Ceramic resource use in the Sand Canyon locality

Material Information

Ceramic resource use in the Sand Canyon locality implications for late Pueblo III community organization in southwest Colorado
Bevilacqua, Christopher Michael
Publication Date:
Physical Description:
247 leaves : illustrations ; 28 cm

Thesis/Dissertation Information

Master's ( Master of Arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Anthropology, CU Denver
Degree Disciplines:


Subjects / Keywords:
Pueblo pottery ( lcsh )
Pueblo Indians -- Antiquities -- Colorado -- Sand Canyon Pueblo ( lcsh )
Pueblo Indians -- Social life and customs ( lcsh )
Antiquities ( fast )
Pueblo Indians -- Antiquities ( fast )
Pueblo Indians -- Social life and customs ( fast )
Pueblo pottery ( fast )
Antiquities -- Sand Canyon Pueblo (Colo.) ( lcsh )
Colorado -- Sand Canyon Pueblo ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 226-247).
General Note:
Department of Anthropology
Statement of Responsibility:
by Christopher Michael Bevilacque.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
55222283 ( OCLC )
LD1190.L43 2003m B48 ( lcc )

Full Text
Christopher Michael Bevilacqua
B.A., University of New Mexico, 1999
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts

This thesis for the Master of Arts
degree by
Christopher Michael Bevilacqua
has been approved by
Catherine M. Cameron

Bevilacqua, Christopher Michael
Ceramic Resource Use in the Sand Canyon Locality:
Implications for Late Pueblo III Community Organization in
Southwest Colorado
Thesis directed by Professor Tammy Stone
This thesis examines ceramic resource use and its
implications for late Pueblo III (A.D. 1225-1300) community
organization in the Sand Canyon locality of southwest
Colorados central Mesa Verde region. The theoretical and
conceptual underpinnings of the research are first presented.
This is followed by a brief synopsis of the regions environment
and culture history, with attention given to the dynamic
character of Ancestral Puebloan society. The remainder of the
thesis describes the clay collection study, oxidation
experiments, and petrographic analysis that are used to address
the issues of ceramic resource use and community organization.
Gray ware assemblages from households that resided in village
and hamlet sites are compared in order to determine patterns of
ceramic resource use. The thesis results suggest that
socioeconomic differences existed between residents of large
and small sites, and are evident through analysis of well-
provenienced ceramic assemblages. The results also support the
notion that late Pueblo III gray ware production was a
household-level economy.
This abstract accurately represents the content of the candidates
thesis. I recommend its publication

This thesis is dedicated to my wonderful family. Their love, humor, and
support made this project possible.

Many generous individuals contributed their time, talent, and ideas
to this project. Id like to thank my committee for their comments and
guidance in this project. My advisor and committee chair, Tammy Stone,
deserves a special thanks for the time she gave to the project, and for her
guidance throughout my time at CU-Denver. Many folks currently or
formerly affiliated with Crow Canyon Archaeological Center influenced
this research. Richard Wilshusen, Andrew Duff, and Donna Glowacki
provided useful comments early in the design of this research. Susan Ryan
taught me a great amount about the archaeology of southwestern Colorado.
I am especially indebted to Mark Varien and Scott Ortman for all of their
advice and assistance with my research design and sampling strategy. Rob
Leitz of Metropolitan State College and Andrea Carpenter of The Space
Within lent me their petrographic expertise and provided invaluable
training, suggestions, and technical assistance. Rob Leitz was especially
helpful in accommodating my use of his institutions petrographic
microscope. Andrea Carpenter provided excellent digital images of thin
sections for a poster I presented at the 2002 conference of the Society for
American Archaeology. Michelle Hegmon of Arizona State University
also provided advice on my petrographic approach. And Mark Mercer of
Petrographic Services in Montrose, Colorado produced great thin sections
at a generous price. Susan Thomas at the Anasazi Heritage Center
promptly provided requested samples. This project was facilitated by a
grant from the Ward Weakly Scholarship Fund, provided by the Colorado
Council of Professional Archaeologists. The encouragement and financial
support I received from my parents is greatly appreciated. My wife Megan
wandered around Montezuma County with me in search of clay samples,
and was a constant reminder that there are more important things in life
than work. Many other individuals, friends and colleagues alike, gave me
their welcomed insights during this research, but I cant name them all. To
everyone that contributed to this research, I thank you. Any flaws in this
thesis are mine of course.

1. INTRODUCTION..............................................1
Definition of Concepts and Terms........................9
Land Tenure.......................................9
Archaeological/Architectural Correlates of the
Archaeological/Architectural Correlates of the
Cross-cultural Studies of Land Tenure..................23
Land Tenure and Resource Access in Southwestern
Resource Access, Ceramic Production, and Exchange......36

ENVIRONMENT AND CULTURE....................................55
Modem Environment................................64
Culture History........................................73
Paleoindian (pre-5500 B.C).......................74
Archaic (ca. 5500 B.C.-500 B.C)..................76
Basketmaker II (ca. 1000 B.C.-A.D. 500)..........79
Basketmaker III (A.D. 500-750)...................82
Pueblo I (A.D. 750-900)..........................85
Pueblo H (A.D. 900-1150).........................95
Pueblo HI (A.D. 1150-1300)......................105
4. METHODS..................................................119
Social Restriction of Ceramic Resource Use............119
General Methodology...................................124
Oxidation Experiments...........................124
Petrographic Analysis...........................127
Project-Specific Methodology..........................136

Sites and Sampling...............................136
Clay Collection and Firing Study.................147
Oxidation Experiment.............................152
Petrographic Temper Analysis.....................153
5. RESULTS....................................................156
Clay Collection and Firing Study........................156
Oxidation Experiment....................................162
Petrographic Analysis...................................169
Temper-Color Comparison.................................178
Oxidation Experiment.............................186
Petrographic Analysis............................188
Temper-Color Comparison..........................196
6. CONCLUSION.................................................200
Questions and Suggestions for Future Research...........208
1. Proveniences of analyzed sherds............................212

2. Descriptions of igneous rock types found
in the Ute Mountains......................................213
3. Flow chart used to determine igneous temper type..........221
4. Summary table of petrographic results.....................223
REFERENCES CITED....................................................226

1.1 The Mesa Verde Region and major geographic features....................4
4.1 Relative locations of the sampled sites..............................137
4.2 Plan map of Sand Canyon Pueblo......................................139
4.3 Plan map of Castle Rock Pueblo......................................140
4.4 Plan map of Lookout House...........................................142
4.5 Plan map of Saddlehom Hamlet........................................143
4.6 Clay collection locations...........................................149
5.1 Average percent shrinkage of each sampled geologic clay source.......157
5.2 Average apparent porosity of sampled geologic clay sources..........159
5.3 Frequency of fired clay colors according to geologic source.........161
5.4 Frequency of post-oxidation vessel color for each sampled site......163
5.5 Percentage of post-oxidation vessel color for each sampled site.....164
5.6 Percentage of site assemblages according to possible geologic source.166
5.7 Average apparent porosity of sampled vessels from each site.........169
5.8 Frequency of temper types according to site...,......................173
5.9 Relative percentages of sedimentary and igneous temper in each sampled

5.10 Percentages of igneous temper in the total assemblage
of each sampled site...............................................176
5.11 Percentages of sedimentary temper in the total assemblage
of each sampled site...............................................177
5.12 Relative percentages of sedimentary and igneous tempers in the
various color categories identified in the oxidation experiment....179
5.13 Percentages of various temper types in the color categories identified
in the oxidation experiment........................................180
5.14 Lookout House Frequency of temper types according to color
categories identified in the oxidation experiment..................182
5.15 Castle Rock Pueblo Frequency of temper types according to color
categories identified in the oxidation experiment..................183
5.16 Saddlehom Hamlet Frequency of temper types according to color
categories identified in the oxidation experiment..................184
5.17 Sand Canyon Pueblo Frequency of temper types according to color
categories identified in the oxidation experiment..................185

This thesis addresses two questions:
1. ) Did late Pueblo III households residing in different
settlement types in the Sand Canyon locality differ in terms of
ceramic resource use?
2. ) If differences in ceramic resource use did exist, are these
differences suggestive of socially restricted access to such
In the present context, ceramic resource use refers to both the acquisition
of finished vessels through exchange and the procurement of raw materials for
ceramic production. Thus the first question asks whether village households
differed from hamlet households in terms of either participation in exchange or
ceramic raw material procurement, since these two processes are largely
responsible for the characteristics of the sampled assemblages discussed below.
Preference for specific raw materials and geographic accessibility also may have
roles in creating differences among household assemblages. By addressing the
second question above, an attempt is made to determine whether assemblage
differences are the result of physical accessibility or personal preference, or if such

differences are related to variable participation in exchange or socially restricted
access to raw materials.
The questions posed in this thesis are important because their answers
allow inferences to be made regarding community organization. More
specifically, they provide a mechanism with which to examine how different
households within a community may have related to one another socially and
economically. Such community-oriented research also has implications for larger
anthropological issues such as aggregation, regional and local abandonment,
migration, and political organization. The inferences and implications drawn from
the research presented herein are discussed in the following chapters.
Archaeologists working in the southwest have noted the importance of
addressing community organization at the proper scale. Such an analytical scale is
larger than a single site, but smaller than a typical archaeological region, such as
the Mesa Verde region. By using such a scale, researchers emphasize the local
processes (Snead 1995: 31) that shape the social, political, and economic contexts
in which community members live. The concept of locality (Willey and Phillips
1958) is the precedent scale for studying community in the central Mesa Verde
region. In particular, the Sand Canyon locality has been an analytical unit for a
considerable body of research. The analytical units of this thesis are household
gray ware assemblages. However, the Sand Canyon locality is relevant here

because all of the sites considered in this study are within the locality and have
been studied with an explicit consideration of the locality concept (e. g. Lipe 1992;
Varien 1999a, 1999b). Furthermore, the gray ware assemblages analyzed in this
study are of interest not in themselves, but for what they can say about community
Lipe (1992: 2-3) defines the boundaries of the Sand Canyon locality
(Figure 1.1). The area largely is bounded on the north by Yellow Jacket Canyon,
and on the south by McElmo Creek. The eastern boundary is formed in part by
Alkali Canyon. The remaining boundaries are identified by arcs of approximately
7.5 km. around the community center sites of Sand Canyon Pueblo (5MT765) and
Goodman Point Pueblo (5MT604). The rationale for the 7.5 km. arcs is that this
distance represents half the average distance between Sand Canyon Pueblo,
Goodman Point Pueblo, and the six neighboring contemporaneous community
center sites. The Sand Canyon locality provides an excellent area for the research
presented herein, not only because of the considerable body of previous research,
but because it is a diverse geologic setting.
As noted above, the Sand Canyon locality is in the central Mesa Verde
region. This region is contained with the larger Mesa Verde, or Northern San Juan,

Figure 1.1. The Mesa Verde Region and major geographic features.
region. The Mesa Verde region has been variously defined. Recent work (Varien
1999: 13) defines the Mesa Verde region as extending from the La Plata River in
the east to Cedar Mesa in the west. The region is further defined by the
Abajo Mountains to the northwest, as well as the La Plata Mountains and Dolores
River to the northeast, and the San Juan River to the south. The central Mesa
Verde region is defined as the portion of Colorado between Mesa Verde proper
and the Utah border (Lipe 1995: 143-144). The central Mesa Verde region
contains both Canyons of the Ancients National Monument and Mesa Verde

National Park. Also mentioned in this thesis is the Totah region, which is the area
of northwest New Mexico that is north of the San Juan River and south of the
Colorado border (Lipe 1995: 144).
Chapter 2 provides a discussion of the theoretical background of this
research. The chapter begins by explicitly defining a few concepts that are
important to this research. Land tenure is first discussed. Then the concepts of
household and community are discussed in terms of their functions and their
archaeologically recognized correlates. Following is a brief discussion of previous
ethnographic research that focuses on land tenure. Then land tenure is discussed
in terms of past research in the Southwest. Finally, ceramic resource access and
how it may relate to production and exchange is considered.
An overview of the central Mesa Verde region is provided in Chapter 3,
which contains two main sections. The first section describes the physical
environment. This section begins with a brief description of the local geology,
followed by a description of modem plant and animal communities. The
environment section concludes with a description of paleoenvironmental research
in the region. The second section of Chapter 3 provides a synopsis of central Mesa
Verde culture history. Some relevant research done in other areas is also
discussed. This section is limited to the Paleoindian through Pueblo III periods.

