INTRASITE FEATURE ANALYSIS of the CRESCENT BAY HUNT CLUB SITE (47JE904), an ONEOTA SITE in SOUTHEASTERN WISCONSIN by James D. M

Home , Oak Forest Site, Porte des Morts

INTRASITE FEATURE ANALYSIS

OF THE CRESCENT BAY HUNT CLUB SITE (47JE904),

AN ONEOTA SITE IN SOUTHEASTERN WISCONSIN

James D. Moss

A Thesis Submitted in

Partial Fulfillment of the

Requirements for the Degree of

Master of Science

In Anthropology

The University of Wisconsin-Milwaukee

May 2010

INTRASITE FEATURE ANALYSIS

OF THE CRESCENT BAY HUNT CLUB SITE (47JE904),

AN ONEOTA SITE IN SOUTHEASTERN WISCONSIN

James D. Moss

A Thesis Submitted in

Partial Fulfillment of the

Requirements for the Degree of

Master of Science

In Anthropology

The University of Wisconsin-Milwaukee

May 2010

______Major Professor Date

______Graduate School Approval Date

INTRASITE FEATURE ANALYSIS

OF THE CRESCENT BAY HUNT CLUB SITE (47JE904),

AN ONEOTA SITE IN SOUTHEASTERN WISCONSIN

James D. Moss

The University of Wisconsin-Milwaukee, 2010 Under the Supervision of Dr. Robert J. Jeske

The Crescent Bay Hunt Club Site is an Oneota site located on Lake Koshkonong in

southeast Wisconsin. This site was occupied circa A.D. 1200-1400. Feature remains representing two wigwam style structures and at least one post-in-trench longhouse structure have been excavated at the site. Using radiocarbon dates and artifact inventories, a Geographic Information System is employed to examine these two distinctly different structure types in order to provide an explanation for their differences.

______Major Professor Date

iii

TABLE OF CONTENTS

Chapter 1...... 1 Introduction ...... 1 GIS Methods and Archaeology...... 2 Oneota in Wisconsin ...... 4 Emergent Horizon A.D. 950 – 1150 ...... 8 Developmental Horizon A.D. 1150 – 1350...... 10 Classic Horizon A.D. 1350 – 1650 ...... 12 Connections to Historical Tribes...... 13 History of Investigations at Lake Koshkonong...... 15 Carcajou Point (47JE002; 47JE812; 47JE813; 47JE814)...... 20 Crabapple Point (47JE93)...... 20 Hearthstone (47JE89) ...... 21 Twin Knolls (47JE379) ...... 22 Crescent Bay Hunt Club (47JE904) ...... 22

Chapter 2...... 29 Crescent Bay GIS Analysis...... 29 Defining functions and ethnographic analogies...... 30 Shovel Probe versus Feature Distribution...... 40 Site Layout ...... 47 Problem ...... 65

Chapter 3...... 67 Radiocarbon Dates ...... 67 Ceramic and Lithic Distribution...... 76 Floral and Faunal Distribution ...... 98

Chapter 4...... 125 Discussion and Conclusion ...... 125 GIS as a Tool for Managing Archaeological Data...... 129

References Cited...... 132

LIST OF FIGURES

Figure 1.1 Oneota around Lake Koshkonong (Shapefile and attribute tables derived from the Wisconsin Historic Preservation Database were provided by the Historic Preservation-Public History Division of the Wisconsin Historical Society; Hearthstone from Rodell (1984)...... 19

Figure 1.2 1887 plat map (H.R. Page & Co. 1887:86)...... 24

Figure 1.3 Topographic relief of the Crescent Bay Hunt Club site...... 26

Figure 1.4 Agricultural potential near the Crescent Bay Hunt Club site (Edwards 2010).

...... 27

Figure 1.5 Location of excavation units on aerial photograph (circa 1996)...... 28

Figure 2.1 Depth (cm) of pit features below plow zone...... 32

Figure 2.2 Feature distribution by type...... 34

Figure 2.3 Formulae for feature volume estimation (McElrath et al. 1987:74)...... 36

Figure 2.4 Distribution of estimated feature volumes in liters...... 38

Figure 2.5 Approximate distribution of positive shovel probes versus areas excavated..41

Figure 2.6 Approximation of site extent based upon positive shovel probes...... 42

Figure 2.7 Distribution of ceramic materials: shovel probe versus unit density...... 44

Figure 2.8 Distribution of lithic debitage: shovel probe versus unit density...... 45

Figure 2.9 Shovel probes containing both ceramics and lithic debitage...... 46

Figure 2.10 1968 map of southeast house (Gibbon 1968). Dotted line represents extent of excavations...... 48

Figure 2.11 1998 block excavations. Unidentified features are from 1968 excavations (Gaff 1998b:Figure 7)...... 49

Figure 2.12 Western house...... 51

Figure 2.13 Central portion of site...... 55

vii

Figure 2.14 Top is a north-south trending trench, interpreted as the western wall of the house. Bottom is an east-west trending trench interpreted as the southern wall of the house...... 56

Figure 2.15 Wall trenches and postmolds...... 58

Figure 2.16 House 3 from Tremaine (Redrawn from O’Gorman 1996: Figure 8)...... 61

Figure 2.17 Central Structure (Longhouse 1)...... 62

Figure 3.1 Spatial distribution of radiocarbon dates at Crescent Bay...... 68

Figure 3.2 Dates from Wood Charcoal (Calibrated with Stuiver and Reimer 1993)...... 73

Figure 3.3 Dates from Annuals (Calibrated by Stuiver and Reimer 1993)...... 73

Figure 3.4 Dates from food residue (calibrated byStuiver and Reimer 1993) ...... 74

Figure 3.5 Nineteen of the 20 Crescent Bay Radiocarbon Dates at 99% Probability (Calibrated by Stuiver and Reimer 1993)...... 75

Figure 3.6 Ceramic density plotted against Liters of soil processed...... 77

Figure 3.7 Debitage density plotted against Liters of soil processed...... 77

Figure 3.8 Debitage density of feature flotation...... 79

Figure 3.9 Total debitage recovered from features...... 80

Figure 3.10 Chipped stone tools recovered from features...... 82

Figure 3.11 Density of ceramics from feature flotation...... 84

Figure 3.12 Ceramics recovered from features by weight...... 85

Figure 3.13 Distribution of Late Woodland ceramics...... 88

Figure 3.14 Distribution of Allamakee...... 89

Figure 3.15 Distribution of Carcajou Ceramics...... 90

Figure 3.16 Distribution of Fisher ceramics...... 91

Figure 3.17 Distribution of Crescent Bay ceramics...... 92

viii

Figure 3.18 Distribution of Busseyville ceramics...... 93

Figure 3.19 Distribution of Grand River ceramics...... 94

Figure 3.20 Distribution of Unclassified Oneota ceramics...... 95

Figure 3.21 House 2 with associated features...... 97

Figure 3.22 Density of animal bone recovered from feature flotation...... 99

Figure 3.23 Distribution of animal bone recovered from features...... 100

Figure 3.24 Density of fish scales recovered from feature flotation...... 101

Figure 3.25 Distribution of fish scales...... 102

Figure 3.26 Density of shell recovered from feature flotation...... 103

Figure 3.27 Distribution of shell...... 104

Figure 3.28 Density of Zea mays cupules and glumes recovered in feature flotation.... 106

Figure 3.29 Distribution of Zea mays cupules...... 107

Figure 3.30 Density of Zea mays kernals recovered in feature flotation...... 108

Figure 3.31 Distribution of Zea mays kernels...... 109

Figure 3.32 Density of Zizania aquatic recovered in feature flotation...... 114

Figure 3.33 Distribution of Zizania aquatic...... 115

Figure 3.34 Density of Chenopodium berlandieri recovered in feature flotation...... 117

Figure 3.35 Distribution of Chenopodium berlandieri...... 118

Figure 3.36 Comparison of feature blocks at Crescent Bay Hunt Club...... 121

Figure 3.37 Locations of human remains found at Crescent Bay...... 124

LIST OF TABLES

Table 1.1 Oneota Horizons (Overstreet 1997)…………………...……………………..…6

Table 2.1 Feature type counts recorded in GIS…………………………………….…….33

Table 2.2 Approximate dimensions of pit features……………………………….……...36

Table 3.1 Radiocarbon dates from the Crescent Bay Hunt Club…………………...……69

Table 3.2 Features containing wild-rice………………………………………………...113

Table 3.3 Distribution of material by area…………………….………………..………122

Table 3.4 Distribution of material by volume…..………………………………………122

Table 4.1 House floor areas…………………………………………………………….128

ACKNOWLEDGEMENTS

I’d like to thank my committee for their professional help and guidance: Dr. Robert

Jeske, Dr. John Richards and Brian Nichols. In addition, thanks to Seth Schneider, Rick

Edwards, and Dr. Patricia Richards for their help and encouragement, and Dr. Robert

Sasso for introducing me to archaeology. Thanks to Mom and Dad for providing parental support and finally a special thanks to all of the members of Anthro-House.

Chapter 1

Introduction

Intrasite feature analysis has long been used by archaeologists to study the distribution of

artifacts and remnants of human activities across a site: “All human activities involving material objects have a spatial component” (Stapert and Street 1997:173). It is the job of the archaeologist to analyze this spatial component and to understand the human behaviors that could account for the observed patterns. Concentrations of particular

classes of materials may indicate special activity areas, or hierarchical social structures,

while even distributions of artifacts may indicate more egalitarian social structures and/or

the lack of task specialization.

The focus of this thesis is an intrasite feature analysis of the Crescent Bay Hunt

Club Site using a Geographical Information System (GIS) to organize, manipulate, and

graphically display data. Artifact densities recovered from prehistoric contexts, as well

as from plow zone contexts, are utilized to examine temporal, functional and spatial

differences within and among activity areas across the site. Of particular interest is the

comparison of two different forms of structures at the site. The feature remains of two

rectangular wigwam type structures (House 1 and House 2) and at least one larger longhouse type structure (Longhouse 1) have been identified at the site. To account for the difference between these markedly different structures, three hypotheses are proposed:

1) The pattern is the result of a seasonal shift in residence patterns, where the

smaller Houses 1 and 2 are single family winter dwellings, and the larger

Longhouse 1 is an extended family summer dwelling (see McKusick 1973;

Skinner 1921).

2) The difference is functional, with Longhouse 1 representing a communal

structure, and Houses 1 and 2, residential ones (see McKusick 1973; Skinner

1921).

3) The two patterns represent a diachronic shift in post marital residence

patterns, where Houses 1 and 2 represent an early, patrilineal post-marital

residence pattern and Longhouse 1 represents a later, matrilineal post-marital

residence pattern (Hollinger 1995).

Using the 20 radiocarbon dates that have been retrieved from the site, along with the distribution of the cultural material, the GIS was utilized to organize the data in order to search for spatial and temporal patterns to determine which, if any, of these three hypotheses best explains the variation in structure forms.

GIS Methods and Archaeology

A GIS was chosen as the best method to interpret the Crescent Bay archaeological data because of its ability to provide “a dynamic and flexible environment within which to integrate, express, analyze and explore the full range of data, both spatial and attribute…. an environment in which to think and explore ideas” (Wheatley and Gillings 2002:18).

One of the primary goals of this thesis was to examine the spatial distribution of different

classes of artifacts across the site, with the goal of providing an explanation, rather than

mere description. The GIS serves as a database management system, allowing the

synthesis of both spatial and attribute data (Heywood, et al. 2006). By using the GIS as a relational database, the user is able to select attribute data quickly and to display it spatially, providing an efficient and flexible means to create multiple maps. Once the GIS and associated databases are established, they can be added to and expanded upon in the future as further excavations and analysis are carried out. Importantly, because it is a relational database, data may be grouped together based on multiple variables, allowing one to assess complex associations, rather than producing simple descriptive maps.

GIS has been utilized in a wide range archaeological contexts, including creating

predictive models (Brandt, et al. 1992), catchment analysis (Hunt 1992), viewshed analysis (Jones 2006) as well as the creation of cognitive models (Maschner 1996). GIS is often employed in intrasite analysis to study the distribution of piece plotted artifacts, usually employing some form of cluster analysis (Kroll and Isaac 1984; Moyes 2002).

There have been few studies that have used a GIS to examine the distribution of the contents of subsurface features within a site (Snell 2005).

With any spatial analysis, it is important to identify the scale and potential problems of the data with which one is working (Stapert and Street 1997). As the result of historic plowing, data on in situ artifacts located on prehistoric ground surfaces that would allow us to perform more advanced spatial analysis has been lost (Moyes 2002; Stapert and

Street 1997). Although all plow zone material has lost its vertical provenience, this

material can be used to examine if it correlates with the features below. The feature fill

contents, which are largely the result of intentional human activities in the past, may be

4 largely intact. Excavations at Crescent Bay suggest that many of the pits found there contain discrete dumping episodes, likely the remains of activities that were carried out in the vicinity. The intentional dumping within many features suggests that an analysis of the contents in and near the architectural structures may indicate the types of activities associated with the structures. If there were different activities being performed near or within each structure type, then different types of refuse would be generated, reflecting either functional or seasonal differences. The GIS’s ability to display the distribution of multiple artifact types in association with radiocarbon dates, and other data, may allow the observation of any temporal or functional differences that may occur between the two structure types.

Oneota in Wisconsin

The term Oneota is broadly used to refer to the material culture of the people that lived in the Upper Midwest during late pre-history. Charles R. Keyes was the first to use the term in an archaeological context, when he described a distinct material culture in Iowa. He chose the term Oneota as it “is the Indian name for the Upper Iowa River, where the culture is most fully represented and has longest been known. This good, original name having been rejected for the river, it may serve here to designate the people that once lived there” (Keyes 1927:224). The term “Oneota Limestone” had been previously used by geologist W. J. McGee in 1891. Oneota also referred to a rock outcrop just north of the mouth of the Upper Iowa, and Henry R. Schoolcraft stated that the term means

“people who sprung from a rock” (McGee 1891:204). Schoolcraft used the term as the title of his collection of Native American notes and manuscripts:

In making this essay, it was thought appropriate that a title for it should be selected from the language of the people, whose history and traits are brought into discussion. The term Oneóta is the name of one of these aboriginal tribes (the Oneidas). It signifies, in the Mohawk dialect, the people who are sprung from a Rock. It is a term which will do as well as any for the entire race, until we obtain better lights (Schoolcraft 1845:v).

Oneota in Wisconsin is divided into four horizons: Emergent, Developmental,

Classic, and Historic (Hall 1962; Overstreet 1976). Though specific dates are often assigned to these divisions (Table 1.1), these are best thought of as relative time periods, and not specific cultural divisions that can be clearly defined. As McKusick (1973:2) points out, these classifications hold little regional continuity and that, “the purpose served seems to be explanation to non-regional specialists rather than providing working definitions for critical analysis of relationships.” In fact, recent research is making it increasingly clear that these horizon designations have little utility for discussing the actual movements and lifestyles of people during the late prehistoric time periods.

However, since these divisions are conventionally used in Wisconsin, they will be described here.

Geographically, Oneota material has been described as far east as Ohio, west into

Nebraska, south into Oklahoma, and north into Ontario and Manitoba (Staeck 1995:4), with the main focus being Michigan, Indiana, Illinois, Missouri, Iowa, Minnesota,

Wisconsin and Nebraska (Overstreet 1997:251). As Oneota does span a large geographic area, it is difficult to describe it as a coherent whole. Henning (1969:318) pointed out

“This is due to the immense spread of Oneota cultural remains over the entire prairie peninsula which results in great difficulty for any one individual to study and encompass the whole of Oneota. One tends to see Oneota development and interrelationships in

terms of the region he knows best.” This task has not gotten any easier in the past 40

years, and so focus of this paper is on Wisconsin, specifically eastern Wisconsin, and it is to this region that is being referred to, when the term Oneota is used.

Table 1.1. Oneota Horizons (Overstreet 1997) Emergent A.D. 950 – 1150 Developmental A.D. 1150 – 1350 Classic A.D. 1350 – 1650 Historic Post A.D. 1650

Oneota is primarily defined by ceramics, and it is based largely on this attribute

that Oneota becomes such a wide-ranging entity in time and space. Oneota ceramics are

generally shell-tempered globular pots with either cordmarked or smoothed surfaces.

Many of these pots are plain, but decorations may consist of a series of rectilinear or

curvilinear patterns, with and without punctates, repeated about the shoulders. Some pots

also contain regular notching of the lip. These traits are recognized as broadly

generalized temporal markers, developing from earlier curvilinear designs, after which

punctates are added, and finally curvilinear changes to rectilinear motifs (Hall 1962:128).

