured and took on new roles. An unmarried male likely only occupied one area in his mother’s house, and upon moving into his wife’s house, immediately occupied those areas utilized by the adult male head of the household.

SUMMARY AND CONCLUSIONS

To summarize, through the use of statistical analysis and intuitive pattern recognition techniques, I have demonstrated that there is evidence of discrete activity areas within winter domestic structures at the Little Egypt site. By further examining the activities commonly performed by each gender in Southeastern Indian societies and discerning the tools, materials, processes, and expected artifacts for each activity, I have identified these activity areas with specific genders.

My findings at Little Egypt were used to develop a model of activity-area structuring for Barnett phase households in northwest Georgia. This model is comparable to Polhemus’s models of Dallas phase households in east Tennessee and suggests a wider pattern of activity-area structuring may have been in place in other regions during the Late Mississippian period.

Opening the “black box” of Late Mississippian households has important implications for the study of production within chiefdom-level societies, namely by considering the contributions of individuals to household production activities. As seen in the analysis of house floors at Little Egypt, nearly all of the activities that occurred within domestic winter structures were involved in production for domestic needs and consumption. The evidence strongly suggests that artifacts normally recovered from house floors are associated primarily with activities performed by females, namely those involved with food production. My research on activity areas paints a picture of individuals performing very different tasks for the common good of the household. Within domestic structures, however, it appears that females were the dominant forces behind production.

REFERENCES CITED