An enormous amount of research has been done in this area, and Chapter 3 really
provides just a glimpse.
Chapter 4 describes the methods used to address the questions posed
above. These methods include vessel oxidation experiments and petrographic
analysis. The methods are first discussed in general, including descriptions of
previous research, then their specific applications in this project are discussed. A
clay collection study, also described in this chapter, was completed in order to
facilitate interpretation of oxidation experiment results. Chapter 4 also includes
descriptions of the sampled sites, as well as of the specific proveniences of the
sampled assemblages.
Chapter 5 provides the results of the clay collection study, oxidation
experiments, and petrographic analysis. The clay collection results include
measurements of porosity, shrinkage, and color in order to characterize the clays
available in the Sand Canyon locality. The results of the oxidation experiments
include measurements of color and porosity. It should be kept in mind that the
porosity results of the clay collection study and oxidation experiments are not
directly comparable, because of the effects of likely differences in clay paste
preparation between Pueblo III potters and the author. The petrographic results
provide frequencies and percentages of various temper types and classes identified
in the sampled assemblages. Results include statistical tests where appropriate.

The description of the results is followed by explanations of the characteristics of
the sampled assemblage. The results are then summarized with consideration of
their relationship to community organization.
This thesis concludes with Chapter 6, which discusses additional
implications for community organization. The chapter also provides a critique of
this research and mentions some ways in which it could have been improved.
Finally, some suggestions for future research are offered.

The issue of resource access in non-capitalist, agricultural societies is of
considerable anthropological interest because of its relationship to social, political,
and economic organization. Through an understanding of who has, and conversely
who does not have, access to important resources, knowledge can be gained
concerning organization at scales ranging from the household to regional level.
However, resource access is most relevant at the scale of the local-level
community because this is the scale at which land tenure systems are negotiated
and social roles defined (Adler 1996: 350).
The theoretical discussion presented in this chapter is intended to aid in the
understanding of how household settlement context in the Sand Canyon locality
during the late Pueblo III period may have influenced ceramic resource use. The
degree of aggregation witnessed during this time (discussed in more detail in the
following chapter) certainly required some reorganization of social, political, and
economic relationships. This chapter is intended to shed light on how this
reorganization may have influenced ceramic resource access and use.

Definition of Concepts and Terms
Land Tenure
Land tenure refers to the system through which individuals socially define
and assert their rights to access and use various resources (Adler 1996: 337; Varien
1999b: 4). Adler (1996) uses a land tenure model to explain changes in food
production, settlement size, and community organization. Importantly, Adler
(1996: 338) points out that land tenure systems are socially negotiated, historically
contingent, and are found in all types of societies.
Anthropological discussions of land tenure among middle-range
agricultural societies are, in general, limited to a discussion of access to productive
agricultural lands. This is not surprising given the obvious importance of
maintaining access to these lands. However, it is argued here that in such societies
the land tenure system that governs access to agricultural land also influences
access to other vital resources. Although the maintenance of access to these non-
agricultural resources may not be as critical as the access to food-producing lands,
the potential symbolic, spiritual, historical, and/or functional importance of such
resources should not be underestimated.
Land tenure systems are negotiated by individuals, as well as by a variety
of other social units. Cross-cultural data suggest that the social unit that maintains
and defends resource access rights may be influenced by the nature of the

resources themselves (Wilk and Rathje 1982: 627). Although land tenure maybe
governed by groups such as clans or lineages, the analytical units of particular
relevance in this study are the household and the community. These social units
are particularly relevant because they have well-established correlates in the
archaeological record of the study area. These concepts are discussed below.
The household is a crucial component of this and many other studies of
social organization. The importance of the household in understanding social
organization rests in the fact that the household is the basic social, political, and
economic group in most societies (T. Stone and Howell 1994: 103). Consideration
of the household as an analytical social unit may also be useful in narrowing the
gap between grand theories of culture change and the archaeological record (Wilk
and Rathje 1982: 618). This is because grand theories generally deal with society
at large, but archaeologists more commonly study contexts in the archaeological
record that were created by smaller social units such as households. Although this
study focuses on community organization, it is necessary to explicitly consider the
community's essential building blocks.
Rather than defining a household as a nuclear family or some other real or
fictive kin group (i.e., a compositional definition), it is more appropriate for

archaeologists to define the household behaviorally (Lightfoot 1993: 12-13). This
distinction is necessary because the archaeologist has the ability to find and
interpret behavioral correlates in the archaeological record, but the ability to infer
kinship from material remains is limited. Although a thorough discussion is
beyond the scope of this project, it is interesting to note that archaeologists in the
Southwest have also worked to identify the composition of household groups (e.g.,
Lightfoot 1993; Ortman 1998).
The work of Netting, Rathje, and Wilk (Netting et al. 1984; Wilk and
Rathje 1982; Wilk and Netting 1984) has been quite influential in the way
archaeologists conceptualize the household (e.g., Lightfoot 1993, Varien 1999a).
These authors have noted that the behavioral and compositional variability of
households around the world contributes to the difficulty of precisely defining the
concept. Netting, Rathje, and Wilk identify four functions of households that are
used to define this basic social unit. These functions include production,
distribution, transmission, and reproduction. Each of these functions is interrelated
with one another and with a group's composition, resource availability, and mode
of production (Wilk and Rathje 1982: 631-633). These four household functions
are described in detail by Wilk and Rathje (1982: 622-631) and Wilk and Netting
(1984: 6-17). A summary is provided here.

Production refers to the procurement of resources or to the manipulation of
resources with the intent of increasing their value. Different types of production
are distinguished based on the number of individuals that participate. Production
by a single individual is referred to as being linear, and production by multiple
individuals is referred to as simultaneous. If a potter were to harvest her own raw
materials, process them, shape vessels and fire them, this would be an example of
linear production.
Simultaneous production can be further divided into that which is simple or
complex. Simple simultaneous production occurs when all individuals in the
producing group do the same activity at the same time. For example, a group of
individuals pulling weeds together in an agricultural field would be simple
production. Complex simultaneous production involves a division of labor in
which different individuals are responsible for different tasks that are part of the
larger production process. The building of a masonry structure would be a good
example of complex simultaneous production if some individuals quarry stone,
while others shape the stone, prepare the mortar, and place the stone.
Distribution refers to the movement of goods from producers to consumers.
In the scheme of Netting, Rathje, and Wilk, distribution also includes the
consumption of these goods. The function of distribution can be divided into two
strategies: pooling and exchange. Pooling and exchange are not mutually

exclusive distributive functions and most households probably combine the two to
some degree. Whether a household focuses on pooling or exchange depends
largely on the group's mode of subsistence and production (Wilk and Rathje 1982:
625). Some households may not be willing to pool resources if some individual
members are not able to contribute to the pool. Likewise, it would not be expected
for a household to rely on exchange as a distributive function if all of its members
participate in the same productive enterprise.
Transmission refers to the passing of resource rights and ownership from
generation to generation (Lightfoot 1993: 157; Varien 1999a: 17; Wilk and Netting
1984; Wilk and Rathje 1982). Transmission also includes the maintenance and
defense of resource use and access. This is likely to be an important household
function in circumstances of restrictive resource access. Netting, Rathje, and Wilk
(Wilk and Netting 1984: 11-14; Wilk and Rathje 1982: 627-630) offer cross-
cultural data to suggest that transmission occurs at a smaller social scale with
resource scarcity and with agricultural intensification. In a situation of high
resource availability, a lineage or entire community may be responsible for the role
of transmission. Conversely, when resources are more limited transmission is
more likely to occur at the household level or even between individuals. The
ethnographic data discussed by Netting, Rathje, and Wilk suggest that an

association between resource availability and social scale of transmission is a
common trend in agricultural societies.
Reproduction includes both biological procreation and the socialization of
children. Of all the social functions discussed here, reproduction is the one that is
most strictly limited to the household. However, this is not to underestimate the
dynamic and variable nature of marriage and human child rearing.
This leads directly to a discussion of coresidence. Coresidence of
household members seems to be the norm cross-culturally and even more so in the
mainstream culture of the contemporary United States. However, it is not a
prerequisite for defining a social group as a household. Obviously, parental
coresidence is not necessary for biological reproduction. However, the
socialization of offspring favors coresidence, not necessarily by the biological
parents, but by some group of household members. Wilk and Netting (1984: 17-
19) discuss several cases in which individuals do not reside with other household
members but continue to play an important role in the household. These authors
suggest a few factors that influence coresidence, including the availability of
housing (or the materials to construct housing), marriage norms, architectural
style, and labor demands that encourage absenteeism.
It is important to emphasize that household functions are interrelated with
one another, and also with household composition and the larger social structure.

The way in which one function is carried out has a considerable effect on the
execution of the other functions. For example, a variety of effects may occur if the
mode of production encourages intermittent coresidence (Wilk and Netting 1984:
19). Intermittent coresidence means that there are at times one or more individuals
that do not need to be supported by the distribution function. Fewer consumers
can translate into decreased production requirements and/or the possibility of
surplus accumulation. The possibility of household surplus accumulation resulting
from an absentee household member relies on a context in which the absentee did
not produce more than he/she consumed when present. Otherwise, lost labor
would offset the potential for a higher ratio of production to consumption.
Intermittent coresidence also implies that there are, at times, individuals not
available to assist in household reproduction. However, if the absentee household
member provides a greater number of resources than the coresident household
members are able to acquire, this may allow the coresident household members to
allot more time to the socialization of children. The number of possible scenarios
that can be used to exemplify the interrelationship between household functions,
household composition, and social context is virtually unlimited (see Wilk and
Netting 1984: 19-21).
As briefly mentioned above, household functions are, for the most part, not
limited to the household. Production, distribution, transmission, and even

reproduction (since it includes childhood socialization) can be and often are
accomplished by other social units, such as a clan, lineage, community, etc.
Furthermore, although all of these functions must occur in any enduring social
group, one social unit need not accomplish all the functions. There are also
differences in the degree to which different social units may participate in the same
function simultaneously. For example, a community may maintain and defend the
resource access rights of its constituent households (i.e., transmission), although
the household may be responsible for the inheritance (also transmission),
production and distribution of these resources, without a significant contribution of
communal labor.
Considering the discussion above, a household can be defined as the
smallest scale social group whose members interact most frequently (Lightfoot
1993: 12), typically reside with one another, and are to some degree responsible
for the functions of production, distribution, transmission, and reproduction. The
composition of households is dynamic and varies considerably across time and
space. Consequently, no single definition of the household may be suitable for
every question of interest to the social scientist. Although households may have
other important aspects pertaining to ritual or politics, for example, the particular
definition offered here is a largely economic conception. This approach is suitable
for the research presented herein.

Archaeological/Architectural Correlates of the Household
The association of a kiva, surface structure, and midden is referred to as a
unit pueblo (Prudden 1903) or kiva suite (e. g. Bradley 1992). Unit pueblos are a
basic component of the Mesa Verde region settlement and architectural patterns,
and were common from Pueblo I times until the residential abandonment of the
area. Unit pueblos occur in a variety of architectural associations. They can be
found in relative isolation, in dispersed clusters, or in aggregated villages such as
Sand Canyon Pueblo and Castle Rock Pueblo (Adler 1992: 22). Varien (1999a:18-
19) makes the argument that unit pueblos, at least in the Sand Canyon locality, are
the architectural manifestations of households. Lightfoot (1993: 129) and Huber
(1993: 236) also define the household as the residential group that occupies a unit
pueblo. Although not all archaeologists would agree (e.g., Reid and Whittlesey
1982), this interpretation appears to be acceptable to many archaeologists working
in the central Mesa Verde region. This precedent is followed in this study.
The implications of how the household and unit pueblos are defined
archaeologically is worth noting here. For example, if a unit pueblo is used to
infer a single household, this would give a different population estimate than
would using another household proxy (see Lightfoot 1993: 145-162).
Furthermore, if kivas are interpreted as integrating social units other than
households, this could be at the expense of household autonomy. Conversely, a

high level of household integration, suggested by the use of a unit pueblo by a
single household, may inhibit the integration of larger social groups such as the
A considerable amount of anthropological research has been done with
explicit consideration of agricultural communities (e.g., Abbott 2000; Adler et al.
1996; Bauer 1987; Duff 1994; Huber 1993; Lipe 1992; Mahoney et al. 2000; G.
Stone 1991; Snead 1995; Tannenbaum 1992; Varien et al. 2000). Like most
concepts in anthropology, the way in which communities, especially prehistoric
ones, are conceptualized is subject to debate (Abbott 2000: 11-15). However,
most conceptions of community have in common the notion of a territorial social
unit held together by some degree of interdependence.
Murdock (1949) and Murdock and Wilson (1972) have been perhaps the
most influential in the way researchers in the Mesa Verde region define the
community. According to these authors a community is defined as the maximum
number of individuals that reside together and have face-to-face interaction. This
interaction occurs on a daily basis and provides a source of social identity through
shared group membership. This type of community has been referred to in the