Many of the other artifacts found at Oneota sites are similar to Late Woodland types, but in addition to the pottery certain artifacts are considered particular to Oneota. These include fish lures made from mussel shells, disc pipes, end scrapers, paired sandstone abraders (shaft smoothers), bison and elk scapula hoes, and copper beads and ornaments

(Overstreet 1997:251). Hall points out that, “Outside of ceramics, there are few cultural traits which are both diagnostic of Oneota and found only in Oneota complexes”

(1962:108). Lithic assemblages consist of bifaces, irregularly modified flake tools,

Madison triangular projectile points and small unifacial scrapers (Hall 1962, Overstreet

1997).

Overstreet asserts that the Oneota sites in Eastern Wisconsin represent a cohesive

region, occupying a similar ecological niche, comprising the Koshkonong, Grand River

and Lake Winnebago foci (Overstreet 1976). These sites are located within the 150 frost-

free day zone (with Walker-Hooper on the southern edge of the 140 day frost-free zone),

which would have been even warmer during the Neo-Atlantic Period A.D. 800-1200

(Baerreis and Bryson 1965). A 140-day frost-free period is required for productive maize

horticulture (Yarnell 1964). Yet there is cohesion of Oneota material from east to west

through different ecological niches throughout the western Great Lakes region, where

Eastern Oneota developed within a lake-woods ecotype and those in the Mississippi

Valley and west, developed within a grassland environment (Overstreet 1976:6).

Because of this, Oneota must be defined as more than a successful adaptation to

environmental conditions, as it developed in different ecological environments. There

was likely a cohesive cognitive aspect that accompanied the spread of this style of

pottery, possibly linked to the spread of the importance of maize, but it is stressed that

Oneota is a broad term that is used to define a style of ceramics, and that other aspects of

Oneota culture vary from region to region. What we call Oneota in southeastern

Wisconsin has significant variation in other material life (e.g. subsistence and settlement) than what is called Oneota in neighboring northern Illinois, Iowa, or the La Crosse terrace in western Wisconsin.

The dating of Oneota sites in Wisconsin has been controversial, with relatively few radiocarbon dates. In addition the interpretations of those radiocarbon dates are contested (Boszhardt 2004; Dobbs 1982). More recent research (Edwards 2010;

Schneider 2008) indicates that the horizons defined by Hall (1962) and expanded upon by

Overstreet (1997) are not particularly useful in explaining Oneota culture or chronology.

Emergent Horizon A.D. 950 – 1150

Although Overstreet (1997) has argued that Oneota sites may be as early as A.D. 950, his interpretation is based on conventional radiocarbon dates rather than calibrated dates.

When appropriately corrected, all but one of the early dates have one sigma error terms that overlap A. D. 1150 (R. Jeske 2006). Bozhardt (2004) similarly challenges the interpretation of these early dates. The Emergent Horizon, if the early dates are accepted, overlaps with the rise of Cahokia in the American Bottom, which rose to prominence during the Lohmann phase A.D. 1050-1100, at which point it became an important

Mississippian center (Pauketat and Emerson 1997:5).

There are four phases that are considered Emergent Horizon in Wisconsin: the early Mero Complex on the Door Peninsula, the McKern Phase along the Middle Fox

River Passageway, the Silvernale Phase on Lake Pepin (on the Mississippi River), and

Lake Koshkonong, evidenced by sites at Crabapple and Carcajou Points (Hall 1962;

Overstreet 1997). Apple River in northwest Illinois is also considered as Emergent

Oneota (Emerson 1999). The site of Aztalan begins as a Late Woodland farming village around A.D. 800, and becomes a fortified Mississippian town at A.D. 1100 and is abandoned by A.D. 1250 (Goldstein and Richards 1999; Richards 1992). Aztalan is

viewed as important because it represents an isolated intrusion of what is considered a

more structured, hierarchical Middle Mississippian culture into southeast Wisconsin,

where Oneota is defined as a less hierarchical Upper Mississippian culture. Located on the Crawfish River, a tributary of the Rock River that runs through Lake Koshkonong,

Aztalan is located approximately 24 kilometers northeast of Lake Koshkonong, or

approximately 35 kilometers by river. Because some of the designs on Oneota ceramics,

specifically interlocking scrolls (e.g. Carcajou Curvilinear), appear to mimic those on

Middle Mississippian wares (specifically the Stirling phase type Ramey Incised), it has

been postulated that the former derived from the later, though it is still likely that shell

tempered Oneota ceramics appeared before the intrusion of Aztalan (Hall 2004).

Subsistence patterns during this period include heavy utilization of aquatic

resources, mixed with maize horticulture (though maize is absent at the Mero Complex

sites on the Door Peninsula) and the hunting of large mammals, including deer.

Overstreet (1997:260) summarizes the archaeological remains from this Horizon:

In summary, eastern Emergent horizon components all possess ceramic assemblages that are primarily undecorated except for lip modification: 88-91% plain at Mero, 95% at Carcajou Point, 99% at Crabapple Point, and 98% at Old Spring. At Carcajou Point, Crabapple Point, and Old Spring, the number of end scrappers is quite low. This ratio has significance both for time and geography since the number of end scrapers relative to projectile points increases through time moving to the west (Hall 1962). At Mero and Porte des Morts the data are less compelling because it was not possible to discern which scrapers were Oneota and which were Late Woodland. These criteria – high frequency of plain ceramics, curvilinear motifs, low frequency of end scrapers, and pit houses – are sufficient to differentiate Emergent horizon components from this associated with later horizons in eastern Wisconsin. It is also notable that these early Oneota populations appear to have rapidly replaced local Woodland residents, and there is little evidence of interaction with or transition from one to the other (Hall 1962; Mason 1966, 1993; Overstreet 1989).

In western Wisconsin and Minnesota, the Silvernale phase ceramics are almost all

decorated, and closely resemble Cahokia’s Ramey Incised pottery. The sites around Lake

Pepin are adjacent to large mound groups, many of which contain burials, but mounds are

completely absent from all of the eastern Wisconsin sites (Overstreet 1997).

Developmental Horizon A.D. 1150 – 1350

It has been erroneously proposed that in Eastern Wisconsin, there is a time gap in radiocarbon dates occurring between A.D. 1050 and 1150 (Goldstein and Richards 1991;

Overstreet 1995), which is attributed to the presence of Aztalan. However, others have demonstrated that a correct interpretation of radiocarbon dates shows no evidence for any gap during the period in question, specifically at Carcajou Point and the Crescent Bay

Hunt Club (Boszhardt 2004; Richards and Jeske 2002). The presence of Oneota sites at

Lake Koshkonong during this time period may explain the location of Aztalan further north on the Crawfish River, a tributary to the larger Rock River rather than at Lake

Koshkonong, through which the Rock River runs and which would seemingly have been a more strategic position for the region (Birmingham and Goldstein 2005:53).

The Developmental Horizon overlaps with the Stirling Phase, during which

Cahokia dominated the American Bottom (A.D. 1100 – 1200), the Moorehead Phase

(A.D. 1200 – 1275) which saw the decline of Cahokia, and the Sand Prairie Phase (A.D.

1275 – 1400) in which the Cahokian population largely disappears (Pauketat and

Emerson 1997:8-9).

Developmental Horizon Phases include: the Late Mero Complex and Green Bay

Phase on the Door Peninsula, the Grand River Phase along the Middle Fox Passageway,

The Adams Phase at Lake Pepin, the Early Brice Prairie Phase at La Crosse, and the Lake

Koshkonong Phase represented by: the Crescent Bay Hunt Club, Schmeling, Twin

Knolls, and Carcajou Point sites (Hall 1962; Overstreet 1997). Again, recent research

indicates that these phase names may have little utility in explaining the variation in

Oneota material culture (Schneider 2008).

Archaeological traits of the Developmental Horizon in eastern Wisconsin include

a slight increase in ceramic decorations, where approximately 3-5% are decorated with

curvilinear trailed lines, often bordered by punctates, but less stylized with sloppier

execution than the Emergent vessels (Overstreet 1997:266). Evidence for trade or

significant population movements between eastern and western Wisconsin begins to

appear. Perrot Punctate pottery, catlinite disk pipes and bison scapula hoes, which are

associated with western Wisconsin, are found on eastern sites and Lake Winnebago

Trailed and Koshkonong Bold ceramics, associated with eastern Wisconsin, are found

near La Crosse. Bipolar core lithic technology is more prevalent and there appears to be

a shift from pit houses to wigwams. Subsistence practices are similar to The Emergent

Horizon, with beans being added. A palisade and mound burial are reported at Walker-

Hooper, on Green Lake (Gibbon 1972; Jeske 1927), though as this study indicates, it is

possible that the palisade is misidentified. Overstreet concludes that “there are no radical

modifications of life-ways from approximately A.D 950 to A.D. 1300 (Overstreet

1997:274), which by itself suggests there is little reason for a distinction between

Emergent and Developmental Horizons.

Classic Horizon A.D. 1350 – 1650

During the Classic Horizon Oneota dispersal becomes much more concentrated, being densely clustered along the Middle Fox River passageway and along the La Crosse terrace, with sites being larger villages (Boszhardt 1994; Overstreet 1985). There is also an Oneota presence on Lake Koshkonong, but it appears that occupation is not as extensive as at the other two localities (Hall 1962; Overstreet 1997). By the end of the

Sand Prairie phase (A.D. 1275-1350) the Cahokia system had completely collapsed

(Whiting Young and Fowler 2000:310)

Ceramics from this horizon are distinct from earlier wares, with approximately

90% of vessels exhibiting shoulder decoration: “These motifs most commonly consist of vertical and horizontal lines, alone or in combination. Zones are often bordered by bold, deep punctates. These ceramic styles clearly distinguish the Lake Winnebago and Late

Lake Koshkonong Phases from earlier Grand River Phase ceramic assemblages”

(Overstreet 1997:276). Additional material markers are a higher frequency of end scrapers to small triangular projectile points (Hall 1962), a well-executed 3-or-4 sided bifacial knife, bifacially chipped stone disks of dolomite or poor quality chert, as well as shell implements made from fresh water mussels, catlinite pipes and a higher incidence of bison scapula hoes (Overstreet 1997:276). The higher incidence of end scrapers is attributed to increased processing of buffalo hides. Subsistence patterns appear similar to earlier horizons and there is evidence for garden bed construction (Boszhardt, et al. 1985;

Gallagher, et al. 1985; Sasso, et al. 1985).

Connections to Historical Tribes

“If the Lake Winnebago Phase of Oneota culture does not represent the archaeological remains of the Winnebago tribe, it ought to” (Mason 1993). However, as Mason goes on to point out, there is little evidence of this association. The only site in Wisconsin that has provided substantial evidence of Oneota ceramics in the same context as European artifacts is the Astor site, located on mouth of the Fox River in Green Bay. This single component is the basis for Overstreet’s provisional Dandy Phase (1993), and as Mason

(1993:410) points out, “a phase with only one component is ‘provisional’ indeed.” The

La Crosse terrace was abandoned before the introduction of European artifacts, and it appears that these people moved further west into northeast Iowa and southeast

Minnesota around A.D. 1625, where they are clearly identified with trade goods

(Boszhardt 1994; Stevenson 1994).

The first direct European contact with the Winnebago (Ho-Chunk), Menominee and Potawatomi tribes of Wisconsin occurred in 1634. Jean Nicolet, a French trader sent by Samuel de Champlain, governor of New France, landed at Red Banks near present day

Green Bay.

He wore a grand robe of China damask, all strewn with flowers and birds of many colors. No sooner did they perceive him than the women and children fled, at the sight of a man who carried thunder in both hands, -- for thus they called the two pistols that he held. The news of his coming quickly spread to the places round about, and there assembled four or five thousand men (Thwaites, et al. 1925:279).

Nicolet had been sent to negotiate a peace between the Winnebago (Ho-Chunk) and the

Ottawa, whose hostilities were preventing the French fur trade from expanding into the

Western Great Lakes. Hostilities continued, as refugees from the warfare in the east

poured into the area along with pressure from the Illinois Confederacy in the south.

Through warfare and disease the Winnebago were decimated, and 30 years later, when

French fur traders arrived in Green Bay, they found a village of only about 500

Winnebago, suffering from starvation and disease (Hall 1993; Loew 2001:12-14). “Some

time around the year 1570, a tribal faction that would evolve into the Iowa, Oto, and

Missouri left the main body of Ho-Chunk and headed west down the Wisconsin River to

present-day Iowa, where they separated and evolved into three distinct tribes. The

weakened main body of Ho-Chunk concentrated into large villages near Green Bay in

order to defend their homeland against the encroaching Anishinabe from the north and

the Illinois Confederacy from the south” (Loew 2001:42).

Hall (1993:54) points out that, “The Lake Winnebago phase may well have been

late prehistoric Winnebago, but in its climax it was two centuries too early to be

protohistoric Winnebago.” However, he does provide an argument for Winnebago

symbolic continuity reaching back to the Middle Woodland Havana-Hopewell in eastern

Wisconsin, through the effigy mound culture, which has elsewhere been attributed to

Winnebago ancestors (Hall 1993; Radin 1911; 1970:28).

People who make and do things in similar ways do not always speak the same native tongue. This seems to be particularly true along the borders between major language and cultural groups. If the Oneota way of life was a development by Woodland peoples from ideas that were widespread in the Midwest, this does not mean that this development had to be confined to a single language group. Some Oneota groups have been identified with non-Siouan speaking historic peoples in Illinois, Indiana,

and Missouri (Gibbon 1986:335).

History of Investigations at Lake Koshkonong

The name Koshkonong has been attributed to many different meanings and derivations and has at least 18 different spellings from early accounts, but the one most commonly used is “the lake we live on” (Cassidy 1948). The first occurrence of the name is from

Dr. Jedidiah Morse’s Report to the War Department from the summer of 1829:

On Rock river and its branches they [the Winnebago] have fourteen villages, one of which, the largest, the lowest down Rock river, three hundred miles by water from its mouth, is called Kus-kou-o-nog…. Their village is on the west side of a Lake of the name of their village, six miles long by three wide…. On this lake are three other Winnebago villages (Morse 1822:Appendix 59).

Frederic G. Cassidy, a lexicographer in the English Department at UW-Madison, wrote in 1948, that Koshkonong does not translate in the Siouan dialect of the

Winnebago, and that the phrase “lake we live on” was likely due to cartographer error

(Cassidy 1948:431). Cassidy reasons that from the early spellings, the word is clearly

Algonquian (most likely from a Fox or Ojibwa informer elsewhere in the Rock River

Valley), and that it was the Ojibwa form that was ultimately adopted. Just as the term

“Winnebago” was never used by the Ho-Chunk themselves, Koshkonong was derived from someone else’s name for their village. The conclusion is that the term originally meant something along the lines of “where there is heavy fog”.

The lake was used by Native Americans for thousands of years, and sites dating from Paleo-Indian times, through the present are located along its margins (Brubaker and

Goldstein 1991; Dillemuth 1999; Foley Winkler 2009; Hall 1962; Jeske, Foley-Winkler,

16 et al. 2003; Jeske and Winkler 2008; Musil and Kolb 1987; Richards and Jeske 2002;

Stout and Skavlem 1908). The land around Lake Koshkonong was surveyed between

1834 and 1835, and Euro-American settlers started to move in soon afterwards. One of these early first settlers, Lucian B. Caswell, arrived in the fall of 1837. Caswell was ten at the time. In his memoir, written in 1914, he recalled,

At this season the wild-rice literally covered the entire surface. It looked like a vast meadow. We could see nothing in it for us but from what we could learn it was everything for the Indians. From far and near, they depended largely on the rice they gathered for their winter food; and the ducks (no one can tell or half describe the varieties) came in millions and millions to feed upon the great field of rice of which they seemed so fond. They would light down all over the vast rice fields and feed on the unlimited quantity till they were fat and most delicious food. It was only a question of ammunition in the number one would kill at a shooting. When a gun was fired there followed every time a sight to behold. The noise of the gun would stir them up and they would rise out of this field of rice in such quantities that the roar was like distant thunder. The atmosphere overhead would be filled till the sun at times would be almost darkened. No tongue or pen can describe the number or quantity (Highsmith 1997:2- 3).

In fact, during the late nineteenth century, Koshkonong was said to be one of the best lakes in the country for hunting Canvasback ducks (Frautschi 1945; Highsmith 1997:5).

In 1846 construction began on a dam four miles downstream from the lake at

Indianford. This dam was of wood construction and used to operate a sawmill. The dam was replaced twice by concrete dams in 1917 and 1931 (Sill 2003). “Before the dam near

Indian Ford in Rock County was built, the Rock River, now a wide, slow-flowing stream, then justified its name with large rocks, a gravel bed and a swift current” (Highsmith

1997:3). Increase A. Lapham visited the area in 1850 to survey the mounds around the lake. The resulting Antiquities of Wisconsin, published in 1855 is the first scientific

description of the archaeological remains. At this point in time, there did not appear to

have been much of an effect from the dam upstream, when he describes the lake:

It is eight miles long, with an average breadth of two miles and five eighths; the periphery, measuring all the sinuosities of the shore, is twenty-eight miles and three quarters; the area, twenty-one square miles.. According to the report of Captain T.J. Cram, there is a rapid current, extending about six hundred feet into the lake, with a depth of water of only from two to three feet. In the other portion of the lake, on the usual channel or track for boats and rafts, the water is from four to twelve feet deep. At the time of our visit (July, 1850), wild-rice was growing abundantly over almost its whole surface, giving to it more the appearance of a meadow than a lake. Fish and mollusks also abound in its waters, finding plenty of food in the warm mud beneath, and among the roots and stems of the grass and rushes (Lapham 1855:34).