literature as a "residential," "face-to-face," or first-order community (e.g., Lipe
1992: 3; Varien et al. 2000: 47).
It is generally assumed that the community provides the primary social
context in which land tenure systems are negotiated. In particular, the shared
group identity of community members plays a considerable role in defining access
to community resources (Snead 1995: 27). Community resources can include both
natural and social resources (Varien 1999a: 20).
Kolb and Snead (1997) emphasize the need to define communities in a way
that facilitates comparison of multiple cases. These authors define community as
"a minimal, spatially defined locus of human activity that incorporates social
reproduction, subsistence production, and self-identification" (Kolb and Snead
1997: 611). This definition is primarily applicable to small-scale agricultural
societies, but may be extended to describe more complex groups as well.
The concept of community as it is used here largely follows the
aforementioned definitions. However, it is argued here that the community may
also play a considerable role in non-subsistence production by protecting non-
subsistence resource rights of its constituency. Additionally, the position taken
here is that daily face-to face interaction is neither necessary for the formulation
and maintenance of common group identity, nor for the development and
perpetuation of land tenure systems. This conceptualization of the community

expands the potential spatial scale of the community and allows for the possibility
that Sand Canyon Pueblo and Castle Rock Pueblo represent a single community in
terms of access to ceramic resources. The concept of community as it is used here
includes the notion that communities are not only groups of individuals with a
shared identity but also a group of individuals with shared risks and a considerable
degree of interdependence (Adler et al. 1996: 403).
In sum, community can be defined as a group of individuals with a shared
system of land tenure, group identity, and economic risks that reside near enough
to one another to have sufficient interaction for the maintenance and defense of a
territory that provides the majority of the group's resources. It is not the intention
here to imply that all individuals or households that comprise a community have
equal access to community resources (Adler and Varien 1994: 84). Although
communities function to assert access rights, all households and individuals that
comprise the community do not necessarily appropriate these resources in the
same way. In addition, communities do not exist in isolation; community
members have social ties to individuals in other communities that influence their
social roles and relations within their own community, as well as with neighboring
communities. Furthermore, it is important to emphasize that communities are
dynamic entities with the potential for changes in membership and social
organization. The definition given above is not intended to apply to all

communities at all points in time. Rather it is meant to be a heuristic device
formulated for the purposes of this project.
Archaeological/Architectural Correlates of the Community
Settlement and architectural patterns typically provide archaeologists with
the material correlates of communities. Archaeologist in the central Mesa Verde
region and elsewhere in the Southwest have defined the archaeological correlate of
the community during the Pueblo II and Pueblo III periods as a community center
site (recognized by the presence of public architecture) and the surrounding cluster
of smaller residential sites (Adler 1992: 22; Lipe and Varien 1999: 272, 291 ;
Snead 1995: 26; Varien 1999a: 23; Varien et al. 1996: 101). The construction and
use of public architecture simultaneously provided both integrative and
differentiating functions via inclusion or exclusion in related activities (Varien
2002, personal communication).
The Pueblo III period experienced the highest degree of population
aggregation in the prehistoric central Mesa Verde region. In such a context of
increasing agricultural intensification and social circumscription of resource
access, the integrative function of public architecture would have strengthened
community identity. As mentioned above, this identity plays an important role in
land tenure. Although there were major changes in architecture and settlement

patterns throughout the Pueblo periods, integrative facilities were functionally
persistent for hundreds of years (Adler 1992: 23).
Ethnographic research by Tannenbaum (1992) brings up an important issue
for anthropologists that study social organization at the community level. The
point made by Tannenbaum is especially relevant for archaeologists. She argues
that the functionalist perspective of early ethnographers has led anthropologists to
assume a priori that clusters of households represent a cohesive organizational
form that is referred to as a village. Tannenbaum emphasizes that there is an
organizational range in household clusters. At one end of this continuum are fully
autonomous, independent households (even though they are clustered spatially).
At the other end are clusters of individual households that have "little independent
reality" (Tannenbaum 1992: 259). The latter groups have the village or other
social entity as their primary organizational form. In her study of Tai
communities, she found that most household clusters were somewhere in the
middle of this continuum.
It is assumed here that Pueblo HI communities in the Mesa Verde region
were also not at one or another extreme of the continuum discussed by
Tannenbaum. In accord with this assumption, late Pueblo III communities were
important organizational entities that were realized as such by their members,
although households retained some degree of independence. It is argued here that

these assumptions are justified by the presence of public architecture, as well as by
the continuity of the unit pueblo architectural pattern at aggregated community
center sites. This assumption and its justification have been expressed elsewhere
(Varien 1996: 102).
Cross-cultural Studies of Land Tenure
Ethnographic studies have documented the variable and dynamic nature of
land tenure in agricultural groups around the world. These studies have also
linked land tenure with a variety of behaviors and social phenomena. Although
these phenomena are not mutually exclusive, they include aggregation, population
movement, changes in village size, socioeconomic differentiation, warfare,
communal production, settlement patterns, construction of landscape features, and
agricultural risk minimization. These social activities typically leave material
remains in the archaeological record. Therefore, a consideration of these studies
can be useful to archaeologists hoping to understand the organization of resource
access in past societies. Although the use of ethnographic studies in understanding
the archaeological record may be criticized because of implied assumptions of
historical continuity and uniformitarianism, the purpose here is quite to the
contrary. These studies highlight the dynamic nature of land tenure and make

clear the variability in the ways agricultural groups deal with issues of resource
Bauer (1987) describes land tenure among the Tigray of Ethiopia as a
malleable system that can be changed by community members when the need
arises. The Tigray practice two forms of land tenure at different times: a
communal system known as chigurafgoses and a hereditary system called risti.
The chigurafgoses system allows all members of a village access to the productive
lands of the village. This communal system is in place when community members
wish to increase their village population. Bauer (1987: 225) mentions that the
reasons why villagers would want to increase the size of their population include
attempts to increase political power and to better defend economic resources.
There comes a point, however, when these villages can not take on greater
numbers. Land tenure shifts to the risti system when the village population grows
to the point at which new arrivals can only gain access to communal agricultural
lands by taking it from preexisting households. Although not described explicitly
by Bauer, the decision to change the land tenure system seems to be based on
village consensus. This shift in the land tenure system limits population influx and
allows villagers with the necessary hereditary status to maintain access to
productive lands. Furthermore, the shift in Tigray land tenure affects population
movement not just at the local level but also on a regional scale.

Curry (1997) also documents the dynamic nature of land tenure among the
Wosera Abelam of Papua New Guinea. The Abelam have traditionally practiced a
system of patrifiliation in which men inherit land rights from their fathers (rather
than from more distant ancestors). In this system, resource access rights are
defined by an individual's status as an agnate or non-agnate. Agnates have access
to productive lands inherited from their fathers, and they have social power that
stems from the ability to control the land tenure system (Curry 1997: 195).
Curry (1997: 201) argues that a demand for land among the Abelam has
lead to a culture of warfare. The common occurrence of warfare, in combination
with natural causes, served to limit population before the impact of the modem
state. Before the effects of government rule, the Abelam were regularly
attempting to increase the size of their communities. Larger group size allowed
the Abelam to defend their resources more effectively. In such a social context,
the system of patrifiliation allowed non-agnates to become assimilated into the
community as agnates, with the corresponding rights of resource access. Non-
agnates are accepted as agnates because fellow community members socially
acknowledge them as such.
With state impact, both warfare and natural deaths among the Abelam have
decreased. This has lead to a reduced need to increase group size while
simultaneously doing just that. In addition, the growing importance of cash crop

agriculture has lead to increasing land shortages. As a consequence, the Abelam
have become more restrictive in terms of land tenure, making it more difficult for
non-agnates to gain agnate status. This has, in effect, led to a decrease in
intervillage residential mobility and a decrease in the socioeconomic status of non-
agnates. One can imagine how this increasingly restricted access to agricultural
lands may provide an impetus for the development of specialized craft production.
Ethnographic research has also documented the use of physical markers to
assert rights to certain resources. The use of such symbols appears to be
characteristic of societies in which decreased mobility and more permanent land
tenure accompany agricultural intensification. G. Stone (1994), for example,
describes the use of perimetric features by Kofyar farmers in Nigeria. These
perimetries can occur in a variety of forms, including stone alignments or walls,
rows of vegetation, or worn footpaths. These features function to mark the
boundaries and defend the ownership of agricultural fields. Their usefulness in
defending resource claims comes not from a physical ability to keep potential
intruders at bay, but from the recognized symbolic meaning of such features.
Stone mentions that land boundaries can also be enforced socially, without the use
of perimetric features. This social boundary enforcement, like perimetries,
depends on shared group ideas concerning land tenure and authority.

In a discussion of land tenure and community organization, it is important
to realize that groups that assert and defend resource access rights may or may not
necessarily participate in communal production. Among the Kofyar (G. Stone
1994: 321), for example, individual farmers (rather than larger corporate groups)
construct perimetric features although the agricultural strategy in this area relies
greatly on communal labor (G. Stone 1991). Conversely, at the Tiv hamlet of
Ukwese in central Nigeria, farmers collaborate in acts of intimidation to defend
their land, but do not cultivate communally (G. Stone and Downum 1999: 121).
Perhaps analogous to the large, mesa-top oriented sites of the central Mesa Verde
region, the Tiv explicitly place their settlements at prominent points on the
landscape as part of their strategy of intimidation (G. Stone and Downmum 1999:
The ethnographic studies described above have considerable archaeological
implications and can be useful in the formulation of hypotheses that can be tested
in the archaeological record. The work of Bauer (1987) emphasizes the fact that
land tenure systems are malleable and socially negotiated. Bauer shows that
community members can actively manipulate land tenure systems in order to
create social circumstances that allow for more effective defense of household
resource claims. Together with the research of Curry (1997), these works show
how changes in land tenure can influence social phenomena such as village

growth, kinship, local and regional population movement, violence and warfare,
and changes in settlement location. Many of these behaviors have material
correlates that can be identified in the archaeological record. As these material
correlates are identified, scholars can create research designs aimed at determining
the nature of community organization and whether or not these social phenomena
were related to changes in local land tenure.
Land Tenure and Resource Access
in Southwestern Archaeology
As is the case in most ethnographic studies, archaeologists working in the
Southwest have also focused on agricultural land when discussing issues of
resource use and access. Several influential works from the Southwest are
described below, focusing on the central Mesa Verde region but also including
some other southwestern cases. Finally, this chapter concludes with a discussion
on how the land tenure/resource access issue relates to ceramic production and
When dealing with the issue of resource access, one of the most important
things to consider is the social scale at which the land tenure system operates. A
society's primary resource access group (sensu Adler 1994: 85) can be the
household, community or some other corporate group {sensu T. Stone and Howell
1994: 104). As suggested in the above discussion, both households and

communities play similar economic roles. In an economic sense, their primary
difference is largely a matter of scale. Considering the scale at which the land
tenure system is negotiated is important because of the implications concerning
social, political, and economic organization.
Adler (1992: 22) describes the later occupation of the central Mesa Verde
region by ancestral Puebloans as being characterized by the landscape becoming
increasingly crowded with both people and claims of resource access rights.
Although not all authors would agree (this seems to be a direct contradiction of
Wilk and Rathje 1982), Adler (1994) argues that there is a positive correlation
between resource scarcity and size of a primary resource access group. Adler
(1994: 87) also argues that the primary resource access group tends to be the
household among groups with a low level of agricultural intensification. Adler
supports this argument with cross-cultural data from ten southwestern cases and
twenty-five worldwide cases.
Continuing along this line, Adler further argues that aggregation into large
villages (such as Sand Canyon Pueblo and Castle Rock Pueblo) was the result of
an increase in the size of the primary resource access group. An implicit
assumption that village coresidence is associated with communal resource access
is made. Thus land tenure is seen as a risk-buffering mechanism (Adler 1996), and

aggregation is seen as an attempt to "insure community access to the increasingly
circumscribed productive environment" (Adler 1994: 99).
To summarize my understanding of Adler (1992,1994,1996), when
resources are abundant it is expected that small-scale corporate groups such as the
household control resource use and access. Conversely, larger social groups such
as communities dominate the land tenure system in contexts of population-
resource imbalance. In such a community-oriented land tenure system, it is
expected that community members must recognize one another's status as such in
order to have access to the productive resources of the community (Adler 1996:
344). It is asserted here that an effective means of maintaining a shared group
identity is to live with other community members in the same village.
Research in other areas of the Southwest suggests that the relationship
between land tenure, aggregation, and agricultural intensification varies across
time and space. G. Stone and Downum (1999), for example, suggest that in the
Wupatki area of northern Arizona aggregation occurred as a conscious effort to
avoid subsistence intensification in a context of major population influx. These
authors suggest that environmental conditions imposed constraints that required
the inhabitants of the Wupatki area to maintain an extensive agricultural strategy.
Like Adler, G. Stone and Downum argue that aggregation can function to
define resource use and exclude others from access to important resources. They

note that in order for a social group to effectively exclude others from claimed
resources, some notion of authority is necessary. This authority can be
strengthened and legitimized by including more people in the social group. At
Wupatki, aggregated social groups built large pueblos that served as monuments of
group authority, thus increasing the community's ability to threaten to defend its
resources as a group (G. Stone and Downum 1999: 121). In the Wupatki case, the
primary resource being defended was not prime agricultural land suitable for
intensification. Rather it was the broad expanses of land used as part of an
extensive agricultural strategy.
G. Stone and Downum suggest that not all individuals were welcomed
equally under the authority of aggregating social units. Socially defined groups
are most easily defined in contrast to other groups. Basing their argument largely
on ceramic typology, G. Stone and Downum argue that aggregation in the Wupatki
area occurred along Sinagua, Kayenta, and Cohonina ethnic lines.
Although aggregation has been cited as one way communities can maintain
and defend resource rights, research suggests that this same phenomenon (i.e.,
aggregation) can also function to alleviate resource competition. Walsh (1998)
documents such a situation in his study of lithic material use and competition on
the Pajarito Plateau of northern New Mexico. Walsh emphasizes the relevance of
resource competition to the issue of aggregation. He argues that the poor

agricultural soils and low density (but high variability) of wild resources on the
plateau was conducive to the maintenance of large subsistence territories by the
plateau's inhabitants. Walsh assumes that the acquisition of lithic materials was
embedded in the exploitation of these large subsistence territories. These resource
territories are identified using the spatial distribution of geologically discrete lithic
materials. During the Late Coalition period (A.D. 1250-1325), the population
density on the Pajarito Plateau increased considerably, largely as the result of
immigration (possibly by emigrants of the Mesa Verde region?). This population
increase coincided with an increase in resource competition. Faced with such a
situation, Walsh argues that the plateau's inhabitants alleviated resource
competition during the Classic period (A.D. 1325+) by implementing two
solutions: emigration and aggregation.
In addition to aggregation, there are other archaeologically documented
ways in which prehistoric inhabitants of the Southwest seem to have asserted and
defended resource access rights. G. Stone and Downum (1999) have documented
the use of perimetric features at Wupatki that are somewhat analogous to those
used by the Kofyar (G. Stone 1994), as discussed above. G. Stone and Downum
argue that these features served as visual indicators of resource use rights.
The actual and/or past use of a resource is another strategy individuals and
social groups can use to maintain access to a particular resource. The use of a

resource is frequently accompanied by material symbols that let others know that
resource claims have been made. Field houses in the northern Southwest may
have been one such symbol. Preucel (1988: 236) argues that field houses on the
Pajarito Plateau could have been built as explicit expressions of land ownership
(Kohler 1992: 620).
Kohler (1992) suggests that field house construction in the central Mesa
Verde region also represents the increasingly scarce availability of agricultural
lands. He points out that the shift from canyon-oriented to mesa-top agricultural
practices is associated with the initial appearance of both field houses and
aggregated villages. Kohler (1992: 621) argues that these field houses could have
served two functions, which are not mutually exclusive. First, the construction of
field houses would have served to minimize the costs of working agricultural
fields that, through time, became increasingly distant from habitations. Second,
field houses may have served to define use ownership of productive lands. This
use ownership may have been defined before potential agricultural lands were put
into production, and may have superseded the use of this land for agricultural
purposes. Kohler further argues that the cooccurrence of field houses and
household aggregation is indicative of land tenure system in which resource use is
in theory controlled by a community, although resource rights are in practice
maintained by households. Kohler supports this interpretation with ethnographic

evidence, and archaeological evidence in other areas of the Southwest suggests a
similar scenario (Kolb and Snead 1997).
Research in the northern Rio Grande region (Kolb and Snead 1997; Snead
1995) suggests additional symbolic mechanisms that were used to defend
community territory in the prehistoric Southwest. For example, Kolb and Snead
(1997: 620) cite both ethnographic and archaeological evidence to argue that
shrines served to demarcate territorial boundaries, to both community members
and non-residents. These authors further argue (Kolb and Snead 1997: 621-622)
that the concentration of petroglyph sites along routes leading to the core of the
Tsikwaiye community suggests that these visual symbols functioned to maintain
community boundaries and emphasize group identity. Perhaps shrines and rock art
sites in the central Mesa Verde region also functioned as symbols of territory and
group identity.
In his discussion of prehistoric warfare in the Southwest, LeBlanc (1999:
11-13) describes "scarce-resources" models as those that describe warfare as the
result of competition for resources. Although productive land is often the focus of
such models, LeBlanc suggests that competition for resources such as exotic goods
or wives may also stimulate warfare. The evidence for warfare in the Mesa Verde
region is quite considerable (see Kuckelman et al. 2000; Kuckelman et al. 2002),

and perhaps reflects an extreme strategy to manipulate the land tenure system of
local communities.
As mentioned above, an individual's social status may influence his/her
ability to access resources. Citing several Pueblo ethnographic sources, Brandt
(1994) warns archaeologists against assuming that members of Ancestral Puebloan
communities shared equal access to resources. Brandt (1994: 15) argues that
differential access to resources was not uncommon in ethnographically
documented Pueblos. This differential access was most pronounced in terms of
ceremonial knowledge and materials associated with such knowledge. According
to Brandt, there were clear status distinctions in these communities, and secrecy
was a primary mechanism used to control knowledge of, and access to, particular
plants, minerals, and animals with ritual significance. Brandt (1994: 19)
summarizes her main argument most succinctly in stating that "open access to
what we would consider basic resources, free for the taking characteristic of less-
centralized societies does not occur."
Archaeological evidence of possible differential resource access has been
documented in the Southwest. Research by Driver (1993), for example, is
suggestive of differential access to valued faunal resources in the study area during
the late Pueblo in period (Lipe and Varien 1999: 336-337). Driver found
considerably higher frequencies of artiodactyl remains at one of the kiva-

dominated architectural blocks at Sand Canyon Pueblo. In addition, although the
faunal assemblage from Sand Canyon Pueblo includes roughly 14 per cent
artiodactyl remains, deer is virtually non-existent in the assemblages from nearby,
small, contemporaneous sites (Lipe and Varien 1999: 337).
In sum, research in the Southwest has suggested that there are many ways
in which land tenure articulates with other social phenomena, including
aggregation, agricultural intensification, population movement, ethnic
consolidation, and changes in the social units that govern resource use and access.
It has also been shown that the prehistoric people of the Southwest likely asserted
and defended their resource rights in a variety of ways. These mechanisms may
include field house and shrine construction, the use of perimetric features, rock art,
intimidation, and violence. Finally, previous research provides both ethnographic
and archaeological evidence for differential access to resources in the Southwest.
The primary goal of this thesis research is to determine if household settlement
context in the Sand Canyon locality during the late Pueblo III period is interrelated
with household access to, and use of, ceramic resources.
Resource Access. Ceramic Production, and Exchange
The compositional variability in ceramics from any particular
archaeological provenience is largely the result of either the raw materials used in

local production or a reflection of exchange relationships. Consequently, this
research must consider ceramic resource access in terms of both production and
exchange. Although other mechanisms, such as population movement, may
influence archaeological ceramic assemblages, production and exchange are
considered here to be the most influential.
All economic systems consist of three interrelated components: production,
distribution, and consumption (Costin 1991; Pool 1992). Similarly, production
itself is comprised of multiple stages. In the case of ceramic production, these
stages include raw material procurement, preparation of these materials, vessel
formation, drying, surface treatment, and firing. Previous research (Bemardini
2000) suggests that it may be useful for understanding production organization to
consider these stages as potentially discrete social activities. This study considers
the procurement of raw materials, the first stage in the production process that
leaves a trace in the archaeological record.
A considerable amount of anthropological research has been conducted in
attempts to derive models that explain and/or describe why potters use particular
resources. Consequently, a variety of both economic, social, and political factors

have been found to influence raw material procurement. Following is a brief
discussion of previous research related to this issue.
Traditionally, studies of ceramic raw material selection have focused
largely on technological and economic factors (Neupert 2000: 251). Bishop et al.
(1982: 315-318), for example, outline several possible clay procurement strategies.
Bishop et al. focus on the ways these procurement strategies influence ceramic
compositional variability. However they argue that each strategy is enacted within
a preferred resource procurement zone, the borders of which are largely
determined by the costs and benefits involved. Bishop et al. briefly acknowledge
the role of social factors in their typology of ceramic resource procurement
Arnold (1985: 35-60) also argues that non-industrial potters utilize
particular resources according to the efficiency with which such resources can be
obtained. Using available ethnographic data, Arnold finds that 84 per cent of
documented potters obtain clay resources from within 7 km. of their residence
(N=l 11). Similarly, 97 per cent of the ethnographic cases (N=31) obtained temper
materials from within 9 km. Although very little ethnographic data exist on
distances traveled to obtain slip, paint, and glaze resources, potters tend to travel
more to obtain these resources. Nevertheless, 36 per cent of the cases for which
information was available (N=36) obtained these resources from within 10 km

(Arnold 1985: 56). Arnold suggests that potters are willing to procure slip, paint,
and glaze resources from greater distances because they are used in lesser
quantities than clay or temper.
Arnold argues that non-industrial potters tend not to exceed the
aforementioned distances because procurement costs begin to exceed returns.
Although Arnold (1985: 57) briefly mentions the effect of local geology on
ceramic production, he does not explicitly consider the relationship between the
distances traveled to obtain ceramic resources and the geographic distribution of
such resources. Consequently, the observation that most groups do not exceed
certain distances to obtain resources may say just as much about the natural
distribution of suitable resources as it does about attempts to maximize returns and
minimize costs.
More recently, anthropologists have begun to move away from strictly
economic models of ceramic resource procurement and have begun to examine the
social and political factors that influence what materials potters choose. Costin
(2000: 380), for example, points out that there are many natural and cultural
factors that influence raw material selection (and, consequently, variability in paste
composition). Costin (2000) notes that potters frequently make selection decisions
based on the performance of raw materials, as well as on the efficiency with which
they can be obtained. Resource procurement also can be influenced by claims of

ownership, or other social restrictions on resource access. Furthermore, the
organization and technology of production play a considerable role in the decisions
potters make regarding raw material use. Costin's (2000) most important point for
the purpose of this thesis is that ceramic resource procurement strategies are cross-
culturally quite complex.
Several recent ethnoarchaeological publications illustrate this complexity.
In a study of ceramic production and social boundaries among the Kalinga, Stark
et al. (2000) argue that the compositional variability found in locally-made
ceramic samples reflects social boundaries between the villages of Dalupa and
Dangtalan. Although these villages are only 2 km. from one another they use
discrete clay sources. The use of different clays is argued by the authors to reflect
the sociopolitical affiliation of each village.
Although the potters from these villages select clays based on variables
such as shrinkage and plasticity (performance and workability characteristics), the
social relations between potters and source owners play a potentially overriding
role. In the case of the Dalupa potters, clay sources occur on agricultural fields
owned by non-potters. Field owners receive no compensation for mined clay, and
they have the ability to cut-off access to clay sources if there is a perceived threat
to the productivity of their land. Stark et al. (2000: 307-308) also mention that

clay resource owners can terminate access for other socially or politically
motivated reasons.
Ethnographic work in the Paradijon community, southern Luzon, the
Philippines, also emphasizes the social and political factors that influence ceramic
raw material procurement. Neupert (2000: 251) argues that in Paradijon the
particular resources used by potters are determined largely by social and political
constraints. The Paradijon community is characterized by dichotomous political
factionalism. Access to a particular ceramic resource is in exchange for allegiance
to one of the two competing political factions.
According to Neupert (2000: 253-255), although both men and women
participate in Paradijon ceramic production, each gender is typically responsible
for different stages in the production process. In Paradijon, male relatives collect
clay and fire vessels shaped by their female kin. Women without suitable male
relatives may ask male relatives of friends, or a number of men that collect clay for
money, to fetch their clay. Women with no suitable male relatives, friends, nor
cash may lack access to clay resources. One such case was documented in
Paradijon. Although very poor, this potter was known as the most-skilled in the
community and worked only for other potters. She also had no allegiance to either
of the political factions in the community and worked for potters on both sides of
the political schism (Neupert 2000: 254).