Thirty years later, Stephen D. Peet also visited Lake Koshkonong to survey its mounds from 1880-1890. His report is published in volume II: Emblematic Mounds and

Animal Effigies in his Prehistoric America series (Peet 1905).

In 1906, with the help of a local resident, H.L. Skavlem, A.B. Stout systematically surveyed the entire lake, including not only descriptions of the mounds, but also descriptions of other archaeological features and artifacts in Volume 7 Number 2 of The

Wisconsin Archeologist, “The Archaeology of the Lake Koshkonong Region” published in 1908. By this time, the Indianford dam had caused the lake to rise significantly, “The gradual raising of the Indian Ford dam, which is some four miles below the outlet of the lake, from its original charted height of 4 feet to its present height of 6 or 7 feet had radically changed the appearance of the lake from what it was in the days of Indian occupation…. Today the wild-rice is confined to the shallower parts of the bays” (Stout and Skavlem 1908:49).

Under the direction of Lynn Goldstein, UWM conducted a systematic survey of the lake using a two mile radius from 1983 to 1986, concluding that Oneota sites are only found on the northwest side of the lake, and along Koshkonong Creek, however, large areas of land are still unsurveyed (Musil and Kolb 1987; Rodell 1984, 1987). To date, six Oneota sites have been identified in this area (Figure 1.1).

Figure 1.1 Oneota sites around Lake Koshkonong (Shapefile and attribute tables derived from the Wisconsin Historic Preservation Database were provided by the Historic Preservation-Public History Division of the Wisconsin Historical Society; Hearthstone from Rodell (1984).

Carcajou Point (47JE002; 47JE812; 47JE813; 47JE814).

Robert Hall’s 1962 book The Archaeology of Carcajou Point, was an expansion of his

1960 dissertation, based on his 1955 to 1959 excavations at the Carcajou Point site.

Carcajou Point (47JE002) is located on the next piece of high ground on Lake

Koshkonong east of the Crescent Bay Hunt Club site. It is in this volume that Oneota is

first divided into Emergent, Developmental and Classic periods. Hall provides a detailed

analysis of the ceramics from the site, including introduction of several new types,

including Carcajou Curvilinear, Carcajou Plain, Busseyville Grooved Paddle and

Koshkonong Bold. The area was surveyed in 1983 by UWM, and three more areas of artifact scatter were identified as containing Oneota debris, but which are likely an extension of 47JE002 (Rodell 1984). A series of contract investigations were performed in 1989, 1990, 1997 and 2002 (Brubaker and Goldstein 1991; Gaff 1998a; Goldstein

1990; Jeske, Hunter, et al. 2003; Richards, et al. 1998). In 2004 a series of shovel tests and excavations were conducted in this area, under the auspices of the Wisconsin State

Tax Credit/Exemption Program by the Office of the State Archaeologist (Rosebrough and

Broihahn 2005).

Crabapple Point (47JE93)

The site is a large habitation site, surrounded by corn hills, conical and linear mounds, as well as one effigy mound. The site has yielded multiple components, including Middle

Archaic, Late Woodland, Oneota and historic Ho-Chunk. The historic material makes up the majority of the site. The site also contained the fur trader Le Sellier’s Cabin. First

identified by Stout and Skavlem (1908), the site was excavated in 1972 by Janet Spector

(1974). Excavations uncovered evidence for both historic Winnebago occupation and prehistoric Oneota occupation, but a transition from one to the other could not be demonstrated. Hall (1962:72) also described some ceramic sherds that had been collected earlier from the site in his Carcajou Point book.

Hearthstone (47JE89)

The initial description of this site comes from Stout and Skavlem but its exact location

was unknown. The area was surveyed in 1983 by UWM which located several scatters of

lithics and shell tempered pottery (Rodell 1984).

Schmeling(47JE833)

This site consists of a field scatter located on a large sandy knoll about 100 meters north

of the Crescent Bay Hunt Club site. It appears that the site was first noted by Stout and

Skavlem (1908). The site was relocated during a UWM survey in 1987, and recorded in

the Archaeological Site Inventory (ASI) files for the state of Wisconsin. Excavations

were carried out by UWM in 2006 and 2008, uncovering a series of features, including

three burials. Two brief publications have been written describing the Paleo-Indian

projectile points collected by the landowner (Jeske and Winkler 2008; Winkler and Jeske

2009) and excavations of the Oneota occupation (Foley Winkler 2009).

Twin Knolls (47JE379)

Originally noted by Stout and Skavlem as the Koshkonong Village site. This site was

located by the 1983-1986 UWM survey, and is represented by a 9.5 acre scatter of Late

Woodland (6%) and Oneota (94%) ceramics along with other archaeological materials and a few outlying lithic scatters likely associated with the site (Musil and Kolb

1987:144). Further investigation and redefinition of the site has been undertaken by

UWM since 2006. Lithic and ceramic analyses of a systematic survey of the site have

been reported (Cowell, et al. 2008).

Crescent Bay Hunt Club (47JE904)

The Crescent Bay Hunt Club Site is an Oneota site located on the northwest shore of

Lake Koshkonong, with a dated occupation of circa A.D. 1200 to 1400. Stout and

Skavlem appear to be referring to Crescent Bay in the following passage, “In this field

and on the adjoining cultivated land the usual indications of an aboriginal village site are

found. Potsherds and burned camp-fire stones are common. To the north of this site, in

WM. Goldthorp’s field, fragments of broken pottery occur in quantity” (1908:80) The

Crescent Bay site sits on what was formerly Goldthorp’s farm, not where the potsherds

and burnt stones were found on Rufus Bingham’s farm to the south (Figure 1.2). This

site was originally designated as JE244 (University of Wisconsin field number).

However 47Je244 belongs to another site in the Archaeological Sites Inventory, resulting

in some confusion among later researchers (Rodell 1984:148). The Crescent Bay Hunt

Club site’s number in the ASI is 47JE904. In 1968, the University of Wisconsin-

Madison excavated a portion of the site over the course of a weekend as part of a fall

field session. Under the direction of David Baerreis, an Oneota house and several nearby

features were uncovered, resulting in an unpublished report by Guy Gibbon (1968), as

well as a Master’s thesis that examined the fish scale and skeletal material (Fortier 1972).

The only published material from these early excavations was the four radiocarbon dates

reported in Radiocarbon (Bender, et al. 1970).

In 1995, University of Wisconsin-Milwaukee (UWM) located a site with a few shell tempered sherds in a small strip of cultivated field in an area thought to coincide with the 1968 excavations, which was labeled site RO-7 (Hanson 1996). In 1998, UWM under the direction of Robert Jeske, definitively relocated Baerreis’s 1968 excavations and have maintained a sustained program of survey and excavation at the site by (Gaff

1998b; Jeske 2000, 2001; Jeske, Foley-Winkler, et al. 2003). In addition to the present effort, four other Master’s theses have been produced using data from the site (Edwards

2010; Foley-Winkler 2004; Hunter 2002; Olsen 2003).

Figure 1.2 1887 plat map (H.R. Page & Co. 1887:86).

The 1998 UWM field school excavation goals were to locate the 1968 excavations and define site boundaries. The crew re-excavated portions of the 1968 house and two features west of the house. It is data from these 1998 excavations that are used in the GIS. The 1968 features not within the 1998 block were added to the GIS, based on Gaff’s map reconstructed from Baerreis’s 1968 field notes and so their position is tentative (Gaff 1998b). Shovel probe data from 1995, 1998 and 2000 indicate that the site measures approximately 160 meters north-south by 140 meters east-west. These dimensions were estimated at 22,400 m2 (5.1 acres) (Jeske 2001:6).

The contemporaneous Schmeling site is located north of Crescent Bay, separated by a draw that was likely a former spring fed wetland that cut between the sites. An

25 approximately 100 meter wide swath of land devoid of cultural material separates the sites. Whether Schmeling was occupied concurrently with Crescent Bay or occupied sequentially is currently unknown. In addition, there are reported Oneota and other archaeological materials from fields south of Crescent Bay and north of Crabapple Point.

Determining the relationship of Crescent Bay to these other materials is beyond the scope of this thesis.

The site occupies the crest of a till-covered limestone bedrock ridge that rises eight meters above the lower wetlands and shore of Lake Koshkonong. This ridge runs along the western shore of Crescent Bay, effectively creating a boundary to the east (Figure

1.3). The site is a prime location for exploiting lacustrine, fluvial, wetland and upland resources and is located on the border between these ecotones and soils ideally suited for horticulture (Edwards 2010) (Figure 1.4). Several factors have affected feature preservation at the site. The entire area has been in cultivation to varying degrees since the mid-nineteenth century leaving only midden remnants and subsurface features. The central portion of the site has been partially protected from cultivation by a 300 x 150 m strip of pine trees (Figure 1.5) that are estimated to have been planted sometime between

1920 and 1940 before the present owners turned the farm into a hunting club (Jeske et al.

2003:20). While the trees cannot be seen in an aerial photo from 1950, the site area appears to be an uncultivated plot surrounded by plowed fields. While the trees protected the site from continued plowing, cultural disturbance and bioturbation associated with pine silviculture has also impacted the cultural materials within the planted pine field.

The northern and southern extents of the site were plowed regularly until circa A.D.

2000, resulting in significant downslope erosion of topsoil to the north, east, and south.

Figure 1.3 Topographic relief of the Crescent Bay Hunt Club site.

Figure 1.4 Agricultural potential near the Crescent Bay Hunt Club site (Edwards 2010).

Figure 1.5 Location of excavation units on aerial photograph (circa 1996).

Chapter 2

Crescent Bay GIS Analysis

The initial task of the project was to convert four 23.5” x 32” hand drawn pencil and

paper site maps into a computerized geographic information system (GIS). The original

data for the hand drawn maps came from field and preliminary maps produced by site supervisors and students over the course of the project. The four site maps were scanned as PDF files on a large format scanner, which were then projected using ESRI’s ArcGIS

9.2. The four maps were geo-referenced to the site coordinate system used by the hand drawn maps, with 500 North, 500 East as the site datum. The site was then geo- referenced to a real world coordinate system (NAD 1983 UTM zone 16), so that topographic maps and aerial photos from the United States Geologic Survey (USGS) could be overlain. The datum coordinates were calculated by averaging numerous readings from Garmin GPS 12 hand held receivers.

Three spatial layers, or feature classes, were created in a Personal Geodatabase, where the datasets are stored within a Microsoft Access data file. Features and units were entered as polygon layers while the shovel probes were entered as points. Using the edit drawing tool, the GIS features were digitized into shapefiles by tracing over the paper and pencil PDFs. When a feature is created within the shapefile in this way, its circumference and area are automatically calculated. It is possible that greater detail of the feature planviews and unit positions could be obtained by importing the individual unit planview sketches into the GIS. It was from these sketches that the site maps were created, but since most features do not have sharply defined boundaries in reality, this method was deemed impractical and tedious, however, going forward, this will likely be

the method for adding new excavation units. The purpose of using the GIS is to be able

to display spatial relationships. However at the level of analysis conducted in this paper, no mathematical calculations were employed to determine relationships and the feature

dimensions used to estimate volume were obtained from the field notes, not the GIS. The

GIS is used to organize the data in order to be able to display artifact distributions

quickly, in relation to the prehistoric structures. Data fields were added for unit and

feature number as well as for feature type.

Defining functions and ethnographic analogies

Archaeologists interpret the function of features as part of their overall interpretation of

past behaviors (Binford 1967). These interpretations are generally informed by

ethnographic analogy, where the archaeologist often infers (or worse, assumes) function

based on a feature’s morphological similarity to historic or ethnographically described

features (Binford 1967; Munson 1969). Ethnographic analogy, however, is useful only as

a basis for making hypotheses that need to be examined using archaeological evidence

(Binford 1967). Keeping in mind Wobst’s (1978) “tyranny of the ethnographic record” it

is critical that we do not let ethnographic observations from a limited time and place

constrain our explanations. Ethnographic analogies provide possibilities for generating

our hypotheses, also known as source side criteria or arguments of relevance (Kuznar and

Jeske 2006:38). They are not useful as explanations for what we find archaeologically

(Binford 1967; Kuznar and Jeske 2006; Stahl 1993). Subject side criteria (e.g. quality of

data, observer bias) are another concern with ethnographic analogy (Kuznar and Jeske

2006:38) and must be taken into account when evaluating our inferences of function from

archaeological forms. The functional labels placed on features used in this thesis bear in

mind the fact that we cannot use analogy from historic sources as interpretations, but

must also relate the archaeological features to other aspects of the site and other lines of

archaeological data to make our inferences.

Archaeological feature types were assigned to each digitized feature based upon morphological characteristics. Larger pit features had previously been designated as

storage pits, multi-use pits, maize storage pits, hearths, wild-rice threshing pits, basins and shallow basins (Jeske, Foley-Winkler, et al. 2003; Mollet and Jeske 2001). These

seven functional/interpretive terms were reduced to three morphological types:

cylindrical pits, basins and shallow basins. Cylindrical pits are circular in planview, with

straight-walls and a flat-floor in profile, the latter being the determining factor, if the

walls are not completely straight as defined. In planview, basins tend to be more

irregular than cylindrical pits, and have parabola-like profiles. Shallow basins are

relatively shallow versions of the basin, arbitrarily defined as extending 10 centimeters or

less below the plow zone. Shallow basins were designated because there does appear to

be a natural break in the depths of features (Figure 2.1) and because little cultural

material is found in them, they are often thought to be the result of non-human activities,

such as dog or wolf pits (Jeske and Kuznar 2001). By comparison, eight morphological

profile types are described for Tremaine, a Classic Oneota site located on the Mississippi

River near La Crosse, Wisconsin (O'Gorman 1995:92). Crescent Bay also lacks

undercut, or bell shaped pits as described at Walker-Hooper (47GL65), a Developmental

Oneota site located along the Grand River in Green Lake County, Wisconsin (Gibbon

1969:63; Hall 1962:16).

Figure 2.1 Depth (cm) of pit features below plow zone.

Four other feature types at Crescent Bay were designated using functional

terminology. Hearths are shallow, defined areas of burnt earth, charcoal and ash. Burials

are defined as a pit containing human remains. Burials are designated for both primary

inhumation and secondary inclusion of partial human remains. Postmolds are defined as

tapered, conical features that are circular to oblong in cross section, while wall trenches

are defined as linear features with flat floors and straight walls with postmolds piercing

into the subsoil below. We know that postmolds are the remains of posts based on

microscopic and chemical cross-section analysis of several samples (Rawling, et al.

1999). In addition to these feature types, eight have been designated as indeterminate

(Table 2.1). Features F68-02, F68-09, have no profile recorded from the 1968 excavations, and features F98-03, F02-13, F06-09, F06-20, F06-178, F06-185 were left unexcavated, due to time constraints.

Table 2.1: Feature type counts recorded in GIS.

Type Number % Basin 35 7.3 Burial 61.3 Cylindrical 26 5.4 Hearth 40.8 Indeterminate 81.7 Postmold 381 79.7 Shallow basin 13 2.7 Wall trench 51.1 Total 478 100

Figure 2.2 Feature distribution by type.

The next step was to calculate an estimate for feature and plow zone volumes.

Following Hunter et al. (2003) geometric formulas were used to estimate volumes for all

non-postmold features, where the cylinder formula was used for the cylindrical features

and the basin formula was used for all other pit feature types (Figures 2.3, 2.4; Table 2.2).

In order to control for differing excavation procedures through the years, and the fact that

different proportions of features were taken for flotation samples, feature artifact density

and unit plow zone density provide for a more accurate unit of comparison, rather than

artifact weights or numbers. For feature artifact densities, liters of sample processed

during flotation was used for the volume. For example, the weight of total lithic debitage

recovered during flotation for a feature was divided by the liters floted for that feature, in

order to produce grams per liter of lithic debitage density. For estimating plow zone

densities, a similar tactic is used, because during the 1998 and 2000 field seasons, the entire 2-x-2 meter unit was screened through ¼ inch mesh, but from 2002 on, only the

southwest quadrant of the unit was screened. In order to be able to compare the

occurrence of lithic and ceramic debris across the site, plow zone density is used, rather

than counts or weights.

Figure 2.3 Formulae for feature volume estimation (McElrath, et al. 1987:74).