The mode of ceramic production also influences resource procurement.
Costin (1991) has developed a typological system for the mode of production
based on the variables of context, concentration, scale (composition), and
intensity. Due to its detail, this is perhaps the best typology for understanding the
relationship between mode of production and access to ceramic raw materials.
Although this typology is designed specifically for specialized production, other
authors (e.g., Hegmon et al. 1995) have found it useful for describing
unspeciahzed household production as well.
Context refers to the affiliation of producers and the sociopolitical
component of the demand for their wares (Costin 1991: 11). Context is described
as either attached or independent, using Earles (1981) categories. A prerequisite
for attached specialization is the presence of elite(s), but independent specialists
may produce in a wider variety of political circumstances. Context identifies the
source of demand and determines, to some degree, what products are produced.
Assuming independent rather than attached production in the prehistoric
Southwest, Hegmon et al. (1995) disregard the context variable in their adaptation
of Costins typology.
Concentration refers to the geographic distribution of production,
considering the spatial relationship between producers and between producers and
consumers (Costin 1991: 13). Concentration of production is described as either

nucleated or dispersed. The distribution of raw materials has perhaps the greatest
influence on concentration. Transportation costs also have a considerable
influence on the concentration of production, as goods with high transportation
costs will tend to be made by dispersed producers. In addition, informal
cooperation among producers, as well as the presence or absence of a formal
marketplace will influence the concentration of production.
The composition of the production unit (Costin 1991: 15) is referred to
as scale. Scale is determined by both the number of people that comprise the
production unit and the method in which labor is recruited. Labor recruitment
occurs on a continuum with kin-based recruitment at one end and wage labor at the
other. Scale is not necessarily associated with demand or context. Scale is largely
determined by efficiency in situations of independent production (Costin 1991: 15-
16), at least in cases where the goods produced are subject to the economy of scale
(e.g., ceramics). Because of a considerable array of intervening variables
(occupation span, sediment accumulation rates, breakage rates, etc.), scale of
production at particular locales, operationalized as volume of goods produced, is
difficult to establish from the archaeological record (Pool 1992: 306-307)
Intensity refers to the amount of time producers spend on their craft
(Costin 1991: 16-18) and is described generally as either full-time or part-time.
Obviously, intensification involves an increasing degree of exclusion from other

activities. There are three economic factors that influence intensity of production:
efficiency (output to input ratio; Pool 1992: 278), risk, and scheduling. Efficiency
provides independent producers with a competitive advantage. Therefore, if
intensification allows for more efficient production, producers will favor the
process. For example, if the production of a particular product requires a
significant degree of skill or training it is more efficient to have fewer full-time
producers than a greater number of part-time producers. Whether or not increased
efficiency results from intensification also depends on the product. If production
involves multiple steps that can be accomplished simultaneously, intensification
will generally lead to increased efficiency.
Risk also influences the intensity of production because more time spent in
craft production equals less time devoted to activities directly related to
subsistence. In order to reduce subsistence risk, independent producers may favor
part-time production so that they may devote time not only to craft production but
subsistence production as well. Attached specialists, on the other hand, are
supported by wealthy patrons and therefore do not encounter the same sort of risk.
It has been suggested (e.g., Costin 1991: 17) that full-time specialization typically
occurs when the competitive advantage of craft intensification outweighs the risks
of not participating directly in subsistence production. This might be the case

when production technology is expensive and/or when the product is subject to the
economy of scale (i.e., the more you make the cheaper it is).
At a theoretical level, each of the variables considered in Costin's typology
of the mode of production can be related to the decisions made by potters when
acquiring raw materials. If producers function in an attached context, for example,
they could be guaranteed access to particular raw materials by the elites for whom
they produce. If the concentration of producers is great and ceramic resources are
limited, a greater degree of resource competition would be expected than if
concentration were low and resources abundant. Furthermore, if the scale of
production is at the individual or household level with kin-based labor, each
producing unit would be expected to have a similar degree of leverage when it
comes to making resource claims. However, if production groups are larger, it
would be expected that each individual in a group would share access to resources,
and would have a greater ability to exclude non-group members from those
resources. Finally, intensity can also be expected to influence ceramic resource
procurement. Full-time specialists may be able to easily maintain resource rights
if they are providing a service that is in high demand. Or if a producer is a full-
time craft specialist because he/she has limited access to subsistence resources,
he/she may have limited access to ceramic materials for the same reasons that
cause inadequate access to subsistence resources (e.g., immigrant status,

inadequate kin relations, etc.) Of course, the articulation of these variables with
access to ceramic resources is influenced largely by the geographic distribution of
those resources.
Research suggests that the mode of ceramic production in the northern
Southwest varied through time and from place to place. Hegmon et al. (1995: 33)
consider three possible modes of production in the central Mesa Verde region.
The most basic mode of production considered is unspecialized household
production for household consumption. In addition, Hegmon et al. suggest the
mode of production may have been characterized by dispersed individual
specialization, in which a few individuals or households makk pottery for an entire
community (1995: 33). Alternatively, the mode of production may have been one
of community specialization, in which production specialists aggregate in a limited
number of communities and produce for an entire region.
The production organization of white, gray, and red wares differed in the
central Mesa Verde region (Wilson and Blinman 1995). Although ideological
factors should also be considered, Wilson and Blinman suggest that vessel
function and the properties of raw materials largely explain the differences in
production organization of the different wares. In addition to the production
differences between the wares, each ware demonstrates changes in the
organization of their production through time (Wilson and Blinman 1995: 70-76).

To summarize the observations of Wilson and Blinman, in the
Basketmaker m period (A.D. 575-725), whiteware production seems to have been
specialized and geographically restricted. There is insignificant data regarding
white-ware production during the Pueblo I period (A.D. 725-900). During Pueblo
II (A.D. 900-1125) there was an increase in the percentage of white wares in total
assemblages as well as increases in technical standardization, stylistic diversity,
and the movement of white-ware vessels over long distances. Wilson and Blinman
(1995: 76) suggest that the increase in white-ware production in Pueblo II may be
at least partially explained by unstable red ware production and exchange. During
Pueblo III (A.D. 1125-13000), there seems to have been a continuing increase in
the specialization and scale of white ware production, as evidenced by large trench
kilns associated with considerable amounts of white ware and increased
production technology. Glowacki (1995) suggests that some white wares
continued to be produced at all sites in the Mesa Verde area throughout Pueblo III.
The production of gray ware experienced less significant diachronic
changes in organization (Wilson and Blinman 1995: 70-77). The evidence
suggests that many households produced gray wares during Basketmaker III,
although there does appear to have been a low-level of production specialization.
In Pueblo I, there was technical improvement in gray ware production and a
concomitant decrease in the already low level (compared with white ware) of

specialization. There is a lack of evidence for non-producing households during
Pueblo I. Decreased specialization and continuation of household level production
and consumption of gray ware continued throughout the Pueblo II and Pueblo III
periods. The lack of evidence for production workshops or large kilns prior to
A.D. 1150 suggests that gray and white ware pottery production appears to fall
within the range of household production or household industries (Blinman and
Wilson 1992: 161). In comparison to other wares, gray ware has received much
less attention in the literature and there is a need for specific consideration of these
Anthropologists have proposed various models of exchange to explain and
describe the nature of this socioeconomic interaction in the Southwest. These
include the pochteca model (DiPeso 1974), prestige goods economy (McGuire
1986, Bradley 1996), peer polity interaction (Renfrew and Cherry 1986), and
World Systems Theory (Wallerstein 1974). The applicability of these models to
various regional and interregional systems of exchange in the prehistoric
Southwest is debatable (see Hegmon [ed.] 2000). These models are, for the most
part, appropriate to exchange that occurs on a larger geographic and social scale

than there is archaeological evidence for in the central Mesa Verde region during
the late Pueblo in period (Lipe and Varien 1999: 337).
It is argued here and elsewhere (Kohler et al. 2000), that in the context of
this study, exchange is most appropriately considered in terms of reciprocity.
Reciprocity has received explicit attention from economic anthropologists (e.g.,
Polanyi 1957; Sahlins 1972), cultural anthropologists (e.g., Carrier and Carrier
1989; Malinowski 1922), and archaeologists alike (e.g., Kohler et al. 2000). A
brief discussion of reciprocity and its implications for the current study follows.
Reciprocal exchange is very diverse both within and between cultures.
Similarly, anthropologists have provided a considerable number of definitions and
typologies of reciprocal exchange that span the substantivist/formalist divide (see
Wilk 1996 for a summary of the substantivist/formalist debate; see Kohler et al.
2000, Pryor 1977, Sahlins 1972 for a wide range of viewpoints on reciprocal
exchange). Influenced by the work of Sahlins (1972: 185-230), reciprocal
exchange, as used herein, simply refers to the vice versa movement of material
and/or nonmaterial goods. Although dominant in small-scale societies, reciprocal
exchange occurs in all socioeconomic systems.
Sahlins (1972: 193-196) develops a heuristic typology for describing the
range of reciprocal interactions. At the positive end of this continuum is what
Sahlins refers to as generalized reciprocity, "the solidary extreme" (Sahlins 1972:

193). This type of exchange is considered to be fairly altruistic and there is no
guarantee that the initial receiver in the transaction will make a return. At the
negative end of this continuum is negative reciprocity, the "unsociable extreme"
(Sahlins 1972: 195). In this type of exchange, the parties involved have competing
interests and each tries to make some sort of gain at the expense of the other.
Balanced reciprocity is the "midpoint" (Sahlins 1972: 194) on this continuum. It
is characterized by the simultaneous exchange of equally valued goods.
According to Sahlins (1972: 204-205), the type of reciprocal relations in which
individuals and social groups are involved is largely a function of the social
distance between the parties involved, as well as, to lesser extents, differences in
social rank, relative wealth and need, and the types of goods exchanged.
Reciprocity between kin tends to occur at the positive end of the spectrum, and
between strangers or enemies it tends to occur at the negative end of the spectrum.
Reciprocal exchange tends to be balanced more often in theory than in practice
(Pryor 1977).
Although exchange, especially non-market prehistoric exchange, is
commonly considered to occur primarily for the acquisition of material goods, it is
important to emphasize the aspects of reciprocity that are not formally economic.
Reciprocal exchange is a mechanism for initiating and perpetuating social relations
(LeBlanc 1999: 55). It is a way to form social, political, and economic alliances.

Furthermore, reciprocal exchange creates social obligations and perceptions of
indebtedness between those involved. It can be further argued that such relations
of indebtedness can create the ability to appropriate the labor of others and thereby
increase individual or group power (see Saitta 1994). Considered in this light,
reciprocal exchange is as important politically and socially as it is economically.
Economies based on reciprocal exchange have been characterized as
comparatively (i.e., when compared to market economies) rigid systems (Kohler
1992). Individuals are not guaranteed relations of reciprocal exchange, and such
relations are often historically contingent. Furthermore, one's position or degree of
participation in a system of reciprocal exchange is influenced considerably by
her/his kin relations. Consequently, individuals that lack appropriate social
relations are likely to also lack relationships of reciprocal exchange, as well as its
related social, political, and economic consequences. Of course, the notion should
not be ignored that individuals or social groups can actively create such social
It was mentioned above that exchange must be considered in this thesis
because it is a factor that may have contributed to the archaeological assemblages
analyzed herein. From this brief discussion of reciprocal exchange, however, one
can also see a number of ways in which exchange relates to the question of
ceramic resource access. For example, individuals or households that have many

local kin relations may participate to a greater degree in a system of reciprocal
exchange than other individuals or households without such relations, such as
might be expected of small immigrant populations. Also, individuals or groups
that are involved in a greater number of reciprocal transactions should be
characterized by a greater diversity of ceramic resources if these resources, either
finished vessels or raw materials, were subject to reciprocal exchange. Exchange
relations can be considered a resource, like raw materials, with differential rights
of use and access between individuals.
Furthermore, an individual's relationship to the production of ceramic
vessels and the acquisition of ceramic raw materials is also expected to influence
her/his likelihood in being involved in the exchange of these items. A non-
producing household, for example, may be characterized by greater ceramic
resource diversity in the archaeological record if they exchange other goods for
ceramics manufactured by multiple households that use different raw materials. A
household that specializes in ceramic production may be expected to have a much
lower diversity of ceramic resources.
It is appropriate to conclude this chapter with a few words on the likely
importance of ceramics in the prehistory of the central Mesa Verde region. The

There are several goals of this chapter. First, it is hoped that the
ethnographic cases described above adequately illustrate the cross-cultural
diversity of land tenure and the social negotiation of resource access. Secondly, it
is the author's intent to convince the reader that prehistoric southwestemers
actively participated in the negotiation of resource use and access through a variety
of mechanisms. Finally, and most importantly, it is the author's goal to also
convince the reader that ceramic resource use and access in the late Pueblo HI
period of the central Mesa Verde region was socially and historically contingent.
It should not be assumed that such behavior operates according to formal.
economic expectations.