Table 2.2 Approximate Dimensions of Pit Features. Notes Length Width Depth Volume Feature Type NEE=not entirely (cm) (cm) (cm) (L) excavated F68-01 shallow basin 120 90 7.5 30 F68-03 basin 110 100 11.5 48 F68-04 basin 145 115 11.5 71 F68-05 basin 15 15 12.5 1 F68-06 cylindrical 35 40 18 16 F68-07 shallow basin 90 85 7 20 F68-08 shallow basin 75 85 3 7 F68-10 indeterminate 200 180 no profile, no location F98-01 basin NEE NEE 15 unexcavated, length F98-03 indeterminate 120 40 estimated F98-06 basin 75 70 2 3 B00-01 burial not mapped F00-01 basin 80 50 32 64 F00-02 basin 60 NEE 50 F00-06 basin NEE NEE 45 F00-07 basin 20 NEE 28 F00-08 shallow basin 47 47 7.5 6 F00-11 cylindrical 120 120 65 735 F00-13 basin 110 55 17.5 41 estimated length F00-14 shallow basin 160 75 5 22 F00-15 cylindrical NEE NEE 30 F00-17 basin 145 110 35 230

Notes Length Width Depth Volume Feature Type NEE=not entirely (cm) (cm) (cm) (L) excavated F00-21 cylindrical 115 95 30 306 estimated width F00-25 cylindrical 85 85 40 221 F00-26 cylindrical NEE NEE 23 F00-48 basin NEE NEE B02-01 burial 160 75 45 246 F02-01 basin 190 150 45 529 overlaps F02-32 F02-03 hearth 95 70 4 9 F02-04 hearth 90 75 12 30 F02-10 basin NEE 90 18 F02-12 basin 100 105 45 222 F02-23 basin 160 170 65 804 F02-25 basin 75 80 18 43 F02-28 cylindrical 140 140 65 1000 F02-31 basin NEE 40 28 F02-32 basin 160 140 40 369 length estimated F02-34 shallow basin NEE 35 10 overlaps F02-01 F02-40 basin 175 130 35 320 F04-01 basin NEE NEE 35 F04-02 shallow basin 125 100 10 47 F04-03 cylindrical 85 85 30 166 F04-04 cylindrical 65 65 42 135 width estimated F04-10 cylindrical 175 160 56 1331 width estimated F04-11 cylindrical NEE NEE 70 F04-12 basin NEE NEE 40 F04-14 cylindrical 125 125 45 543 width estimated F04-15 cylindrical 150 140 40 706 F04-22 cylindrical 170 170 60 1361 width estimated F04-26 basin NEE NEE 45 F04-27 basin 100 70 25 73 width estimated F04-28 basin 80 60 20 40 width estimated F04-33 cylindrical 140 110 35 538 F04-34 cylindrical 145 140 40 651 F04-35 basin 180 145 12 118 F04-36 basin NEE NEE 40 F06-02 cylindrical 150 95 15 265 F06-03 basin 160 140 24 209 F06-05 cylindrical 170 160 38 862 F06-07 cylindrical NEE NEE 38 F06-08 basin NEE NEE 26 F06-10 basin NEE NEE no notes F06-20 indeterminate NEE NEE unexcavated F06-31 shallow basin no notes F06-36 shallow basin 85 70 4 8 F06-39 basin 125 110 16 84 F06-42 shallow basin 30 24 10 3 F06-63 cylindrical 130 135 28 371 F06-64 hearth 185 125 5 43 F06-65 hearth 65 65 5 7

Notes Length Width Depth Volume Feature Type NEE=not entirely (cm) (cm) (cm) (L) excavated F06-172 basin 190 NEE 25 F06-173 cylindrical 180 180 60 1526 F06-185 indeterminate 100 100 unexcavated F08-03 shallow basin NEE 6.5 10 F08-06 shallow basin NEE NEE 10 F08-07 cylindrical 150 150 58 1024 F08-13 cylindrical NEE NEE 32 F08-14 basin NEE NEE 30 F08-20 cylindrical no notes F08-21 cylindrical NEE NEE 55 F08-22 cylindrical NEE NEE no profile F08-31 shallow basin 90 90 10 31 F08-61 basin NEE 55 18

Figure 2.4 Distribution of estimated feature volumes in liters.

Once the site map and feature volume were completed, the artifact inventory forms were entered into Microsoft Excel spreadsheets. One difficulty with this process was that seven different forms have been utilized over the past ten years of artifact analysis, including field tabulation forms, preliminary forms, and final forms-not all of

which are completed for recent years at this time. Recovered artifact classes are currently

undergoing analysis (e.g., lithics, faunal, ceramics), so the counts and weights used here

are necessarily preliminary. Because of this issue, a different spreadsheet was created for

each year with separate tabs for the material that was dry screened in the field and for the

material that was water screened in the lab. These data were then collated and

standardized, so that it could all be incorporated into a single spreadsheet. Another step

in this process was required because different entries for the same feature or unit needed

to be added together, so that a single entry existed for each feature and for each unit. The

“Join” functionality in ArcGIS requires one entry per spatial entity (feature, unit or

shovel probe) in order to work properly. Consequently, feature areas are not differentiated as they were not deemed useful for the level of analysis being conducted.

It appears that most feature areas are individual dumping episodes, either within a single occupation or as part of a sequential set of similar occupations, and thus should still reflect the types of activities being carried out within the vicinity of the two structure types. Feature and unit volumes calculated above were then entered and artifact densities were calculated as grams per liter of soil processed or screened. For several material types (e.g., fish scales and floral materials), counts were used, rather than weights, as the

materials tend to be in discrete units, and are so small that their weight is often negligible.

The Excel spreadsheets were then imported into ArcGIS. The Join function ties an

entry in a spreadsheet to a feature in the GIS by designating a field in both tables which

indicates the same attribute. For example, the field for the archaeological feature number

in both tables is designated, and all other fields in both tables become one entry in the

GIS for that archaeological feature. Once this is accomplished, these data can be exported as a new layer and the joined data are then permanently saved in the GIS.

Shovel Probe versus Feature Distribution

After the features and units were digitized in the GIS, a schematic of the positive shovel probes was layered over the site map in order to give a sense of the extent of the site, as well as the proportion that has been excavated (Figure 2.5). Several open spaces appear to be present among the positive probes (see R. Jeske 2001, 2003). Voids in the archaeological material can be just as essential to the interpretation of the spatial distribution of a site as the distribution of the features themselves. These apparent open spaces so far seem to represent actual voids in the cultural remains, as they have been tested by excavation and there are few archaeological features found within the excavated units in these portions of the site. Using ArcGIS’s measure tool, a line was drawn around the majority of the shovel probes (Figure 2.6) to obtain a site area of approximately

21,500 square meters. Outlying probes that only contained a single flake were left outside of the site boundary. The total area excavated is approximately 700 square meters, which suggests an estimated 3.5 percent of the site has been excavated.

Figure 2.5 Approximate distribution of positive shovel probes versus areas excavated.

Figure 2.6 Approximation of site extent based upon positive shovel probes.

In examining the correlation of positive shovel probes to the features below the plowzone, it must be remembered that the distribution of shovel probes has been overlain on the excavation map as an ideal distribution spaced at ten meter intervals. The actual shovel probes were placed by students pacing out 10 meter intervals, where the local terrain, especially trees, often interfered, so that it is estimated that many of the probes may be off by as much as several meters. Keeping this in mind, it does appear that positive shovel probes do predict areas of features; though the amount of material recovered during shovel probing does not seem to necessarily correlate to areas of high feature density, nor does there appear to be any predictive power to the density of plow zone material (Figures 2.7; 2.8). It is not clear that the distribution of positive shovel probes containing cultural material appears to corroborate in delineating site boundaries, especially the southern boundary of the site. Shovel probes containing both ceramics and lithic debitage have been plotted, and it will be interesting to see if this is a better indicator of site boundaries as more areas are excavated on the periphery (Figure 2.9).

Figure 2.7 Distribution of ceramic materials: shovel probe versus unit density.

Figure 2.8 Distribution of lithic debitage: shovel probe versus unit density.

Figure 2.9 Shovel probes containing both ceramics and lithic debitage.

Site Layout

The orientation of the site appears to be along a northwest to southeast axis,

perpendicular to the edge of the ridge, overlooking the wetlands below (Figure 1.5). It is

interesting to note that the 1968 field school laid out their excavations in the same

manner, to conform with the topography, rather than along a north-south grid. House 1 at

Carcajou Point also appears to be laid out along the same axis (Hall 1962: Plate 4).

Structures

Three main concentrations of features have been excavated. A subrectangular structure

interpreted as house (House 1), denoted by concentric rings of postmolds, occupies the

southeast extent of the site. A second similar house sits on the western portion of the site

(House 2). The third concentration includes a rectangular enclosure in the central portion

of the site (Longhouse 1).

The southeast house was originally discovered in 1968, and portions of it were re-

excavated in 1998 (Figures 2.10 and 2.11). House 1 is about 300 meters from the modern

shoreline. Its dimensions are estimated to have been 4.3-x-4.3 meters. Several unidentified pit features are in the vicinity; however, since no site map exists from the

1968 excavations, this portion of the site has been reconstructed from Baerreis’s notes and the placement of these units is tentative (Gaff 1998b). As mapped they are over 6 meters from the house. Four radiocarbon dates were obtained from the 1968 excavations, all on wood charcoal from features (Bender et al. 1970).

Figure 2.10 1968 map of House 1 (Gibbon 1968). Dotted line represents extent of excavations.

Figure 2.11 1998 block excavations. Unidentified features are from 1968 excavations (Gaff 1998b:Figure 7).

The western house was discovered in 2006. The postmold pattern is very similar to that of the southeastern house (Figure 2.12). The dimensions are estimated to have been

5-x-5 meters. Both houses contained hearths. These structures are interpreted to be wigwam style houses, where a wigwam structure is defined as a series of branches or saplings are anchored in the ground on one end, and then bent over and tied to a second, similar branch placed to form the opposite wall, creating a domed structure, which can be round or square in outline. These structures are Crescent Bay contain additional rows of posts representing interior benches, possibly for sleeping (Gibbon 1968; Skinner

1921:88-93). Similar postmold patterns were found at Carcajou Point, though only small portions of these structures were excavated (Hall 1962).

Figure 2.12 Western house.

Houses 1 and 2 at Crescent Bay appear to be very similar in construction to the

Menominee winter lodges described by Skinner (1921:88-93). These types of houses

were common throughout the Woodlands of Eastern North America (Skinner 1921:85-

86). Construction of these structures starts with a rectangular pattern of saplings placed

into the ground. The poles are then bent and tied together to form a domed house, with

benches added on the interior (Skinner 1921:88-93). This matches very well with the

pattern seen at Crescent Bay. The frames are then covered with mats of cattail flags, or

birch or cedar bark. For bark construction, McKusick (1973:41) describes a double wall

of posts, where the outer wall is used to hold the bark in place. Schoolcraft (1854:56-57)

states that, “The Winnebago use skins, mats made of flags, and bark, for enclosing their

winter lodges”. Thin section and chemical analysis of postmolds profiled during the

1998 excavations revealed that the posts for the walls of House 1 rotted in place, and

were neither pulled from the ground nor burned (Rawling, et al. 1999). Individual cell

walls of the wood post and termite feces could still be seen microscopically, although the

organic component of the wood had deteriorated to less than two percent. The pattern of

the posts representing House 1 appears to be more jumbled than the pattern representing

House 2, and it appears as if wall angles were shifted at some point. This pattern is

interpreted as a rebuilding episode. House 2 is composed of 122 postmolds, while

excavations of House 1 in 1968 yielded approximately 185 postmolds.

As with House 1, there are several exterior features associated with House 2. To

the northwest there are two cylindrical pits within a meter of the outer line of postmolds.

These are the most likely to be directly associated with occupation of the structure because of their close proximity to the exterior wall. A radiocarbon sample from feature

F06-63, obtained from organic residue adhering to the interior of a Winnebago Trailed ceramic vessel rim sherd, yields a date of cal. A.D. 1225 to 1280 at one sigma. To the south there are two more cylindrical pits within three meters of the projected corner of the house with three consecutive basin features between them. Five radiocarbon dates were taken from organic residue on a variety of vessels found in Feature F04-15. These dates range from cal. A.D. 1050 through A.D. 1435 at the one sigma level. The early date was obtained from the residue on a Grand River Trailed vessel and ranges from cal.

A.D. 1040 to 1160 at one sigma. It is a statistical outlier from the other dates at the site at

95% confidence. The range of radiocarbon dates obtained from these surrounding features seems to indicate that this area was reused throughout an extended period of site occupation, though it is very doubtful that the structure was occupied for this entire duration, especially since it shows no signs of rebuilding.

The third concentration of features contains examples of all of the feature classes and represents a dense area of activity (Figure 2.13). The GIS has been especially helpful in analyzing feature distribution in this area. Several wall trenches appear to run parallel to each other, as well as what appears to be an end of a rectangular enclosure. All of the wall trenches contain postmolds occurring every 20 to 30 centimeters, most of which are only visible in profile (Figure 2.14). The wall trenches themselves extend only about 5 to

10 centimeters below the plow zone. The northern end of this concentration extends beyond the protection of the pine trees, and was exposed for a longer period of time to the affects of plowing and erosion. It is likely that the wall trenches extended further north, but only the postmolds remain. In fact, it is possible that the house extended farther north than the postmolds that remain, and what is now interpreted as a corner of the house may

54 be in internal partition. More excavation work needs to be done to determine the precise length of the structure with certainty.

Figure 2.13 Central portion of site.

Figure 2.14 Top is a north-south trending trench, interpreted as the western wall of the house. Bottom is an east-west trending trench interpreted as the southern wall of the house.

The GIS was helpful in filtering out the palimpsest effect of the paper and pencil

map. Until recently, the wall trenches had been interpreted to be a palisade or walled

enclosure, possibly protecting a garden or food processing area from animals. By

selecting the value field for feature type, each type of feature can be displayed or turned

off. Displaying the wall trenches and postmolds within the GIS, while filtering out the other feature types, reveals what appears to be a rectangular structure that was likely

rebuilt at least once (Figure 2.15). If the assumption that the line of postmolds to the

north is an extension of the western parallel wall trench is correct, and taking the line of

postmolds that form a perpendicular 90 degree angle as the northern end, the resulting enclosure would be approximately 7.5 x 18.5 meters. The floor area would be similar in size to ethnographically and archaeologically (McKusick 1973; O'Gorman 1995; Skinner

1921) described longhouses. The term longhouse in this context is meant as a descriptive term for a long, rectangular structure, which because of its size was likely utilized as a dwelling. It is not clear that the structure is equivalent to historic structures that have been traditionally referred to as longhouses (e.g., an Iroquoian longhouse) nor is any such

direct connection implied.

The postmolds extending below the wall trench exhibit two distinct patterns when

viewed in profile. The wall trenches running roughly north-south contain longer, narrow

postmolds, and the wall trench running roughly east-west contain shorter and wider

postmold profiles (Figure 2.14). One interpretation is that the north-south walls were

formed by smaller branches, bent to arch over the centerline of the structure, while the

east-west walls contained stouter posts, forming a vertical wall, so that the resulting structure resembled a Quonset hut. Future work on the postmolds themselves will help

determine vertical orientation of post placement and other aspects similar to the posts

examined from House 1 (Rawling, et al. 1999).

Figure 2.15 Wall trenches and postmolds.

The Tremaine site (47LC95) is used for comparative purposes because it contains

the best record of archeologically described Oneota longhouses in Wisconsin. However,

because Tremaine was occupied at least 100 years later in the Mississippi Valley no

direct connection is implied or inferred. Tremaine is a Classic Horizon Oneota site

located on a terrace overlooking the Mississippi River just north of La Crosse,

Wisconsin, and was occupied between A.D. 1400 to 1500 (O'Gorman 1996:199). The

Tremaine site contains seven longhouses whose dimensions are 7.4 to 8.5 m wide and 25

to 65 m in length (O'Gorman 1996:81). The Crescent Bay enclosure falls short of this

range. In addition to the lack of wall trenches, Tremaine longhouses have rounded ends,

whereas the Crescent Bay structure has squared ends.

One distinct feature that the Tremaine and Crescent Bay structures do have in common are the burials that were placed under the floor of the houses. At Tremaine, most of the bodies were placed perpendicular to the axis of the house, with the heads

pointing toward the centerline (Figure 2.16). The Crescent Bay enclosure has two undisturbed burials closely associated with it (Figure 2.17). The first burial (B02-01,

center of Figure 17) is an adult male that is semi-flexed, holding both a child and an infant. All three bodies are oriented in a northwest-southeast direction, with their heads

pointing towards the northwest, perpendicular to the centerline of the structure. The second burial (B00-01) is a single adult male that lies outside of the outline of the enclosure, but falls within what appears to be a second structure that may have been built over the first (or vice versa). This burial lies perpendicular to the centerline of the hypothetical second structure; however, the head in this burial points west, away from the

hypothesized centerline. Beyond the number of burials present, there appears to be a significant difference in the distribution of features between the two sites. At Tremaine,

the pit features tend to cluster more outside of the structure, but at Crescent Bay, as far as

can be determined, the features cluster inside the structure. The possibility remains that this distribution is the result of the palimpsest effect, and that the pit features are not contemporaneous with Longhouse 1, but because they do appear to cluster within the

boundaries of the walls, until further areas are excavated around Longhouse 1, it is

assumed that some of the pit features were contemporaneous with the structure. Further evidence for association is the placement of burial B00-01 within what is assumed to be a

preexisting pit feature, located within the boundaries of Longhouse 1.