This chapter summarizes background information of the central Mesa
Verde region in order to provide a context for the research presented herein. First
are brief descriptions of the geology, environment, and paleoenvironment in the
region. This is followed by a synopsis of the culture history of the region from the
Paleoindian through Pueblo m periods. Although the central Mesa Verde region
experienced prehistoric occupation after the Ancestral Puebloan migration from
the area, this body of research is considered beyond the scope of this project.
Interested readers are referred to Wilshusen and Towner (1999: 353-369) for a
summary of post-Puebloan studies.
Much of the information presented in this chapter is derived from the 1999
published context for the Southern Colorado River Basin (Lipe et al. 1999; Adams
and Petersen 1999; Lipe 1999; Lipe and Pitblado 1999; Wilshusen 1999a,b; Lipe
and Varien 1999a,b), an area that overlaps the central Mesa Verde region. The
context largely describes the archaeology and environment specifically in the
Southern Colorado River Basin, but this information can, for the most part, be
accurately extrapolated to the central Mesa Verde region. The amount of
archaeological research that has been done in the central Mesa Verde region is

enormous. Rather than attempt to summarize this research here, the goal of this
chapter is to simply provide an environmental and historical framework in which
the reader can consider the research that is the focus of this thesis.
As noted in the previous chapter, methods of ceramic compositional
analysis are used to address the research questions presented in this thesis. Since
ceramic raw materials have geologic origins, the local geology of the central Mesa
Verde region is especially relevant to this thesis. This section begins with a
general description of the geology and geography of the northern Southwest,
followed with more detailed descriptions of the local sedimentary and igneous
The central Mesa Verde region is located on the Colorado Plateau. This
large geologic province covers large portions of the states of Arizona, Colorado,
New Mexico, and Utah. The Colorado River and its tributaries drain the province.
Major tributaries include the San Juan, Little Colorado, and Green Rivers.
Generally speaking, the Colorado Plateau is bounded by the Rio Grande
rift to the east, the Jemez fault zone to the southeast, the Texas-Walker Lane

structure of Arizona to the southwest, the Wasatch Line to the west, and the Uinta
Mountains to the north (Baars 2000: viii-ix). The province is an uplifted area,
separated from the rest of the North American continent by horizontal wrench
faults (Baars 1995: 1-2). The geology of the Colorado Plateau is characterized by
vast basins of sedimentary rock, separated topographically by step-forming
monoclines and faults. Across the province, these sedimentary formations are
punctured by both intrusive and extrusive igneous formations.
The geology of the central Mesa Verde region is not atypical for the
Colorado Plateau, with mountain forming igneous formations and thick
sedimentary deposits cut by deep canyons. The sedimentary geology of the area is
relevant to this project because it provided most of the clay resources for
prehistoric pottery production. The oldest exposed sedimentary formation in the
central Mesa Verde region is the Navajo Sandstone, which may date to either the
Triassic1 (240-200 million years ago [m.y.a.]) or Jurassic (200-135 m.y.a.) period
(Ekren and Houser 1965: 7-8). Above the Navajo Sandstone is the San Rafael
Group, which is comprised of the Entrada Sandstone, Summerville Formation, and
Junction Creek Sandstone from oldest to youngest. Above the San Rafael Group
is the Morrison Formation, which is comprised of the Salt Wash Sandstone
Member, the Recapture Shale Member, the Westwater Canyon Sandstone
Member, and the Brushy Basin Shale Member, again from oldest to youngest.

Both the San Rafael Group and the Morrison Formation are Upper Jurassic in age
(Ekren and Houser 1965: 6-18). The Morrison formation provided a prehistoric
source of raw materials for lithic and ceramic production (Glowacki et al. 1998;
Pierce et al. 1999). The layers of silicified volcanic ash in the Morrison formation
(Gerhardt 2001) are the most likely source of the tuffaceous temper discussed in
the following chapters.
The aforementioned Jurassic formations are overlain by more recent
Cretaceous (135-65 m.y.a.) deposits (Ekren and Houser 1965: 18-26). Lower
Cretaceous formations include the Karla Kay Conglomerate and the overlying
Burro Canyon formation. Upper Cretaceous deposits include the Dakota
Sandstone, Mancos Shale, and the Mesa Verde Group, from oldest to youngest.
The Dakota Sandstone also provided raw materials for lithic production (Pierce et
al. 1999). The Dakota Sandstone and the Mancos Shale are sources of clays that
are suitable for ceramic production (Glowacki et al. 1998). More recent
Quartemary (3 m.y.a. to present) deposits are limited to alluvial, colluvial and
eolian deposits (Ekren and Houser 1965: 26).
As in much of the Colorado Plateau, more recent igneous rocks have
protruded through these sedimentary deposits. Several laccolithic intrusions form
prominent features on the landscape of the Mesa Verde region. These features
include the Ute Mountains (also known as Sleeping Ute Mountain), the Abajo 1
1 Dates for geologic periods are from Chronic (1998: xii).

Mountains, and the La Platas. These mountains are roughly contemporaneous
(formed ca. 65 million years ago), similar structurally and compositionally, and
internally heterogeneous (Ekren and Houser 1965: 27; Hegmon 1995: 375). In the
central Mesa Verde region, igneous rock was a common temper material used in
gray ware ceramics.
In particular, the igneous geology of the Ute Mountains is most relevant to
the current study. Due to its geographic proximity, this mountain range is the most
likely source of igneous temper to have been used by the inhabitants of the four
sites that provide the case studies for this research. Igneous temper may have been
procured directly from the mountain or from igneous cobbles found in McElmo
Creek and its tributaries.
Ekren and Houser (1965) provide the most detailed description of the
geology of the Ute Mountains. Ekren and Houser (1965: 28-33) describe five
classes of igneous rocks that constitute the range. These are microgabbro, diorite
porphyry, granodiorite porphyry, quartz monzonite porphyry, and lamprophyre.
Ekren and Houser's descriptions of these rocks are briefly summarized below.
Diorite porphyry is the most common rock found in the Ute Mountains
(Ekren and Houser 1965: 28-30). This broad type can be divided into two
varieties, which Ekren and Houser refer to as normal diorite porphyry and
leucocratic diorite porphyry. These two varieties are distinguished primarily by

their color, which is a function of the proportion of mafic (i.e., dark-colored) to
felsic (i.e., light colored) minerals in the groundmass of the rock.
Normal diorite porphyry contains a greater proportion of mafic minerals in
the groundmass, and is thus darker in color. This groundmass of normal diorite
porphyry is made up of plagioclase, anhedral quartz, and potassium feldspar, with
smaller amounts of hornblende, augite, and chlorite. Phenocrysts and groundmass
occur in approximately equal proportions. The most common phenocrysts are
anhedral to euhedral plagioclase crystals, which average 2mm in length and make
up 25 to 45 percent of the rock's volume. Other common phenocrysts include
euhedral crystals of augite and hornblende. Combined, augite and hornblende
phenocrysts make up about 14 percent of the volume of normal diorite porphyry.
Approximately equal proportions of groundmass and phenocrysts also
characterize leucocratic diorite porphyry. However, this rock contains no augite or
other pyroxenes, and contains more hornblende than the normal diorite porphyry.
The groundmass of leucocratic diorite porphyry consists of plagioclase, potassium
feldspar, and quartz, and is characterized as dense to extremely dense.
Phenocrysts are of palgioclase (ca. 35 percent of the total rock volume) and
hornblende (ca. 10-20 percent). In hand-specimen, leucocratic diorite porphyry is
similar to granodiorite porphyry but can be distinguished microscopically by its
lack of pyroxene and lower potassium feldspar content.

According to Ekren and Houser (1965: 28), two varieties of microgabbro
occur in the Ute Mountains. Both varieties are black, aphanitic (i.e., so fine
grained that constituent minerals can not be identified with the naked eye) rocks
that are similar in appearance to basalt. The primary difference between the two
varieties of microgabbro is in their texture. One variety is seriate in texture, with
most mineral crystals not exceeding 1 mm. in length. The largest minerals in this
variety are augite (ca. 15 percent) and hornblende (ca. 8 percent), which range in
size from .1 mm to 3mm. This variety also contains plagioclase minerals that
range in size from .1 mm to .6 mm.
The second variety of microgabbro is porphyritic in texture. It contains
phenocrysts of hornblende (ca. 10 percent) and augite (ca. 3-10 percent) that range
in length from 2 to 10 mm. These phenocrysts occur in a groundmass composed
primarily of plagioclase and chlorite. Quartz is ubiquitous, comprising
approximately 4 percent of the rock. Biotite also occurs, although less frequently.
Common accessory minerals include apatite and magnetite. Both varieties show
signs of considerable alteration of parent minerals.
Ekren.and Houser (1965: 30-31) recognize three spatially distinct (as they
occur as bedrock) varieties of granodiorite porphyry in the Ute Mountains. They
refer to these varieties as the southern, northern, and central types, according to
where they are found in the mountain range. The southern variety is characterized

by approximately 50 percent extremely dense groundmass rich in potassium
feldspar, equidimensional phenocrysts of zoned plagioclase as large as 10 mm.,
and phenocryst pseudomorphs after hornblende that have altered to sericite,
chlorite, and calcite on the outward crystal portions. Epidote and garnet are
common in the southern variety.
The central variety of granodiorite porphyry is similar to the southern
variety in that plagioclase and pseudomorphs after hornblende are the most
common phenocrysts. However, the plagioclase phenocrysts in the central variety
are smaller and do not exceed 4 mm. in length. In addition, the central variety of
granodiorite porphyry contains abundant fine grained quartz crystals in the
groundmass as much as .5 mm. in length. The plagioclase phenocrysts of both the
southern and central types show signs of considerable alteration to sericite, calcite,
and chlorite.
The northern variety is referred to as hornblende gradodiorite porphyry.
Hornblende and plagioclase phenocrysts of this variety are less altered than in the
southern and central varieties. Other phenocrysts include augite, magnetite, and
less frequently rounded zircon crystals. The groundmass is rich in fine grained
potassium feldspar, plagioclase, and quartz.
Likewise, Ekren and Houser (1965: 31-32) recognize two geographically
distinct varieties of quartz monzonite porphyry in the Ute Mountains, a northern

variety and a southern variety. Both varieties are characterized by an extremely
dense groundmass of potassium feldspar, and possibly quartz and albite. The
northern variety is characterized by equal proportions of groundmass and
phenocrysts. Phenocrysts are of rounded bipyramidal quartz (ca. 10 percent),
zoned and altered plagioclase (ca. 30-35 percent), and altered biotite.
The southern variety of quartz monzonite porphyry is more similar in
texture to the southern variety of granodiorite porphyry, with large phenocrysts of
plagioclase (ca. 35 percent) and pseudomorphs after hornblende (ca. 3 percent).
However, the southern variety of quartz monzonite porphyry is distinguishable by
its greater abundance of quartz phenocrysts (ca. 4 percent). Smaller phenocrysts
of epidote and apatite are also common.
The final type of igneous rock found in the Ute Mountains is spessartite
(Ekren and Houser 1965: 32-33). Spessartite is a type of rock referred to as a
lamprophyric. Such rocks are generally characterized by a porphyritic texture, as
well as by a groundmass of both light and dark minerals and phenocrysts of dark
minerals (Bates and Jackson 1984: 286). Again, two varieties of spessartite can be
distinguished within the Ute Mountains, a western variety and a variety found at
Black Mountain (a prominent peak in the Ute Mountains). The western variety
spessartite is conspicuously porphyritic, with a groundmass of predominately
plagioclase and chlorite, and large (commonly >10 cm. in length) phenocrysts of

augite and hornblende. Thomsonite, magnetite, and apatite are also present in the
western spessartite.
Black Mountain spessartite is less porphyritic and almost seriate in texture.
A seriate texture is for the most part defied by the presence of large hornblende
phenocrysts, which can measure up to 30 mm. in length. Smaller phenocrysts
include relatively unaltered plagioclase (0.2 -1.5 mm. in length), euhedral augite
(ca. 2 mm. in length), and biotite (0.1 4.0 + mm.). Radiating groups and sheaves
of chlorite with actinolite (0.1 0.3 mm. in diameter) also occur, in addition to
magnetite and pyrite as accessory minerals. Although both spessartite varieties are
similar to micrograbbo in hand-specimen, the spessartites are recognizable by a
lack of quartz and the presence of large hornblende and/or augite phenocrysts.
Modem Environment
This section of the thesis describes the climate and biotic communities of
the central Mesa Verde region, which is discussed here as a sub-region of the
Southern Colorado River Basin. This information is not considered to be directly
relevant to the current study. It is nevertheless included in this chapter in order to
give the reader an adequate description of the modem environment, which is much
like the environment that characterized most of the Archaic period, as well as the
Basketmaker and Pueblo periods (see below).

The high mesas of the central Mesa Verde region are dissected by
numerous canyons and arroyos, creating an area of considerable topographic relief.
Like the rest of the American Southwest the region is arid, with the amount of
water lost to evaporation typically exceeding that gained by precipitation
(VanWest and Dean 2000:20). However, there is a positive correlation between
elevation and precipitation. Precipitation follows a bimodal pattern (Cordell 1997:
37) with most moisture coming in the form of winter snow and summer
thunderstorms, or monsoons. Due to its latitude and location on the elevated
Colorado Plateau, the central Mesa Verde region is relatively cool and moist in
comparison to the rest of the Southwest.
Seven biotic communities exist in the central Mesa Verde region (Adams
and Petersen 1999: 14-18). These biotic communities are characterized by suites
of plants and animals, and vary according to elevation. Following is a description
of these biotic communities, with special attention given to plant and animal
resources that have potential for human exploitation. The descriptions are
summarized from Adams and Petersen (1999).
The saltbrush-sagebrush community is typically found in elevations
between 1200 and 2200 m. This ecological zone is dominated by big sagebrush
(Arteminsia tridentata) but also includes considerable proportions other varieties
of sage (Arteminsia spp), saltbrush (Atriplex) rabbitbrush (Chrysothamnus), and

winterfat (Ceratoides). Cactus species include cholla, prickly pear (Opuntia spp.)
and hedgehog (Echinerocereus). Several grasses are also present including
ricegrass (Stipa), grama (Bouteloua), dropseed (Sporobolus), fendlergrass (Poa),
and galleta (Hilaria). Animal species are dominated by small mammals, coyotes
(Canis), and sparse quantities of antelope (.Antilocarpa) and desert bighorn (Ovis).
The saltbrush-sagebrush community occupies roughly 19.6 percent of the Southern
Colorado River Basin (Adams and Petersen 1999: 14)
The Grassland biotic community also occupies lower elevations, 1200 -
2200 m, in areas more susceptible to range fires (Adams and Petersen 1999: 15).
This zone has been considerably altered and reduced by historic land-use practices.
Grasslands typically grade into saltbrush-sagebrush at lower elevations, and into
pinyon-juniper woodland or Ponderosa pine (Pinus ponderosa) forest at upper
elevations. Grasses include grama, ricegrass, junegrass (Koeleria), and dropseed,
many of the same grasses found with lesser frequency in the saltbrush-sagebrush
community. Shrub species include rabbitbrush, yucca (Yucca), sumac (Rhus), wild
rose (Rosa), winterfat (Ceratoides), as well as sagebrush, and saltbrush. Animals
include a wide variety of small mammals, birds, and antelope. The grassland
community occupies about 3 percent of the Southern Colorado River Basin
(Adams and Petersen 1999: 16).

The pinyon-juniper biotic community typically occupies elevations
between 1500 and 2300 m. This is the spatially most extensive biotic community
of the central Mesa Verde region. It occupies roughly 33.8 percent of the Southern
Colorado River Basin (Adams and Petersen 1999: 16). Plants are dominated by
pinyon (Pinus edulis) and several varieties of juniper {Juniperus scopolorum, J.
osteosperma, and J. monosperma). Shrubby species include sagebrush, Gambel
oak (Quercus gambelli), rabbitbrush, skunkbrush (Rhus aromatica), saltbrush,
mountain mahogany (Cercocarpus), serviceberry (Amelanchier), and yucca.
Prickly pear and hedgehog cacti are present. The pinyon mouse (Peromyscus) is
the dominant animal species, and the pinyon-juniper woodland also provides
winter range for elk and mule deer (Adams and Petersen 1999). In terms of
subsistence and material culture resources, the pinyon-juniper woodland is
arguably the richest of all the biotic communities of the central Mesa Verde
region. All four of the sites that are the focus of this research are located in
pinyon-juniper woodland.
Continuing upward in elevation, the gambel oak scrubland biotic
community is generally found on south-facing slopes between 2300 and 2750 m
(Adams and Petersen 1999: 17). Gambel oak dominates the vegetation of this
biotic community. However, many shrub species are also found here, including
mountain mahogany, sumac, serviceberry, wild rose, hackberry (Celtis),

chokecherry (Primus), currant (Ribes), and elderberry (Sambucus). This
vegetation provides habitat for mule deer (Odocoileus hemionus), as well as wild
turkey (Meleagris gallopovo) in the winter and fall. Approximately 9.9 percent of
the Southern Colorado River Basin is covered by gambel oak scrubland (Adams
and Petersen 1999: 17).
The pine-douglas-fir forest comprises roughly 23.9 percent of the Southern
Colorado River Basin, and occurs from 2000 to 3050 m (Adams and Petersen
1999: 17). The vegetation of this zone is dominated by pines (Pinus spp.), douglas
fir (Pseudotsuga), and less frequently true firs (Abies spp.). Stands of aspen
(Populus tremuloides) are common at higher elevations and on shady, north-facing
slopes. Parks of old growth Ponderosa pine are common at lower elevations where
the pine-douglas-fir forest typically grades into pinyon-juniper woodland or
grassland. These dominant plants are interspersed with gambel oak, diverse
grasses (Gramineae), shrubs (Ribes, Rosa, Sambucus), and additional coniferous
species. Mule deer, elk (Cervus llaphus), gray wolf (Canis lupus), wild turkey,
and a wide variety of smaller mammalian species are among the animals that
occupy the pine-douglas-fir forest (Adams and Petersen 1999: 17).
The spruce-fir forest biotic community occupies elevations from treeline
(3500-3800 m) down to approximately 2450 m. This biotic community is found in
roughly 8.4 percent of the Southern Colorado River Basin (Adams and Petersen

1999:17). Common plants include limber, bristlecone, and ponderosa pines
(Pinus spp.), aspen, willow (Salix), alder (Alnus), maple (Acer), fruit-bearing
shrubs (Sambucus, Ribes, Rubus), dwarf juniper (Juniperus communis), and
diverse smaller herbaceous species. Mule deer, elk, and a variety of birds are
common animals (Adams and Petersen 1999: 17-18).
Alpine tundra is relatively rare in the Southern Colorado River Basin,
located only in the highest elevations of the La Plata and San Juan Mountains
above 3500 m (Adams and Petersen 1999: 18). There is no alpine tundra in the
central Mesa Verde region, although its prehistoric inhabitants likely had access to
the resources found there. Only the hardiest species survive in this community.
Plant species include lichens, mosses, small woody shrubs, and herbaceous plants.
Animals include many small to medium sized mammals. Bighorn sheep (Ovis) is
the most common large mammal.
Large tracts of the central Mesa Verde region are today dedicated to
agricultural production. Much of this farming involves modem irrigation systems,
but dry-farming is still commonly practiced. Although other factors play a role
(e.g., soil type, VanWest and Dean 2000: 21), dry-farming success in this region is
dictated largely by two variables: precipitation and number of frost-free days.
Historically, lands between 6000 and 7100 m have been characterized by a

sufficient degree of both variables to be most successful in dryland agricultural
production (Adams and Petersen 1999: 26).
A considerable amount of research has focused on reconstruction of the
prehistoric environment in the central Mesa Verde region. These environmental
reconstructions are based on proxy evidence from pollen data, tree-rings,
macrobotanical and faunal remains, and geomorphological processes (Adams and
Petersen 1999: 34; Cordell 1997: 50-63). Information gained from
paleoenvironmental studies is frequently used as a component of explanatory
models (see Adams and Petersen 1999: 34 for a list of such studies) or to provide a
contextual framework for research.
The paleoenvironment of the central Mesa Verde region is best
characterized as temporally variable. Proxy data, such as those from pollen
records and packrat middens, show large, low-frequency shifts in climate and
vegetation through time. Other proxies, such as tree-rings, show high-frequency
climatic events such as drought (Adams and Petersen 1999: 36; Dean 1988, 1996).
These changes are related to large-scale atmospheric and geologic phenomena
(Adams and Petersen 1999: 34). Central Mesa Verde paleoenvironment is very
briefly summarized here to provide an environmental context for the culture

history of the region presented below. The reader is referred to Adams and
Petersen (1999: 34-50) for references to more detailed sources.
The paleoclimate of the late Pleistocene (16,000-11,000 B.C.) is
characterized as considerably colder and wetter than today's climate in the central
Mesa Verde region. Waugh and Petersen (1995; cited in Adams and Petersen
1999: 44) suggest an average annual temperature of only 2C and an average
annual precipitation of 80 cm at Monticello, Utah, just north of the region. During
this time, subalpine forests occupied lands that are now covered with pinyon-
juniper woodland. There was no warm season, and consequently no summer
monsoon (Adams and Petersen 1999: 44).
The post-glaciation climate (11,000-8,500 B.C.) was both wetter and
warmer (Adams and Petersen 1999: 44-49). The warming trend continued into the
early Holocene (8,000-6,000 B.C.), but the climate also became drier. The
middle Holocene, or Altithermal (6,000-4,000 B.C.), also saw increasing
temperatures and precipitation. During the middle Holocene winters were dry and
cold however, with most precipitation coming in the form of summer monsoons.
Average annual temperature was approximately 10C, with an average annual
precipitation of roughly 60 cm in Monticello (Waugh and Petersen 1995; cited in
Adams and Petersen 1999: 49). Pinyon-juniper woodland appears to have first

expanded into the central Mesa Verde region after 2,500 B.C. (VanDevender
1984; VanDevender et al. 1984; cited in Adams and Petersen 1999: 49).
After the middle Holocene, the paleoclimatic record of the central Mesa
Verde region becomes more complex (Adams and Petersen 1999: 49). Two low-
frequency, large-scale climatic events are of particular relevance to the current
study. These are the Medieval Warm period (A.D. 800-1200) and the Little Ice
Age (A.D. 1250-1850). The Medieval Warm period is characterized by an
increase in summer precipitation (Adams and Petersen 1999: 41), creating
favorable conditions for agricultural production. The Little Ice Age is
characterized by cooler summer temperatures, decreased strength of the summer
monsoons, southern retreat of the monsoon boundary, and arrival of the monsoon
rains later in the summer (Adams and Petersen 1999: 41). Adams and Petersen
(1999) imply that the Little Ice Age would have made the central Mesa Verde
region an uninhabitable place for agriculturalists. However, other research (Van
West 1994; Van West and Dean 2000) suggests that climatic conditions in the
central Mesa Verde region were never severe enough to provide the sole stimulus
for the Puebloan migration from the region.
High-resolution paleoclimatic reconstruction recently has been published
(VanWest and Dean 2000) for the central Mesa Verde region between A.D. 900-
1300. This reconstruction uses tree-ring data to determine total tree-year

precipitation, mean tree-year temperature, and Palmer Drought Severity Index
(PDSI). This research shows that most dry periods during this time period were
not long in duration, most lasting only one year but occasionally lasting two or
three years (VanWest and Dean 2000: 23). Nevertheless, it also shows that there
were ten episodes between A.D. 900-1300 in which precipitation shortfalls
occurred for at least five consecutive years. The drought that occurred between
1130-1180 was the most severe, lasting 50 years with virtually no relief provided
by years of average or above-average precipitation. What is known as the Great
Drought lasted from 1273-1300 and is second in severity after the 1130-1180
drought (VanWest and Dean 2000: 26). The degree to which the Great Drought
influenced the Puebloan migration from the central Mesa Verde region has been
the subject of recent debate (compare Adams and Petersen 1999 and Van West and
Dean 2000).
Culture History
The remainder of this chapter describes the culture history of the central
Mesa Verde region, with some discussion of relevant research in adjacent areas.
As with the environmental information presented above, this section provides a
context for the research presented herein. Additionally, however, the culture
history of the region is considered to be important because the theoretical approach

taken in this thesis explicitly considers history to be a significant factor in social
life. This thesis deals primarily with the issue of resource access, which was noted
in the previous chapter to be historically contingent. Furthermore, this section
demonstrates the socially, demographically, and politically dynamic character of
the regions prehistory.
Paleoindian (pre-5500 B.C.)
Evidence of Paleoindian occupation in the central Mesa Verde region is
sparse. Fluted Folsom and Clovis points are characteristic of the earlier portion of
the Paleoindian period. Only a handful of sites with Folsom or Clovis components
have been recorded in the central Mesa Verde region (Lipe and Pitblado 1999:
Several point type complexes characterize the latter portion of the
Paleoindian period in the region (Lipe and Pitblado 1999:101-104). The most
common point types are Angostura and Great Basin Stemmed. In a study of a
large portion of southwest Colorado, Pitblado (1999,1993; cited in Lipe and
Pitblado 1999: 97-104) notes 14 Angostura occurrences, nine Great Basin
Stemmed occurrences, and lower frequencies of Agate Basin, Plano complex,
Jimmy Allen/Frederick, Pryor stemmed, and possible Hell Gap and Eden/Firstview
components. Thus, human use of the central Mesa Verde region continues to be

sparse but seems to be considerably greater in the latter portion of the Paleoindian
Lipe and Pitblado (1999: 97) point out that little is known about the
behavioral adaptations (e.g., subsistence, mobility strategies, social organization)
of the Paleoindian inhabitants of the central Mesa Verde region. However, early
Paleoindians are generally considered highly mobile hunter-gatherers that
exploited now extinct Pleistocene megafauna, as well as smaller game, using a
variety of sophisticated hunting, butchering, and processing tools (Cordell 1997:
90-91). Later southwestern Paleoindians diverged into geographically
distinguishable behavioral patterns (Cordell 1997: 96). Paleoindians almost
certainly exploited a large variety of plant foods, although direct evidence (i.e.,
diagnostic plant processing tools or plant remains in dated contexts) is lacking
(Cordell 1997: 91).
Based on ethnographic analogy, both social group size and population
density are usually considered to have been quite low. However, Cordell (1997:
97-98) points out that modem hunter-gatherer analogs are circumscribed by
agricultural groups and have been largely restricted to unproductive lands. Thus it
may be inappropriate to assume that the social circumstances in Paleoindian North
America were similar to those found among today's hunter-gatherers.