While rebuilding has been noted for the Crescent Bay structure, it must be pointed

out that it is of a different pattern than seen at other sites containing longhouses. At both

Tremaine and Grant, there is evidence of not only rebuilding, but also of repeated

extension of existing structures, which thus far, does not appear to have occurred at

Crescent Bay, though further investigation is necessary (McKusick 1973; O'Gorman

1995).

Figure 2.16 House 3 from Tremaine (Redrawn from O’Gorman 1996: Figure 8).

Figure 2.17 Central Structure (Longhouse 1).

Other Wisconsin Oneota sites that may contain longhouses include Diamond Bluff

(47PI2) in western Wisconsin, which appears to contain at least one structure defined by several series of linear postmolds (Rodell 1997:208). Dambroski (47PT160) in central

Wisconsin contains postmolds that appear to outline an oval structure approximately 6 x

21 meters (Dolan and Kuehn 2006) and Carcajou Point which contains several series of parallel lines of postmolds that were uncovered in excavations from the 1960s, which

Hall hypothesizes represent gabled bark summerhouses (Hall 1962:17). In 2002/2003 another line of posts was uncovered, adjacent to a burial, where the posts appear to be slightly curving, as if it were the end of an oval structure (Birmingham 2004:Figure 2).

It is still possible that the Crescent Bay structure is an artifact of incomplete excavation, as there are significant unexcavated portions in the central and northern areas

(Figure 2.13). However, based on presently available data, this feature is most reasonably interpreted as a longhouse structure.

The two wigwam structures (Houses 1 and 2) identified at Crescent Bay resemble other archaeological and ethnographic examples, but there are no exact matches to be found for Longhouse 1. Post-in-trench wall structures have been observed at Carcajou

Point and the Zimmerman site in Illinois (Brown 1961; Hall 1962). However, both of the rectangular structures are much shorter than the example at Crescent Bay. Hall describes the Carcajou Point structure (15 by 20 ft, or 4.6 x 6.1 m), as a gabled bark summer house

(Hall 1962:17). Hall does not describe the wall trenches in profile, so it is not known if the posts resemble either pattern seen at Crescent Bay (Robert Hall, personal communication, 2008). It is noted that post-in-trench wall construction at Carcajou Point

resembles the wall construction of the semi-subterranean houses at Aztalan and other

Mississippian sites (Hall 1962:20).

Skinner’s (1921) ethnography also describes the construction and use of

longhouses; however, at that time they were exclusively used as ceremonial structures, though it was remembered when they had been used as multifamily homes. When used as houses, the interior had been quartered, or sectioned (Skinner 1921:99-100). By bisecting, and dividing the longhouses at Tremaine into sections, O’Gorman (1996) was able to demonstrate that an uneven feature volume existed between these areas, and argued for uneven resource distribution. As a significant portion of the Crescent Bay structure is unexcavated, this particular analysis cannot be replicated here. Schoolcraft described bark summer lodges as used by the Winnebago. This pattern does not match the Crescent Bay structure, notably the lack of evidence of interior benches.

These lodges are built by setting posts or poles in the ground, and covering them with bark. Ash, elm, and linn, are used for this purpose. The shape of the lodge is similar to that of a log cabin, and differing in size according to the number of persons in the family or families who occupy them. Said lodges are from twelve to forty feet in length, and from ten to twenty feet in height from the ground to the top of the roof. These lodges are built near the field or fields they cultivate, and are occupied for several summers. A lodge forty feet in length,, and sixteen in width, will accommodate three families of ten persons each…. The summer lodge is made of lighter materials, and is portable. When on a hunt to a distant location, the frame of the lodge is left standing, and the covering only is removed (Schoolcraft 1854:56-57).

McKusick (1973:39) and Skinner (1921:96) both note that because the longhouses

were harder to heat in winter, they were only used as warm weather, seasonal dwellings

(though the Iroquois of New York and Canada lived in these types of structures year

round (Skinner 1921:87; McKusick 1973:39). Alternatively, Hollinger (1995) proposes

that smaller structures may represent an earlier (Emergent and Developmental) patrilineal

post marital-residence pattern and that larger structures represent the Classic Oneota

period, during which a matrilineal residence pattern was practiced. This hypothesis better

explains what is traditionally held as the Oneota pattern, where smaller house structures

are associated with Emergent and Developmental Oneota sites in Wisconsin, and larger

structures are noted from the Classic Horizon (Hall 1962; Hollinger 1995; Overstreet

1997).

Problem

The question then becomes, “Why are there two distinctly different structures at the

Crescent Bay Hunt Club Site?” Three possible hypotheses are proposed to account for the different structures.

1) First, the houses could be the result of a seasonal shift in residence patterns, as

ethnographic evidence indicates, where the smaller structures are single family

winter dwellings, and the larger structure is an extended family summer dwelling.

2) The second hypothesis, also modeled after ethnographic evidence, is that the

difference is functional, with Longhouse 1 representing a communal structure,

rather than a residential one.

3) The third hypothesis is that the two patterns represent a diachronic shift in post

marital residence patterns, where the smaller structures represent an early,

patrilineal post-marital residence pattern and the larger structure represents a later,

matrilineal post-marital residence pattern.

Artifact densities and distributions were utilized to test the first and second

hypotheses. If the first hypothesis is correct, then it is expected that a differential distribution of seasonal markers will occur between the depositions in pit features associated with the two different structure types. If the second hypothesis is correct, it is expected that a differential distribution in material types not directly associated with seasonal activities will occur in the discarded material within the pit features, indicating different activity areas. If the third hypothesis is correct, it is expected that a differential distribution will occur in the radiocarbon dates associated with the two types of structures, specifically, that the dates associated with the smaller structures are earlier in time, and that the dates associated with the larger structure are later in time.

Chapter 3

Radiocarbon Dates

There are 20 radiocarbon dates (Table 3.1) available from the Crescent Bay site. The dates, read uncritically at the two sigma level, span roughly from cal. A.D. 1000 to 1500.

These dates span Emergent, Developmental and Classic Oneota horizons; however, inspection of the probability estimates demonstrate that the main occupation of the site falls between cal. A.D. 1200 to 1400. By dividing the radiocarbon date ranges at the 1 sigma level into 50 year blocks, the distribution of the dates can be spatially displayed as if they were faces of a clock, with the earliest dates starting at 12:01 and the latest dates ending at 11:59. Keeping in mind that these date ranges represent a 68% probability, an intra-site pattern does emerge (Figure 3.1).

Figure 3.1 Spatial distribution of radiocarbon dates at Crescent Bay.

Table 3.1 Radiocarbon dates from the Crescent Bay Hunt Club

Conventional Context Material δ13C Cal BC-AD 2δ Reference C14 Age [1035 AD:1225 AD] 0.991231 Grand River [1234 AD:1238 AD] R. Jeske F04-14 880±40 -22 Trailed 0.005857 2008 [1248 AD:1251 AD] 0.002912 [1218 AD:1303 AD] Crescent Bay 0.937079 R. Jeske F04-14 730±40 -23.6 Curvilinear [1365 AD:1383 AD] 2008 0.062921 [1027 AD:1191 AD] Grand River 0.970363 R. Jeske F04-15 920±40 -17.5 Trailed [1196 AD:1207 AD] 2008 0.029637 Winnebago R. Jeske F06-63 800±40 -21.9 [1167 AD:1278 AD] 1. Trailed 2008 [1051 AD:1081 AD] 0.042122 Bender et [1127 AD:1135 AD] al. F68-26 wood charcoal 810±50 0 0.008527 1970:337- [1152 AD:1284 AD] 338 0.94935 [1054 AD:1077 AD] Bender et 0.023401 al. F68-10 wood charcoal 800±50 0 [1154 AD:1287 AD] 1970:337- 0.976599 338 Bender et al. F68-06 wood charcoal 780±50 0 [1159 AD:1293 AD] 1. 1970:337- 338 [1168 AD:1299 AD] Bender et 0.982854 al. F68-01 wood charcoal 760±50 0 [1369 AD:1380 AD] 1970:337- 0.017146 338 R. Jeske F00-26 maize/nut 620±80 0 [1262 AD:1439 AD] 1. 2001a:10 [1221 AD:1308 AD] 0.884361 R. Jeske F00-21 maize/nut 720±40 0 [1362 AD:1386 AD] 2001a:10 0.115639 R. Jeske F00-11 maize/nut 600±70 0 [1280 AD:1432 AD] 1. 2001a:10 R. Jeske F00-06 maize/nut 600±40 0 [1294 AD:1411 AD] 1. 2001a:10

Conventional Context Material δ13C Cal BC-AD 2δ Reference C14 Age [1211 AD:1298 AD] 0.990532 Jeske et al. F02-40 Carcajou Plain 750±40 0 [1371 AD:1378 AD] 2003:79 0.009468 [1269 AD:1299 AD] Grand River 0.931523 R. Jeske F04-14 700±20 -24.7 Plain [1369 AD:1380 AD] 2008 0.068477 maize/nut R. Jeske F04-14 Crescent Bay 590±40 -9.5 [1296 AD:1415 AD] 1. 2008 Curvilinear [1312 AD:1359 AD] maize/nut 0.300209 R. Jeske F04-14 Winnebago 530±40 -19.3 [1387 AD:1444 AD] 2008 Trail 0.699791 [1275 AD:1299 AD] 0.911135 R. Jeske F02-01 Carcajou Type 690±15 0 [1370 AD:1380 AD] 2008 0.088865 [994 AD:1047 AD] 0.762223 Grand River [1089 AD:1122 AD] R. Jeske F04-22 990±20 0 Type 0.191527 2008 [1139 AD:1149 AD] 0.04625 [1229 AD:1232 AD] Grand River 0.009941 R. Jeske F04-35 745±20 0 Type [1238 AD:1286 AD] 2008 0.990059 Busseyville R. Jeske F04-03 785±15 0 [1221 AD:1269 AD] 1. Paddle Grooved 2008

The dates from the southern portion of the site range from cal. A.D. 1200 to 1300.

These dates were obtained from wood charcoal, and possible skew early because of the problem of dating old wood, or “presample-growth error” (Ralph 1971). Sampling wood from different parts of the tree will produce different dates, because 14C is fixed during the yearly ring growth (Taylor 1987:44-47). Wood from the inner tree rings yields earlier dates, while wood from outer tree rings will yield later dates.

Feature F68-10 has not been relocated, but it was situated near the southeastern house. The dates are only indirectly associated with the southeastern structure. The dates from the southeastern block excavations are matched by a single date associated with

House 1, dated at cal. A.D. 1200 to 1300. The two dates from the eastern part of the site also fall within the cal. A.D 1200 to 1300 time range.

The central portion of the site dates a little later, from cal. A.D. 1250 to 1400, with a single anomalous date of cal. A.D. 1000 to 1050 from a Grand River Trailed

vessel found in feature F04-22. The central longhouse does appear to have been a later

occupation, which supports the hypothesis for a diachronic difference in house types

stemming from a shift in postmarital residence patterns. Feature F04-14 appears to be cause for concern as the dates for this feature are derived from radiocarbon samples from

five discrete dumping episodes that range from cal. A.D. 1050 to 1450. However, R.

Jeske (2008) has argued based on pit profiles and the observations of Skavlem (1908),

that most pit features were formed by initial excavation and use, were then used as refuse

dumps, or natural siltation infill followed by refuse dumping, followed by more infill,

then capped with plowed over material as the fields were cultivated in the 19th century.

The fact that these pits were still discernable on the surface as late as the end of the

nineteenth century provides for a long time for C14 datable material to be included.

What is of interest is that while we have two early dates on Grand River vessels, we have no dates later than A.D. 1400. These pits were utilized prior to A.D. 1400, but then were

not affected by humans until Europeans plowed the fields.

Taking the overall distribution of dates into account, it appears that the material

sampled has an effect on interpretations. In Figures 3.2, 3.3, and 3.4 the red line

represents A.D. 1300. The four wood charcoal dates (all from the 1968 excavations)

cluster around cal. A.D. 1245, with a 2δ pooled range of cal. A.D. 1215 to 1275 (Figure

3.2). The six maize/nut annual dates cluster around cal. A.D. 1350, with a 2δ pooled

range of cal. A.D. 1300 to 1400 (Figure 3.3). The 10 ceramic residue dates span a wide

range of dates, but their 2δ pooled range is cal. A.D. 1225 to 1270 (Figure 3.4). Only one

of the 20 dates (F04-22) is statistically earlier than the pooled 2δ average of cal. A.D.

1260 to 1275. Removing this date as an outlier, the resulting pooled 2δ average becomes

cal. A.D. 1265 to 1280 and the remaining 19 dates all overlap at a 99% probability

(Figure 3.5). It appears that early and late dates within the range may be a function of the

material type being analyzed, rather than a function of diachronic site usage. We expect

the wood charcoal dates to be earlier than the rest, and they are. The mean food residue

dates do average earlier than annual plant dates, but have a considerable range, and are

statistically the same. This discussion is not meant to conclude that the site was only occupied for 15 years, but rather that radiocarbon dates alone do not provide an analysis fine-grained enough to be able to separate out the different activity areas as being distinct

temporal entities.

Figure 3.2 Dates from Wood Charcoal (Calibrated with Stuiver and Reimer 1993).

Figure 3.3 Dates from Annuals (Calibrated by Stuiver and Reimer 1993).

Figure 3.4 Dates from food residue (calibrated byStuiver and Reimer 1993) .

Figure 3.5 Nineteen of the 20 Crescent Bay Radiocarbon Dates at 99% Probability (Calibrated by Stuiver and Reimer 1993).

Returning to our spatial distribution, the more detailed look at radiocarbon dates

shows that we cannot conclude with certainty that the two house types are the result of a

diachronic pattern of structure types. The dates do not support the third hypothesis, that

the two types of house structures represent a temporal shift in post marital residence

patterns. Further dating using a combination of food residues and annuals from each of

the areas may eventually allow us to further refine the chronology within the site.

Ceramic and Lithic Distribution

If there was a functional difference between the two different house types, as proposed by the first hypothesis, either due to the central being a communal ceremonial structure, as described in Skinner (1921) or by different domestic activities being carried out within the vicinity of each, then we would expect a difference in refuse being deposited into the associated pit features, in particular in ceramics and lithic debitage. Lithic debitage found in the context of feature deposits is thought to be the result of cleaning of nearby floor areas.

Features were compared by artifact densities using weights of artifact types per liter of soil floted, in order to control for variation in field methods, in addition to raw counts and weights, that enable the examination of all of the materials recovered by both screening through ¼ inch mesh and through flotation. Initially there appeared to be a correlation between volume of soil floted and density of artifacts (Figures 3.6; 3.7). By plotting artifact density versus volume of soil floted, it is apparent that there is a non- linear correlation, which tends to be skewed along both axes. Running a Spearman

Correlation test on both sets of data, shows that both exhibit some sort of correlation.

Null: rs=0.00 Alter: rs≠0.00 Ceramics: Reason to doubt the null hypothesis (rs=0.360; n=80; p=0.001). Lithic debitage: Reason to doubt the null hypothesis (rs=0.306; n=80; p=0.006).

However, the features at zero density along the x-axis are postmolds, and thus zero density is not unexpected. As for the distribution along the y-axis, this suggests that the sampling method is working, and that there are features that have few to no cultural remains.

Figure 3.6 Ceramic density plotted against Liters of soil processed.

Figure 3.7 Debitage density plotted against Liters of soil processed.

There is a problem with using feature densities in that the 2000 flotation material was not available for this analysis, and so 16 out of 78 pit features (20 percent) do not have flotation artifact data. It was then determined that raw weights per feature, combining artifacts collected by dry screening in the field and through floatation in the lab would be the best method of displaying distributions, keeping in mind that the 2000 flotation material and the 2008 screened material are not included. The results of both methods are given in Figures 3.8 and 3.9.

Figure 3.8 Debitage density of feature flotation.

Figure 3.9 Total debitage recovered from features.

Remembering that the results from House 1 and the southeast portion of the site are not comparable to later data because of the excavation and recovery strategies used in 1968, the analysis for the remainder of this paper focuses on Longhouse 1 and House 2 with their associated features.

Features F04-14 and F04-15 near the western house, as well as several features on the eastern portion of the site, contain a large amount of lithic debitage. Within

Longhouse 1, only one feature contains more than 75 grams of lithic debitage (F02-23 contains 108 grams). F04-22 northeast of Longhouse 1 does contain 340g of lithic debitage. From this distribution, it appears that less flint knapping occurred within

Longhouse 1 than at other portions of the site. By selecting all of the features within

Longhouse 1, including those closely associated on the southern end we get a total of 562 grams of lithic debitage coming from Longhouse 1, versus 497grams for all of the features surrounding the western house (Figure 3.36). When the palimpsest of

Longhouse 1 is taken into account, and the fact that only approximately half of

Longhouse 1 area has been excavated, it is impossible to state that more flint knapping occurred in one area versus the other, and it appears that there is a relatively even distribution of debitage between the two areas.

The distribution of chipped stone tools, or portions of chipped stone tools (broken tools were not tallied separate from whole tools, when noted on the tabulation forms, and so the GIS is unable to distinguish between the two), suggests a correlation between features that contain high rates of debitage and high rates of chipped tools, with tools largely missing from within the boundaries of Longhouse 1 (Figure 3.10).

Figure 3.10 Chipped stone tools recovered from features.

Ceramic distribution shows a similar pattern to that of the lithic debitage distribution (Figure 3.11, 3.12).

Figure 3.11 Density of ceramics from feature flotation.

Figure 3.12 Ceramics recovered from features by weight.

Selecting the features within Longhouse 1 and those on the outside southern end,

a total of 8472 grams of ceramics were recovered from within these features, as opposed

to 22,322 grams recovered from the features surrounding House 2 (Figure 3.36). For

comparative purposes, material per square meter of excavated features was used as a way

to compare the amount of artifacts between the two areas, while accounting for differential excavations (Table 3.3). The ceramics were over four times as dense around

House 2, compared to Longhouse 1. The lithic debitage is less than two times as dense.

While the majority of the ceramics are plain, shell tempered sherds, those that have been given a type designation can be mapped to see if any patterns occur (Schneider and

Clauter 2007; Schneider, et al. 2006). Because most of these sherds do not occur in high numbers, only their locations have been plotted, and not number or weights. Sherds obtained from the plowzone in 1998 for the southeast portion of the site were included, to give an idea of the types of sherds that were recovered from that general area. Late

Woodland sherds were only recovered from a single feature (Figure 3.13). All other sherds types were recovered from both Longhouse 1 and House 1 and 2 areas (Figures

3.14 through 3.20). An exception is Fisher ceramics (Griffin 1948), which were absent from Longhouse 1 (Figure 3.16). Fisher ceramics are found in northern Illinois and northwest Indiana (Berres 2001; Brown 1961; Griffin 1948; Jeske 2003) and are generally considered to date to the 13th century or earlier. Allamakee Trailed ceramics

(Figure 3.14) are considered post 14th century ceramics, associated with Orr Phase sites

of Iowa, Huber Phase sites in Illinois and classic Oneota horizon sites Illinois and

western Wisconsin (Brown 1990; Henning 1995; O'Gorman 1993). Grand River (Figure

19) and Carcajou wares (Figure 15), which seem to have rather long temporal spans, are

found throughout the site. It is also interesting to note that the Crescent Bay Punctate

(Schneider 2008:Figure 17) variety is largely from the south and east portions of the site, while Busseyville Paddle Groove (Schneider 2008:Figure 18) is largely from the north and east portions of the site. Whether these patterns are the result of shifts in time or are

88 the result of sampling error still needs more dating of ceramics and excavation at the site.

Figure 3.13 Distribution of Late Woodland ceramics.

Figure 3.14 Distribution of Allamakee.

Figure 3.15 Distribution of Carcajou Ceramics.

Figure 3.16 Distribution of Fisher ceramics.

Figure 3.17 Distribution of Crescent Bay ceramics.

Figure 3.18 Distribution of Busseyville ceramics.

Figure 3.19 Distribution of Grand River ceramics.

Figure 3.20 Distribution of Unclassified Oneota ceramics.

One further observation is the distribution of materials in the features surrounding

House 2. The features just to the south of the house have high concentrations of lithic and ceramic materials, especially F04-14 and F04-15, both designated as cylindrical features, while the two cylindrical features just to the west of the house have relatively few artifacts. It is proposed that the entrance to the house faced to the southeast (Figure

3.21). This wall appears to have lacked interior benches like the other three, and the location of two features dense with lithic and ceramic refuse leads us to reason that they were contemporaneous with the occupation of the structure, and served as convenient refuse pits, once their primary function had been fulfilled.

Figure 3.21 House 2 with associated features.

Floral and Faunal Distribution

The examination of the distribution of floral and faunal materials is hampered by the lack of finalized research on these materials from Crescent Bay. While preliminary raw counts and weights are recorded for features, analysis has been conducted for relatively few features at this point (Hudson n.d.; Hunter 2002). Nonetheless, distribution of faunal material appears to be very similar to the distribution of the ceramics (Figure 3.22, 3.23), while the distribution of fish scales (Figure 3.24, 25) and shell (Figure 3.26, 27) are more restricted and largely absent from the central longhouse area. There are several possible activities that could account for the distribution of shell: their use as a food item (mussel) and as an aplastic for making ceramics, or for manufacturing other objects, such as ornaments or fish lures.

Figure 3.22 Density of animal bone recovered from feature flotation.

100

Figure 3.23 Distribution of animal bone recovered from features.

101

Figure 3.24 Density of fish scales recovered from feature flotation.

102

Figure 3.25 Distribution of fish scales.

103

Figure 3.26 Density of shell recovered from feature flotation.

104

Figure 3.27 Distribution of shell.

105

An examination of the distribution of floral materials is skewed by the fact that material from only 36 features has been examined by a paleoethnobotanist. Of these 36,

29 are from the 2000 excavations, though an ongoing analysis is currently underway.

The presence of maize cupules indicates that maize was present on the cob at the site, which in turn indicates that maize was grown nearby, as is expected for Oneota sites.

The distribution of cupules and kernels is very similar (Figures 3.28, 3.29, 3.30, 3.31).

106

Figure 3.28 Density of Zea mays cupules and glumes recovered in feature flotation.

107

Figure 3.29 Distribution of Zea mays cupules.

108

Figure 3.30 Density of Zea mays kernels recovered in feature flotation.

109

Figure 3.31 Distribution of Zea mays kernels.

110

In the past, the cylindrical pits have been described as probable wild-rice

threshing pits, based on Jenks (Jenks 1901) and Stout and Skavlem (1908) (Hunter, et al.

2003; Mollet and Jeske 2001). Keeping in mind the discussion about ethnographic

analogy and archaeological features, we can take a closer look at these pits and their inferred function.

There is considerable variation in size among the examples of wild-rice threshing pits cited by historic sources. Stout and Skavlem observed around 40 pits in 1906, on property neighboring the Crescent Bay site. They ranged from two to four feet in diameter and one to two feet deep. The authors believed that the pits matched Jenks’

(1901) description (Stout and Skavlem 1908:80). Jenks does not give any dimensions of the Lake Koshkonong threshing pits, but most of the pits he does describe are approximately two feet across (Jenks 1901:1066-70). The average diameter of the 29 features designated as cylindrical at Crescent Bay is 1.35 meters (4.43 ft). In fact there is only one of these features that is smaller than two feet in diameter. There are only three other cylindrical features that have a diameter less than 0.95 meters (3.13 ft).

Jenks’ ethnographic descriptions of wild-rice threshing pits describe them as being about two feet across and about one and a half feet deep (Jenks 1901:1066-70).

According to Jenks, the pit was lined with a skin, upon which the rice was placed, and then trod upon by an adult male in order to separate the hull from the grain. However,

O’Gorman (1996:183) has argued that this would have been traditionally done by women, along with every other phase of wild-rice gathering and production. Several variations on this general method are described, including putting the rice into a skin bag, lining the pit with wood, using a stick or two, which are driven into the ground for

111

support, or using staves to beat the rice, rather than stomping on it. Not all methods

included digging a pit. Some laid blankets on the ground where they beat the rice, or

others would roast the grain, which caused the hulls to crack, and then they would be rubbed off (Jenks 1901:1066-70).

Because of the range of variation in historic groups, we cannot assume that people

on Lake Koshkonong in the thirteenth and fourteenth centuries were processing wild-rice in exactly the same was as any one of these examples. We cannot base our decision on ethnographic evidence alone and must look to archaeological data themselves (Binford

1967; Kuznar and Jeske 2006).

The proposed threshing pits at Crescent Bay are larger than most ethnographic

examples. However, other characteristics can be used to further examine this issue. The

morphology of the pits—straight sides with a flat floor—supports an interpretation of wild-rice threshing pits, as this shape provides a flat surface on which to stomp the grain.

In addition, many of the cylinder pits contained large stones in the bottom, which R.

Jeske (2001) interprets as possible chinking for a post that may have been used to steady the thrasher. Jenks describes several methods of posts or stakes being used in this

manner (Jenks 1901:1067). Unfortunately, no internal central postmolds have been identified by the excavators yet.

Pit contents are another line of evidence (Figure 3.32, 3.33). Eight out of the nine cylinder features examined for floral material yielded wild-rice, though four of these eight yielded less than ten grains (Table 3.2). The other four yielded over 100 grains each. Only charred or parched seeds of wild rice are preserved and analyzed (Olsen

112

2003:92). Moffat and Arzigian (2000) argue that burning is not associated with

threshing, therefore carbonized wild-rice would not necessarily be found in threshing

pits. However, modern and ethnographic evidence indicates that wild-rice is dried or

charred to loosen the chaff for threshing (Jenks 1901:1064). Even still, if the pits were

lined or the rice was in bags (Jenks 1901:1067; Vennum 1987:123-125) we might expect

to find few grains of wild rice in the pits themselves. Conversely, if wild-rice refuse is

dumped with other food remains it would end up in former storage pits used as dumps.

Like pit size or shape, the presence or absence of wild-rice alone is not going to be a reliable indicator of the pits’ original function.

However, taken as a whole—pit size, shape, the possible chinking stones, the ubiquity of wild rice in pit contents—it is likely that these features functioned as wild rice threshing pits. The convergence of archaeological data, rather than simple direct analogy, allows the inference to be drawn (Binford 1972; Kuznar and Jeske 2006). Like the interpetation of Longhouse 1 as a longhouse, the cylindrical pits do not precisely match historic analogs of threshing pits. Rather the interpretations are archaeological explanations of Oneota features and structures.

113

Table 3.2 Features containing wild-rice. Feature Feature Type Wild-rice (ct) F02-01 basin 4380 F02-28 cylinder 250 F00-15 cylinder 180 F04-15 cylinder 160 F00-11 cylinder 145 F00-17 basin 29 F00-01 basin 20 F00-05 post 11 F00-03 post 10 F00-21 cylinder 9 F00-13 basin 5 F00-02 basin 3 F00-09 post 3 F08-07 cylinder 3 F00-18 post 2 F06-173 cylinder 2 F00-26 cylinder 1 F00-06 basin 1

114

Figure 3.32 Density of Zizania aquatic recovered in feature flotation.

115

Figure 3.33 Distribution of Zizania aquatic.

116

Small amounts of Chenopodium berlandieri have been recovered from the site

(Figure 3.34, 3.5). The majority of the Chenopodium at the site was appears to be wild, while a small amount was cultivated. The Chenopodium possibly cultivated at the site, but more likely brought in from outside the local area (2003:167). In addition, small amounts of other eastern agricultural complex plants, including squash (Cucurbita pepo) and little barley (Hordeum pusillum), have been recovered, as well as wild purslane

(Portulaca oleracea), but each has only been recovered from one feature, and so their

distribution has not been mapped.

117

Figure 3.34 Density of Chenopodium berlandieri recovered in feature flotation.

118

Figure 3.35 Distribution of Chenopodium berlandieri.

119

From examining the distribution maps above, there does not appear to be a

significant difference in the distribution of the floral material between the two structures;

however, when the totals and densities are examined, a pattern emerges in both the floral

and faunal materials (Table 3.2). While it cannot be stated whether the differences in the

following discussion are statistically significant at this point, there does appear to be a

difference in the material remains that are associated with each structure. Thus far, the

two areas have been compared on a one to one basis. However, there is difference in the

number and volume of features that are being compared (Figure 3.36). The combined

area of the features (postmolds and wall trenches were not included, as they contain little material) was first compared, in order to utilized both floted and screened materials. The

GIS can quickly calculate the area of features selected. Comparing the surface areas of feature profiles, Longhouse 1 contains 20.4 square meters of features, or 14 percent of the area excavated in that vicinity (the wall trenches and unidentified features were removed from these calculations), while the House 2 area contains 11.1 square meters of features,

or 9 percent of the area excavated in that vicinity (Table 3.3). An additional

consideration is that the amount of area excavated in the vicinity of both structures affects

the amount of material recovered to interpret. To compensate for this in order to achieve

a more accurate unit for comparison, all material recovered during flotation analysis were

calculated as material per cubic meter (regardless of feature type), which gives a more

accurate indicator of intensity of use (Table 3.4). When calculating liters of soil floted,

the amount varied between material types, because fewer features were analyzed during

the floral analysis. With these factors in mind, several observations are made. It appears

that maize, wild-rice, and Chenopodium were processed and/or consumed in both areas,

120 with a little more material coming from Longhouse 1 (except for maize kernels). Bone and fish scales are at least eight times as dense, and shell is 69 times as dense in the vicinity of House 2. This suggests that more food was being cooked within the near or within House 2, than within Longhouse 1, as these materials are more likely to be cooked and eaten, rather than stored (Arzigian, et al. 1989:255).

Another line of evidence is ceramic production. If shell that is present at the site was part of the ceramic making process, we would expect a large amount of shell fragments. If the shell is refuse from only mussel consumption, then we would expect shells in larger pieces, virtually whole and in a position that would indicate that the hinge ligament was still holding the two halves together when discarded (Arzigian, et al.

1989:265). Although faunal analysis is not complete, a preliminary examination of the

Crescent Bay shell from a variety of contexts indicates that the majority appears to be calcined. Feature F04-14 contained both whole, or large shell pieces, as well as a quantity of pulverized pieces. F04-15 contains chunks of the translucent interior, as well as a few whole shells. F04-40, the feature within Longhouse 1 that contains the most shell, appears to be the calcined remains of a single bivalve. A detailed analysis needs to be conducted in order to further explore this idea, but it appears that bivalves were processed for food in the vicinity of House 2, and then possibly used as temper for ceramics.

121

Figure 3.36 Comparison of feature blocks at Crescent Bay Hunt Club.

122

Table 3.3 Distribution of material by area. Category Longhouse Per/m2 House 2 Per/m2 delta direction bone (g) 832 40.8 3440 309.9 7.6 + shell (g) 50 2.5 1200 108.1 44.1 + fish (ct) 931 45.6 3192 287.6 6.3 + maize cupules (ct) 1547 75.8 402 36.2 0.5 - maize kernel (ct) 775 38.0 279 25.1 0.7 - Chenopodium sp. (ct) 477 23.4 53 4.8 0.2 - wild rice (ct) 652 32.0 160 14.4 0.5 - lithic debitage (g) 562 27.5 497 44.8 1.6 + ceramics (g) 8472 415.3 22322 2011.0 4.8 + Total Area (m2) 20.4 11.1

Table 3.4 Distribution of material by volume. Category Longhouse 1 Per/m3 House 2 Per/m3 delta direction bone (g) 620.9 0.139 2609.6 1.803 13 + shell (g) 7.9 0.002 199.87 0.138 69 + fish (ct) 849 0.19 2249 1.554 8.2 + maize cupules (ct) 1547 0.741 402 0.697 0.9 - maize kernel (ct) 775 0.371 279 0.483 1.3 + Chenopodium sp. (ct) 477 0.228 53 0.092 0.4 - wild rice (ct) 652 0.312 160 0.277 0.9 - lithic debitage (g) 271.2 0.061 298.1 0.206 3.4 + ceramics (g) 619.2 0.138 4918.9 3.4 24.6 + Total Volume (m3) floted for non-floral analysis 4474.5 1447.3

Total Volume (m3) floted for floral analysis 2088.5 577.0

Finally, the distribution of human remains differs between the two areas, with

Longhouse 1 having the only intact burials that have been located at the site (Figure

123

3.37). There are two burials within the boundaries of Longhouse 1. B00-01 was a single adult male placed within a cylindrical pit (R. Jeske 2000). B02-01 is the remains of three individuals, an adult male (B02-01) holding a child (B02-02) and an infant (B02-03)

(Foley Winkler 2004). B04-01 is an isolated adult male mandible immediately east of, if not within, the structure (Foley Winkler 2009). B98-01 was burial of a probable adult male found in a shovel probe north and east of the house (likely not associated with the structure itself) (Gaff 1998b). B08-01 is a right maxillary deciduous incisor of a six year old child, found just east of the structure (Foley Winkler 2009). B04-02 is the distal end of the right humerus of an adult male that was found during flotation of Feature 04-15, located south of House 2 (Foley Winkler 2009). It is concluded that Longhouse 1 was associated with mortuary ritual. However, beyond containing intact burials beneath the floor, the nature of ritual is unclear.

124

Figure 3.37 Locations of human remains found at Crescent Bay.

125

Chapter 4

Discussion and Conclusion

Three hypotheses were proposed to account for the presence of two distinctly different structures at the Crescent Bay Hunt Club site:

1) The pattern is the result of a seasonal shift in residence patterns, where the smaller

structures (Houses 1 and 2) are single family winter dwellings, and the larger

structure (Longhouse 1) is an extended family summer dwelling.

2) The difference is functional, with Longhouse 1 representing a communal structure,

and Houses 1 and 2, residential ones.

3) The two patterns represent a diachronic shift in post-marital residence patterns,

where the smaller houses represent an early, patrilineal post-marital residence

pattern and the larger longhouse represents a later, matrilineal post-marital

residence pattern.

It must first be stated that these three hypotheses are not mutually exclusive. In fact, it is likely that changes in season or time would bring about changes in function. It is possible that all three were simultaneously responsible for forming the archaeological remains seen at the site. The purpose of this thesis is to examine if any of the three

126 hypotheses have any merit as a provisional explanation. In my conclusions, I shall also propose what further evidence would be required to strengthen each hypothesis.

The first hypothesis is not currently supported on the basis that the seasonal indicators present (e.g. maize, wild-rice and nuts, all indicative of a fall occupation) are fairly evenly distributed between the two structure activity areas. The ability to store these foods complicates any interpretation along these lines, but allows us to infer the possibility of winter occupation as well. It is reasonable to assume that at least some of the population was present during the summer, to plant and tend to the maize fields. At this point, there is no reason to conclude that the structures were differentially occupied seasonally, however, it is still a possibility. Evidence for seasonality include Fortier’s analysis of 500 fish scales recovered from the 1968 excavations which concluded that most of the represented fish were captured during the late summer or fall, though a very few did indicate a spring growth (Fortier 1972:20-21). One potential complication is that

95 percent of his sample came from only two features. The other features were largely devoid of scales, which might suggest a multi-seasonal occupation. Continuing with this line of reasoning, the absence of shell from Longhouse 1 may also argue for a winter occupation, or rather, a not summer occupation, when the shell beds would have been most heavily harvested. Further floral and faunal analysis needs to be conducted to help resolve this issue. One possible interpretation based upon available evidence, is that

Longhouse 1 was a winter dwelling, which contained storage pits that could be easily accessed in the winter. Use as storage pits would also account for the greater density of floral material from this area. The Crescent Bay pattern is not directly supported by ethnographic evidence, where smaller structures are occupied in the winter and larger

127 structures are occupied in the summer (Skinner 1921; McKusick 1973). However, it is still interesting to note that, ethnographically, there was a seasonal difference in structures in the Great Lakes area. Seasonality cannot yet be ruled as a possibility at

Crescent Bay.

The third hypothesis does not seem to be supported based on the 20 radiocarbon dates from the site. The single exception is a very early date from Grand River vessel recovered from a feature near Longhouse 1. However, that aberrant date is the wrong direction to validate the hypothesis, and Longhouse 1is associated with six other dates within its vicinity placing the structure’s use prior to the end of the thirteenth century.

The current radiocarbon chronology does not support diachronic difference of occupation between structures. However, multiple rebuilding suggests some longevity, and wall trench F00-42 clearly cuts through basin F00-17, indicating that Longhouse 1 was built after the site had been occupied for some duration. More dating and a continued examination of feature cross-cutting may give more complete confidence in the future.

The second hypothesis, that the structures served different functions, does have some support. The two different structure types present at the Crescent Bay Hunt Club present two distinct patterns of associated features that--along with the distribution of artifacts within these features--suggest a functional difference. The first significant (and obvious) difference is size (Table 4.1). The floor area of House 2, inside of the benches is eight square meters; the entire interior space is sixteen square meters. The benches and exterior walls are less well defined for House 1, and its interior space is estimated to be approximately thirteen square meters. The interior area of Longhouse 1 is 178 square meters, over eleven times the area of the smaller structures. Size is directly related to the

128

number of individuals that a structure is designed to accommodate, thus Longhouse 1 can

be assumed to have been built to accommodate a substantial number of individuals

compared to Houses 1 & 2.

Table 4.1 House floor areas. Structure Area (m2) House 1 13 House 2 16 Longhouse 1 178

The presence of benches within the Houses 1 and 2 is the second structural

difference between the two. If these are indeed sleeping benches, then we have evidence

for sleeping in one structure type, but not in the other. A third difference is the presence

of pit features associated with food processing and storage within Longhouse 1, but on

the exterior of Houses 1 and 2. It appears that from pit morphology and material

contents, that the pits were used for similar functions, and that similar activities took

place in their vicinity. Though the types of materials are equally represented in both areas, there is a difference in the amounts of these materials, indicating that the frequency of some activities differed between the two structures.

The final difference between the two house types is the presence of burials within

Longhouse 1, and the isolated bone recovered from nearby features. This may be the result of a small sample size, though we can include the House 1 in this discussion, because if there had been an intact burial beneath the house floor, it would have been discovered. The presence of burials beneath Longhouse 1 requires that it be interpreted

129 as being associated with mortuary ritual, though the presence of domestic activities strongly suggests that this was not the only function of the structure (see Hollinger 1995).

Sleeping and cooking of meat, fish and mussels, and if the interpretation of the shell material holds up to further examination, the production of pottery, occurred in the vicinity of House 2, which are all activities representative of a single family structure, or household economies. Because of their design, these structures would be well suited to a fall/winter occupation. They are small and easily heated, with benches for sleeping, up off of the ground, in warmer air. It appears that the processing (and possibly storage) of plant materials in pits occurred both in the open air and within the boundaries of

Longhouse 1. At this point, there is no reason to deduce that these features were functionally different whether they were located inside or out. Longhouse 1 may then be interpreted to be a communal structure, where larger group activities such as mortuary rituals, harvest rituals, or others may have taken place, in addition to regular domestic activities, such as food processing and storage. Though the radiocarbon evidence at this time does not support Hollinger’s hypothesis of a shift in residence patterns through time, the Crescent Bay longhouse may be an example of a very early longhouse built when the shift in marriage residence patterns was first happening.

GIS as a Tool for Managing Archaeological Data

It has been demonstrated that the GIS is a useful tool for managing spatial data, providing a relational database that links artifact attributes to a location in space. However, as with any database, it is only as useful as the data that are collected and entered. If a GIS is to

130

be used in archaeological investigations, then the data collected and entered must be

formalized and remain as consistent as possible. The Crescent Bay materials are being

studied by a number of different researchers for multiple purposes. At this time, a single

inventory form for all screened and floted materials is not available. The ceramics,

lithics, floral and faunal materials are all in different stages of analysis, and will continue

to be so for the near future. Excavations are on-going at the site, and will be partially

contingent upon analyses such as this thesis.

Most critically, data need to be recorded in a systematic and consistent manner.

As any researcher can attest, the longer data are kept unassociated, the more likely they

will get lost, misplaced or forgotten as individuals come and go. It is exactly these types

of problems that the GIS can help solve. The work conducted for this thesis has provided

a beginning to systematize the collections of seven field seasons.

Going forward, the data gathered and collated for this thesis will be placed onto

the computer server at the University of Wisconsin-Milwaukee Archaeological Research

Laboratory. Ideally, the next step would be to hyperlink the digital photos, and scanned

unit and feature forms to the GIS. Many of these are already on the server and it would

be just a matter of prioritization. More importantly, as investigations and analysis go

forward, care should be taken to be sure that all artifact analysis is entered into a

spreadsheet and placed onto the server, using the same feature and unit numbering conventions, so that this data can be graphically displayed with only a few keystrokes.

The excavations at Crescent Bay were undertaken before the digitization of

archaeological field data was accessible to field school level archaeology. However,

131

currently spatial and attribute data can be incorporated into the general inventory and

recording processes much more easily. A GIS can be a useful tool in managing the data,

but implementing this sort of data management in order to analyze artifact distributions

becomes a daunting task at best without standardization of collection and analyses

methods. If data are recorded with a GIS in mind, it then becomes a much easier and

more efficient task to continually incorporate new data into the existing database to be

able to reanalyze distributions and associations. With the advent of digital photography,

it becomes a fairly simple task to hyperlink photos (and scanned documents) with their features within the GIS, further facilitating an interactive analysis by linking all these separate elements (artifact attributes, spatial locations, and associated documentation) together into a single user interface. If GIS is to reach its potential usefulness in archaeology in general, it needs to be incorporated into our methods, methodology, and research design, before a single shovel pierces the ground.

132

References Cited

Arzigian, C. M., J. L. Theler, R. F. Boszhardt, M. Scott and R. Rodell 1989 Feature Fill Depostition and Seasonality at the Pammel Creek Site. The Wisconsin Archeologist 70(1-2):251-272.

Baerreis, D. A. and R. A. Bryson 1965 Climatic Episodes and the Dating of Mississippian Cultures. The Wisconsin Archeologist 47(3):101-131.

Bender, M. M., R. A. Bryson and D. A. Baerries 1970 University of Wisconsin Radiocarbon Dates VII. Radiocarbon 12:335- 345.

Berres, T. E. 2001 Power and Gender in Oneota Culture: A Study of a Late Prehistoric People. Northern Illinois University Press, DeKalb.

Binford, L. R. 1967 Smudge Pits and Hide Smoking: The Use of Analogy in Archaeological Reasoning. American Antiquity 32(1):1-12.

Birmingham, R. A. 2004 Preliminary Report on the Busseyville Component at the Carcajou Point Site: Excavations on the Yates Propery. In A Conundrum at Carcajou Point: 2004 Property Tax Exemption Program Investigations at the Carcajou Point Site (JE- 0002) Lots W8651, W8625, W8537, edited by A. L. Rosebrough and J. H. Broihahan, pp. 59-75. Office of the State Archaeologist Technical Report Series No. 04-0002. Office of the State Archaeologist Historic Preservation Division- Public History Wisconsin Historical Society, Madison.

Boszhardt, R. F. 1994 Oneota Group Continuity at La Crosse: The Brice Prarie, Pammel Creek and Valley View Phases. The Wisconsin Archeologist 75(3-4):173-236.

2004 Blind Dates and Blind Faith: The Timeless Story of the "Emergent" Oneota McKern Phase. The Wisconsin Archeologist 85(1):3-30.

Boszhardt, R. F., T. W. Bailey and J. P. Gallagher 1985 Oneota Ridged Fields at the Sand Lake Site (47-LC-44), La Crosse County, WIsconsin. The Wisconsin Archeologist 66(1):47-67.

Brandt, R., B. J. Groenewoudt and K. L. Kvamme 1992 An Experiment in Archaeological Site Location: Modeling in the Netherlands using GIS Techniques. World Archaeology 24(2):268-282.

133

Brown, J. A. 1961 The Zimmerman Site: A Report on Excavations at the Grand Village of Kaskaskia, La Salle County, Illinois. Report of Investigations No. 9. Illinois State Museum, Springfield, Illinois.

1990 The Oak Forest Site: Investigations into Oneota Subsistence-Settlement in the Cal-Sag Area of Cook County, Illinois. In At the Edge of Prehistory: Huber Phase Archaeology in the Chicago Area, edited by J. A. Brown and P. J. O'Brien, pp. 123-308. Center for American Archaeology, Kampsville.

Brubaker, R. and L. Goldstein 1991 Jefferson County: Archaeological Investigations at Carcajou Point in the Lake Koshkonong Region. In The Southeastern Wisconsin Archaeology Program: 1990-1991, edited by L. Goldstein, pp. 35-83. vol. 107. UW-Milwaukee, Milwaukee.

Cassidy, F. G. 1948 "Koshkonong," a Misunderstood Place-Name. In The Wisconsin Magazine of History, pp. 429-440. vol. 31. Wisconsin Historical Society, Madison, WI.

Cowell, S. H., E. Schuetz and S. A. Schneider 2008 The Oneota Component at the Twin Knolls Site (47JE379). Paper presented at the 54th Annual Midwest Archaeological Conference, Milwaukee, Wisconsin.

Dillemuth, J. A. 1999 A GIS Analysis of the Lake Koshkonong Effigy Mounds. In Southewast Wisconsin Archaeology Program 1998-1999, edited by R. J. Jeske, pp. 95-143. Report of Investigations 141. Archaeological Research Laboratory, University of Wisconsin-Miwlaukee, Milwaukee.

Dobbs, C. A. 1982 Oneota Origins and Development: The Radiocarbon Evidence. In Oneota Studies, edited by G. Gibbon, pp. 91-106. Publications in Anthropology No. 1. University of Minnesota, Minneapolis.

Dolan, S. and S. R. Kuehn 2006 An Oneota Longhouse Structure from East-Central Wisconsin. Paper presented at the Midwest Archaeological Conference, Urbana, Illinois.

Edwards, R. 2010 Oneota Settlement Patterns Around Lake Koshkonong in Southeast Wisconsin: An Environmental Catchment Ansalysis using GIS modeling. Unpublished Master's Thesis, Department of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

134

Emerson, T. E. 1999 The Apple River Mississippian Culture of Northwestern Illinois. In Cahokia and the Hinterlands: Middle Mississippian Cultures of the Midwest, edited by T. E. Emerson and B. Lewis, pp. 164-182. University of Illinois Press, Chicago.

Foley-Winkler, K. M. 2004 Oneota mortuary practices in southeast Wisconsin. Unpublished Master's Thesis, Depratment of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

Foley Winkler, K. M. 2004 Oneota mortuary practices in southeast Wisconsin. Unpublished Master's Thesis, Depratment of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

2009 2008 Investigations at the Schmeling Site. Paper presented at the 74th Annual Society for American Archaeologists, Atlanta, Georgia.

Fortier, A. C. 1972 Fish Bone and Scale Analysis from the Crescent Bay Hunt Club Site (JE- 244), Jefferson County, Wisconsin. Unpublished Master's Thesis, Department of Anthropology, University of Wisconsin, Madison.

Frautschi, W. A. 1945 Early Wisconsin Shooting Clubs. In The Wisconsin Magazine of History, pp. 391-415. vol. 28. Wisconsin Historical Society, Madison, WI.

Gaff, D. G. 1998a Investigations at Carcajou Point, Jefferson County. In Southeastern Wisconsin Archaeology Program: 1997-1998, edited by R. J. Jeske. Report of Investigations No. 134. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

1998b Preliminary Investigations at the Crescent Bay Hunt Club, Jefferson County. In The Southeastern Wisconsin Archaeolgoy Program: 1997-1998, edited by R. J. Jeske, pp. 34-75. Report of Investigations No. 130. Uninversity of Wisconsin-Milwaukee, Milwaukee.

Gallagher, J. P., R. F. Boszhardt, R. F. Sasso and K. Stevenson 1985 oneota Ridged Field Agriculture in Southwestern Wisconsin. American Antiquity 50(3):605-612.

135

Gibbon, G. 1968 Je-244: The Crescent Bay Hunt Club Site. Notes on file, University of Wisconsin-Milwaukee, Milwaukee.

1969 The Walker-Hooper and Bornick Sites: The Grand River Phase Oneota Sites in Central Wisconsin. Unpublished Ph. D. Thesis, Department of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

1972 The Walker-Hooper Site in Marquette County. The Wisconsin Archeologist 53(4):149-290.

1986 The Mississippian Tradition: Oneota Culture. The Wisconsin Archeologist 67(3-4):314-338.

Goldstein, L. 1990 Work at the Carcajou Point Site (47-JE-2). In The Southeastern Wisconsin Archaeology Program: 1989-1990, edited by L. Goldstein and E. Benchley, pp. 39-42. Report of Investigations No. 103. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

Goldstein, L. and J. D. Richards 1999 Ancient Aztalan: The Cultural and Ecological Context of a Late Prehistoric Site in the Midwest. In Cahokia and the Hinterlands: Middle Mississippian Culture of the Midwest, edited by T. E. Emerson and B. Lewis, pp. 193-206. University of Illinois Press, Chicago.

Griffin, J. W. 1948 Upper Mississippi at the Fisher Site. American Antiquity 14(2):124-126.

H.R._Page_&_Co. 1887 Illustrated Atlas of Jefferson County Wisconsin, containing Maps of Every Township in the County, with Village and City Plats. H. R. Page & Co., Chicago.

Hall, R. L. 1962 The Archeology of Carcajou Point 2 vols. University of Wisconsin Press, Madison.

1993 Red Banks, Oneota, and the Winnebago: Views from a Distant Rock. The Wisconsin Archeologist 74(1-4):10-79.

2004 Comment on "Blind Dates and Blind Faith: The Timeless Story of the 'Emergent' Oneota McKern Phase". The Wisconsin Archeologist 85(1):31-35.

Hanson, P. 1996 Jefferson County: Survey of the Crescent Bay Hunt Club Property. In Southeastern Wisconsin Archaeology Representing Region 9: Southeastern

136

Wisconsin Project Report: Wisconsin Regional Archaeology Program June 1995 - June 1996, edited by L. Goldstein, pp. 45-52. Report of Investigations No. 128. University of Wisconsin-Milwaukee, Milwaukee.

Henning, D. R. 1995 Oneota Evolution and Interactions: A Perspective from the Weaver Terrace, Southwest Iowa. In Oneota Archaeology: Past, Present, and Future, edited by W. Green. Report 20. Office of the State Archaeologist, Iowa City.

Heywood, I., S. Cornelius and S. Carver 2006 An introduction to geographical information systems. 3rd ed. New York, Harlow, England.

Highsmith, H. 1997 The mounds of Koshkonong and Rock River : a history of ancient Indian earthworks in Wisconsin. Fort Atkinson Historical Society [and] Highsmith Press, Fort Atkinson, Wis.

Hollinger, R. E. 1995 Residence Patterns and Oneota Cultural Dynamics. In Oneota Archaeology: Past, Present, and Future, edited by W. Green Report 20. Office of the State Archaeologist, The University of Iowa, Iowa City.

Hudson, J. n.d. Title., in press.

Hunt, E. D. 1992 Upgrading Site-Catchment Analyses with the Use of GIS: Investigating the Settlement Patterns of Horticulturalists. World Archaeology 24(2):283-309.

Hunter, C. L. 2002 A Comparative Study of Oneota and Langford Traditions. Unpublished Master's Thesis Department of Anthropology, University of Wisconsin- Milwaukee, Milwaukee.

Hunter, C. L., A. Lampien and C. Zervic 2003 Feature Analysis. In Lake Koshkonong 2002/2003: Archaeological Investigations at Three Sites in Jefferson County, Wisconsin, edited by R. J. Jeske. Report of Investigations No. 153. University of Wisconsin Archaeological Research Laboratory, Milwaukee.

Jenks, A. E. 1901 The wild rice gatherers of the upper lakes; a study in American primitive economics. Govt. print. off., Washington,.

137

Jeske, J. A. 1927 The Grand River Mound Group and Camp Site. Bulletin of the Public Museum of the City of Milwaukee. Vol. 3, No. 2, Milwaukee.

Jeske, R. J. 2000 Continuing Investigations at Crescent Bay Hunt Club Site (47-JE-904), Jefferson County. In The Southeastern Wisconsin Report of Investigations, edited by R. J. Jeske. Report of Investigations No. 145. University of Wisconsin- Milwaukee Archaeological Research Laboratory, Milwaukee.

2001 Program in Midwestern Archaeology (Southeastern Wisconsin Archaeology Program): 2000-2001. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

2003 Langford and Fisher Ceramic Traditions: Moiety, Ethnicity or Power Relationships in the Upper Midwest? The Wisconsin Archeologist 84(1):165-180.

2008 Space and Time at the Crescent Bay Hunt Club. Paper presented at the 52nd Annual Meeting of the Midwestern Archaeological Conference, Milwaukee.

Jeske, R. J., K. M. Foley-Winkler, T. Dahlen and L. C. Lambert 2003 Continuing Investigations at the Crescent Bay Hunt Club Site (47-JE-904), Jefferson County. In Lake Koshkonong 2002/2003: Archaeological Investigations at Three Sites in Jeffereson County, WIsconsin, edited by R. J. Jeske. Report of Investigations. No. 153 vols. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

Jeske, R. J., C. L. Hunter, D. M. Winkler, D. L. Miller and L. Plenener 2003 Preliminary Investigations at Carcajou Point (47-JE-02). In Lake Koshkonong 2002/2003: Archaeological Investigations at Three Sites in Jefferson County, Wisconsin, edited by R. J. Jeske. Report of Investigations No. 153. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

Jeske, R. J. and L. A. Kuznar 2001 Canine Digging Behavior and Archaeological Implications. Journal of Field Archaeology 28(3/4):383-394.

Jeske, R. J. and D. M. Winkler 2008 The Clovis Occupation of the Schmeling Site (47JE833) in Jefferson COunty, Wisconsin. Current Research in the Pleistocene 25:99-102.

Jones, E. E. 2006 Using Viewshed Analysis to Explore Settlement Choice: A Case Study of the Onondaga Iroquois. American Antiquity 71(3):523-538.

138

Keyes, C. R. 1927 Prehistoric Cultures. In The Palimpsest, pp. 215-229. vol. 8. Historical Society of Iowa, Iowa City.

Kroll, E. M. and G. L. Isaac 1984 Configurations of artifacts and bones at early Pleistocene sites in East Africa. In Intrasite Spatial Analysis in Archaeology, edited by H. J. Hietala, pp. 4- 31. Cambridge University Press, New York.

Kuznar, L. A. and R. J. Jeske 2006 Analogic Reasoning, Ethnoarchaeology, and the Impact of Canines on the Archaeological Record. Archeological Papers of the American Anthropological Association 16(1):37-46.

Lapham, I. A. 1855 The antiquities of Wisconsin, As surveyed and described. Antiquities of the New World,; v. 4; Variation: Smithsonian Institution.; Smithsonian contributions to knowledge,; v. 7, art. 4. Smithsonian Institution, Washington, D.C.

Loew, P. 2001 Indian nations of Wisconsin : histories of endurance and renewal. Wisconsin Historical Society Press, Madison.

Maschner, H. D. G. 1996 The Politics of Settlement Choice on the Northwest Coast: COgnition, GIS, and Coastal Landscapes. In Anthropology, space, and geographic information systems, edited by M. S. Aldenderfer and H. D. G. Maschner, pp. 175-189. Oxford University Press, New York.

Mason, R. J. 1966 Two Stratifies Sites on the Door Peninsula of Wisconsin. Anthropological Papers 26. Museum of Anthropology, University of Michigan, Ann Arbor.

1993 Oneota and Winnebago Ethnogenesis: An Overview. The Wisconsin Archeologist 74(1-4):400-421.

McElrath, D. L., J. A. Williams, T. O. Maker and M. C. Meinkoth 1987 No. 1 The Radic Site. University of Illinois Press, Urbana.

McGee, W. J. 1891 The Pleistocene history of northeastern Iowa. United States geological survey to the secretary of the interior 1889-'90.

139

McKusick, M. B. 1973 The Grant Oneota village. Office of State Archaeologist, report; 4; Variation: Report (Iowa. Office of State Archaeologist) ;; 4. University of Iowa, Iowa City.

Moffat, C. R. and C. M. Arzigian 2000 New Data on the Late Woodland Use of Wild Rice in Northern Wisconsin. Midcontinental Journal of Archaeology 25(1):49-81.

Mollet, J. and R. J. Jeske 2001 Feature Analysis of the Crescent Bay Hunt Club Site (47-JE-904). In Program in Midwestern Archaeology (Southeastern Wisconsin Archaeology Program): 2000-2001, edited by R. J. Jeske. Report of Investigations No. 148. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

Morse, J. 1822 A report to the Secretary of War of the United States, on Indian affairs, comprising a narrative of a tour performed in the summer of 1820. S. Converse, New Haven.

Moyes, H. 2002 The use of GIS in the Spatial Analysis of an Archaeological Cave Site. Journal of Cave and Karst Studies 64(1):9-16.

Munson, P. J. 1969 Comments on Binford's "Smudge Pits and Hide Smoking: The Use of Analogy in Archaeological Reasoning". American Antiquity 34(1):83-85.

Musil, J. L. and M. F. Kolb 1987 The Lake Koshkonong Region: Survey & Effigy Mound Research. In The Southeastern Wisconsin Archaeology Project: 1986-87 & Project Summary, edited by L. Goldstein, pp. 121-181. Report of Investigations No. 88. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

O'Gorman, J. 1993 The Tremaine Site Complex: Oneota Occupation in the La Crosse Locality Wisconsin. Volume 1: The Ot Site (47LC262). In Archaeology Research Series. State Historical Society of Wisconsin, Madison.

1995 The Tremaine Site (47-LC-95). Archaeology Research Series No. 3. Museum Archaeology Program State Historical Society of Wisconsin.

1996 Domestic Economics and Mortuary Practices: A Gendered View of Oneota Social Organization. Unpublished Ph. D. Thesis, Department of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

140

Olsen, M. L. 2003 Oneota subsistence in southern Wisconsin : agriculture and plant domestication at the Crescent Bay Hunt Club Site. Unpublished Master's Thesis, Department of Anthropology, UW-Milwaukee, Milwaukee.

Overstreet, D. F. 1976 The Grand River, Lake Koshkonong, Green Bay, and Lake Winnebago Phases: Eight Hundred Years of Oneota Prehistory in Eastern Wisconsin. Unpublished Ph. D. Thesis, Department of Anthropology, Univeersity of Wisconsin-Milwaukee, Milwaukee.

1985 Corn and Community Congregation-The View from Lake Winnebago. Paper presented at the 84th Annual Meeting of the American Anthropological Association, Washington, D.C.

1989 Survey and Testing, South Adams Street Redevelopment Project, Green Bay, Wisconsin. Report of Investigations No. 211. Great Lakes Archaeological Research Center, Inc., Milwaukee.

1993 McCauley, Astor, and Hanson - Candidates for the Provisional Dandy Phase. The Wisconsin Archeologist 74(1-4):120-196.

1997 Oneota Prehistory and History. The Wisconsin Archeologist 78(1-2):250- 297.

Pauketat, T. R. and T. E. Emerson 1997 Introduction: domination and ideology in the Mississippian world. In Cahokia : domination and ideology in the Mississippian world edited by T. R. E. T. E. Pauketat, pp. 1-29. University of Nebraska Press, Lincoln.

Peet, S. D. 1905 Prehistoric America 2 4 vols. Office of the American antiquarian, Chicago.

Radin, P. 1911 Some Aspects of Winnebago Archeology. American Anthropologist 13(4):517-538.

1970 The Winnebago tribe. University of Nebraska Press, Lincoln.

Ralph, E. K. 1971 Carbon-14 dating. In Dating techniques for the archaeologist, edited by H. N. Michael and E. K. Ralph, pp. 1-48. MIT Press, Cambridge, Massechusetts.

141

Rawling, J. E. I., R. J. Jeske and E. Kats 1999 Postmold Morphology, Petrology, and Genesis at the Crescent Bay Hunt Club Oneota Site, Wisconsin. Paper presented at the American Geological Society meetings, Denver.

Richards, J. D. 1992 Ceramics and Cultures at Aztalan, a Late Prehistoric Village in Southeast Wisconsin. Unpublished Ph. D. Thesis, Department of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

Richards, J. D., R. P. B. and B. D. Nicholls 1998 Archaeological Investigation in the Carcajou Point Locale, Sumner Township, Jefferson County, Wisconsin. Report of Investigations No. 129. University of Wisconsin-Milwaukee Historic Resource Management Services Archaeological Research Laboratory, Milwaukee.

Richards, J. D. and R. J. Jeske 2002 Location, Location, Location: The Temporal and Cultural Context of Late Prehistoric Settlement in Southeast Wisconsin. The Wisconsin Archeologist 83(2):32-54.

Rodell, R. 1984 A Survey of Oneota Sites in the Lake Koshkonong Area. In The Southeastern Wisconsin Archaeology Project: 1983-1984, edited by L. Goldstein, pp. 144-167. Report of Investigations No. 77. University of Wisconsin- Milwaukee Archaeological Research Laboratory, Milwaukee.

1987 Oneota Research. In The Southeastern Wisconsin Archaeology Project: 1986-87 & Project Summary, edited by L. Goldstein, pp. 190-199. Report of Investigations No. 88. University of Wisconsin-Milwaukee Archaeological Research Laboratory, Milwaukee.

1997 The Diamond Bluff Site COmplex: Time and Tradition in the Northern Mississippi Valley. Unpublished Ph. D. Thesis, Department of Anthropology, University of Wisconsin, Madison.

Rosebrough, A. L. and J. H. Broihahn 2005 A Cunundrum at Carcajou Point: 2004 Property Tax Exemption Program Investigations at the Carcajou Point Site (JE-0002) Lots W8651, W86252, W8537. Office of the State ArchaeologistTechnical Report Series No. 04-0002. Office of the State Archaeologist Historic Preservation Division-Public History Wisconsin Historical Society, Madison.

142

Sasso, R. F., R. F. Boszhardt, J. C. Knox, J. L. Theler, K. P. Stevenson, J. P. Gallagher and C. Stiles-Hanson 1985 Prehistoric Ridged Field Agriculture in the Upper Mississippi Valley. Report of Investigations No. 38. Mississippi Valley Archaeology Center, La Crosse, Wisconsin.

Schneider, S. A. 2008 Oneota Ceramic Relationships at Lake Koshkonong. Paper presented at the Midwest Archaeologcial Conference, Milwaukee.

Schneider, S. A. and J. A. Clauter 2007 Preliminary Analysis of Upper Mississippian Ceramics from the Crescent Bay Hunt Club and Schmelling Sites on Lake Koshkonog in Southeast Wisconsin. Paper presented at the 84th Annual Meeting of the Central States Anthropological Society, Minneapolis, Minnesota.

Schneider, S. A., J. A. Clauter and M. Brown 2006 The Ceramic Assemblage from the Crescent Bay Hunt Club Site (47-JE- 904) and Its Significance to the Late Prehistoric Southeastern Wisconsin. Paper presented at the 52nd Annual Midwest Archaeological Conference, Urbana, Illinois.

Schoolcraft, H. R. 1845 Oneóta ; or Characteristics of the red race of America from original notes and manuscripts. London, New York.

1854 Information respecting the history, condition and prospects of the Indian Tribes of the United States. Lippincott, Grambo & Co., Philadelphia.

Sill, J. R. 2003 Koskonong - The Lake We Live On. vol. 2010. Electronic document, http://webpages.charter.net/jsill/Koshkonong/index.htm, accessed April 11, 2010.

Skinner, A. B. 1921 Material culture of the Menomini. Indian notes and monographs no 20. Museum of the American Indian, Heye foundation, New York,.

Snell, S. P. 2005 Geographic Information Systems and Intrasite Analysis at the Bell Site (47WN9). Unpublished Master's Thesis, Department of Anthropology, University of Wisconsin-Milwaukee, Milwaukee.

Spector, J. D. 1974 Winnebago Indians, 1634-1829: An Archaeological and Ethnohistoric Investigation. Unpublished Ph. D. dissertation, Department of Anthropology, University Of Wisconsin, Madison.

143

Staeck, J. P. 1995 Oneota Archaeology Past, Present, and Future: In the Beginning, Again. In Oneota archaeology : past, present, and future, edited by W. Green and D. W. Benn, pp. 3-8 20. University of Iowa, Iowa City, Iowa.

Stahl, A. B. 1993 Concepts of Time and Approaches to Analogical Reasoning in Historical Perspective. American Antiquity 58(2):235-260.

Stapert, D. and M. Street 1997 High Resolution or Optimum Resolution? Spatial Analysis of the Federmesser Site at Andernach, Germany. World Archaeology 29(2):172-194.

Stevenson, K. P. 1994 Chronoloigical and Settlement Aspects of the Valley View Site (47-LC- 34). The Wisconsin Archeologist 75(3-4):237-294.

Stout, A. B. and H. L. Skavlem 1908 The Archaeology of the Lake Koshkonong Region. The Wisconsin Archeologist 7(2):47-102.

Stuiver, M. and P. J. Reimer 1993 Extended 14cData Base and Revised Calib 3.0 14c Age Calibration Program. Radiocarbon 35:215-230.

Taylor, R. E. 1987 Radiocarbon dating : an archaeological perspective. Academic Press, Orlando.

Thwaites, R. G., E. Kenton and ed 1925 The Jesuit relations and allied documents; travels and explorations of the Jesuit missionaries in North America (1610-1791). [The American library.; v. 23]; . A. & C. Boni, New York.

Vennum, T. J. 1987 Wild Rice and the Ojibway People. Minnesota Historical SOciety Press, St. Paul.

Wheatley, D. and M. Gillings 2002 Spatial technology and archaeology : the archaeological applications of GIS. Taylor & Francis, New York.

Whiting Young, B. and M. L. Fowler 2000 Cahokia: The Great Native American Metropolis. University of Illinois Press, Chicago.

144

Winkler, D. M. and R. J. Jeske 2009 Late-Pleistoncene Occupations in the Lake Koshkonong Region, Southeastern Wisconsin. Current Research in the Pleistocene 26:126-127.

Wobst, H. M. 1978 The Archaeo-Ethnology of Hunter-Gatherers or the Tyranny of the Ethnographic Record in Archaeology. American Antiquity 43(2):303-309.

Yarnell, R. A. 1964 Aboriginal relationshipsw between culture and plant life in the Upper Great Lakes region. Anthropological Papers 23. University of Michigan Museum of Anthropology.