Archaic (ca. 5500 B.C.- 500 B.C.f
Following Lipe and Pitblado, Archaic is the name given to the period
between the end of the Paleoindian period and the introduction of maize (Lipe and
Pitblado 1999: 106). Although various phase schemes have been proposed for the
Archaic that could apply to the central Mesa Verde region (Lipe and Pitblado
1999: 106-119), a lumping approach is taken here and the period is considered in
general terms. Published information on the Archaic in the central Mesa Verde
region is limited, and very few Archaic components have been excavated.
However, the Archaic has received considerable attention in adjacent areas. In
northwest New Mexico, for example, the development of fossil fuel and water
resources has resulted in the documentation of numerous Archaic sites (e.g.,
Vogler 1993a, 1993b).
A considerable portion of what is known about the Archaic specifically in
the central Mesa Verde region comes from the Ute Mountain Ute Irrigated Lands
Archaeological Project (UMUILAP). UMUDLAP intensively studied 16 probable
Archaic sites (Billman 1997; Lipe and Pitblado 1999: 122). Most of these are
spatially small sites with limited quantities of artifacts. The UMUILAP sites also
lack the features and structures characteristic of Archaic sites to the north of the
central Mesa Verde region (Reed and Metcalf 1999; Lipe and Pitblado 1999: 118-
119). Consequently, they have been interpreted as limited activity areas for

resource procurement and/or processing by small groups (Billman 1997; cited in
Lipe andPitblado 1999: 123).
A significant contribution of the UMUILAP relates to Archaic mobility
practices. In much of western Northern America, Archaic groups are considered
to have had very large multi-year ranges. For example, Vierra (1994) estimates
that Archaic groups in the San Juan Basin of northern New Mexico had a possible
maximum range of up to 39,800 km2 (cited in Lipe and Pitblado 1999: 107).
UMUILAP research found that most raw materials for tool manufacture came
from nearby sources, suggesting that the Archaic groups of the central Mesa Verde
region exploited much smaller territories than Archaic groups in other parts of the
Southwest (Lipe and Pitblado 1999: 123). Billman (1997) suggests that such a
small range was possible due to the great topographic relief in the UMUILAP area
and adjacent areas.
In general, the Archaic subsistence pattern is characterized by an increase
in diet-breath, a trend that appears to have started in the latter portion of the
Paleoindian period. Archaic groups exploited a wider variety of lower-ranked
food sources. This increase in diet-breadth is associated with an increase in the
diversity of ecological zones that are exploited for food as part of a seasonal round.
In the late Archaic and the Archaic/Basketmaker II transition there seems
to be considerable diversity in the acceptance of cultigens (Lipe and Pitblado

1999:123). The earliest cultigens on the Colorado Plateau date to ca. 1000-800
B.C. (Wills and Huckell 1994: 42). Unlike other subregions of the Southwest,
however, the earliest use of cultigens on the Colorado Plateau does not coincide
with an increase in either intensity of site occupation or the use of food storage
features. Wills and Huckell (1994: 50-51) assert that the initial acceptance of
cultigens by forager groups in the Southwest was associated with conditions of
high foraging productivity that allowed labor to be allocated to food production
activities. These authors further argue that cultigens provided a relatively
predictable resource and thus made for more efficient procurement of wild
resources. Although the spread of cultigens in the greater Southwest occurred
during the Archaic, they do not appear in the central Mesa Verde region until the
later Basketmaker II period.
Archaic mobility practices tend to mirror those of subsistence. Mobility
strategies cover the forager/collector continuum, with a logistical collector strategy
becoming more popular than in the previous Paleoindian period. However, there
is synchronous diversity among Archaic southwestern groups. For example,
Feinman et al. (2000: 456) point out that in the Late Archaic of southern Arizona,
large pithouse villages of 500 to 800 structures were occupied at the same time
other groups pursued a nomadic foraging strategy. According to these authors
(2000: 457) the cultural differentiation that led to the formation of groups across

the Southwest now recognized by archaeologists began at this time. Vierra and
Doleman (1994: 79) argue that Archaic groups in the San Juan Basin of northern
New Mexico followed a forager strategy from spring to fall and a collector
strategy during the winter season (sensu Binford 1980). Thus, there was also
seasonal variability in settlement and subsistence within Archaic groups.
Basketmaker II tea. 1000 B.C.-A.D. 500)
Basketmaker II is the first Basketmaker period in the generally accepted
chronology of the central Mesa Verde region. Historical circumstances in the
development of temporal schemes has led to the lack of a Basketmaker I period
(Lipe 1999: 132). Several authors (Lipe 1999:133; Matson 1991; Smiley 1994)
have proposed that a revised chronology is needed to avoid obscuring the spatial
and temporal variation subsumed by the lengthy Basketmaker II period.
Basketmaker II is part of a shared Basketmaker-Puebloan cultural tradition, but
there is debate whether there is cultural continuity from Late Archaic to
Basketmaker II (Lipe and Pitblado 1999: 106). Following Lipe (1999:133), the
Basketmaker II period includes the Early Agricultural period as defined by
Huckell (1996: 343-344).
Basketmaker II archaeological components are defined by the presence of
cultigens, primarily maize, and the absence of pottery. In fact, archaeologists

working in some parts of the Southwest refer to would-be Basketmaker II sites as
"Archaic-with-maize" (see Huckell 1996; Lipe 1999: 133). Hunting and gathering
continued to be an important component of Basketmaker II subsistence and maize
appears to have increased in importance throughout the period (Lipe 1999: 159-
161). In southeast Utah, multiple lines of evidence, including stable carbon-
isotope and coprolite analyses, indicate a reliance on maize similar to that found in
later Pueblo periods (Matson et al. 1988:248).
Whether the spread of cultigens during the Basketmaker II period was the
result of immigrant agricultural groups arriving in the central Mesa Verde region
or a more indigenous development is not currently clear. Matson (1991; cited in
Lipe 1999: 137) argues that west of the central Mesa Verde region, in the Grand
Gulch area of southeast Utah, migrations of Cochise farmers from the Sonoran
desert are responsible for Basketmaker II sites. Meanwhile, Matson suggests that
east of the central Mesa Verde region, in the Durango area, agriculture developed
by the local adoption of cultigens through contact with agricultural migrants. Thus
there may not be a pan-Southwestem, or even pan-Mesa Verde, solution to the
migrant vs. indigenous debate.
Pitstructures are the most common Basketmaker II habitation in the central
Mesa Verde region. Rockshelters were also used. It is likely that many
Basketmaker II rockshelter habitation sites have been obscured by later reuse,

especially by the construction of masonry rooms during the Pueblo III period.
Additionally, many Basketmaker II sites have likely been classified as late Archaic
components on the basis of surface remains (Lipe 1999: 152-153). These factors,
in addition to the lengthy temporal duration of the period, contribute to the
difficulty of establishing population estimates. Nevertheless, Basketmaker II
populations of the central Mesa Verde region are considered to have been quite
low in comparison to later periods. Populations appear to have been settled in
patchy concentrations (Lipe 1999: 155), making use of canyon bottoms rather than
fertile uplands for agricultural production. An occupational hiatus may have
occurred from A.D. 375-575 in the central Mesa Verde region (Wilshusen 1999a:
Basketmaker II mobility has received little attention in the central Mesa
Verde region. However, the situation may have been similar to that proposed by
Varien (1999a) for later periods, that is with communities being fairly sedentary
and constituent households moving more frequently. Survey data from Cedar
Mesa in southeast Utah show repeated use of particular locales by loose clusters of
households and may support this notion (Matson et al. 1988). The Basketmaker II
residents of Cedar Mesa seem to have chosen habitation locales that provided an
agriculturally productive compromise between precipitation and growing season
duration (Matson et al. 1988: 249).

Basketmaker HI (A.D. 500-750)
Basketmaker III is perhaps one of the most interesting and least understood
periods in the culture history of the central Mesa Verde region. During this time
there is continuity in trends that developed during Basketmaker II. Other cultural
developments, many which show continuity throughout the Pueblo periods, make
their first widespread appearance during Basketmaker III.
There is a dramatic increase in the number of sites during this time period,
with almost 2000 Basketmaker El sites recorded in the Southern Colorado River
Basin (Wilshusen 1999a: 166). Habitations are the most commonly recognized
site type. Habitation sites typically consist of pitstructures with antechambers and
several extramural features, including storage cists, hearths, and other vegetal-
covered storage features. Approximately 10 percent of these sites are surrounded
with a stockade (Wilshusen 1999a: 180). Both the stockades and the increase in
storage facilities suggest increasing duration of occupation, prominence of a
logistical mobility strategy, and an increased reliance on stored foods.
There also appears to be a shift in population density from the previous
Basketmaker II period, with populations moving from the eastern and western
margins of the central Mesa Verde region toward its center. This trend of spatially
shifting population density continues throughout the Puebloan occupation of the

region. Settlement is largely dispersed and seems to reflect the basic subsistence
pattern, as sites tend to be located on good agricultural soils, near both permanent
water sources and pinyon-juniper woodlands (Birkedal 1976; Fetterman and
Honeycutt 1987; Rohn 1963; Wilshusen 1999a: 186). Geomorphological research
in McElmo Canyon by Force and Howell (1997) documents that the Basketmaker
III settlement pattern of residence associated with good agricultural lands is related
to periods of sediment aggradation in valley bottoms. These authors also show
that this association between settlement and cycles of aggradation and stream
entrenchment is found throughout the subsequent Pueblo periods in McElmo
Basketmaker m groups had become increasingly reliant on agricultural
production for subsistence (Wilshusen 1999a: 185-187). Research suggests that
maize may have contributed 50-80 percent of the diet (Stiger 1979; Decker and
Tieszen 1989). There also seems to be a concomitant decrease in the amount of
meat in the diet (Blinman 1988; cited in Wilshusen 1999a: 186). Wild plant foods,
especially weedy pioneer species, also played a considerable dietary role.
Minnis (1989) uses coprolite data to demonstrate that local patterns of plant
food consumption in subregions of the northern Southwest developed at this time.
Comparing coprolite data from Mesa Verde, Chaco Canyon, Glen Canyon, and
northeast Arizona, Minnis argues for temporal stability and regional variability in

dietary plant use from Basketmaker III until Pueblo III. Minnis' research shows
that by Basketmaker III agriculture was already an important subsistence
component. By Basketmaker III com is the most abundant and ubiquitous
component in coprolite samples from across the Four Comers region (Minnis
1989: 559). Furthermore, the most important non-domesticated plant foods were
those that prosper in disturbed, agricultural contexts (Minnis 1989: 550).
Perhaps related to these Basketmaker III changes in demographics,
settlement, and subsistence, is the development of first archaeologically
recognizable communities (see Wilshusen 1999a: 187-188). Although the public
architecture found at Basketmaker III sites to the south has not been documented
in the central Mesa Verde region, settlement patterns suggest that the communities
of the later Pueblo periods were beginning to take form during this time. Multiple-
residence hamlets become more common during this time, and have longer
occupation spans than single-residence hamlets (Wilshusen 1999a: 187-188). In
addition, survey suggests that multiple-residence sites are commonly surrounded
by a dispersed settlement of smaller hamlets. The larger, multiple-hamlet sites are
frequently stockaded, and it has been suggested that these sites may have provided
a focal point for community identity and organization similar in function to Pueblo
era sites with public architecture (Wilshusen 1999a: 188).

Finally, the first widespread occurrence of ceramics appears during
Basketmaker III. Early Basketmaker III ceramics are typically brown wares made
of naturally tempered alluvial clays. By the late A.D. 500's there is the appearance
of plain gray wares and unslipped, polished, and painted white wares. Red wares
first appear between A.D. 725 and 740 (Wilshusen 1999a: 173)
Pueblo I (A.D. 750-9001
The Pueblo periods experienced striking cultural change in the Mesa Verde
Region. Early characterizations of the Pueblo I period depicted this time as one of
continuity and stability (Wilshusen 1991: 129). Schachner (2001: 168-169) points
out that prior to the work of the Dolores Archaeological Program (DAP) the
Pueblo I period was seen as one of steadily increasing population and a
homogenous, dispersed settlement system. These early characterizations can be
seen as a product of an early evolutionary approach that emphasized local phase
schemes, cultural centers, and the broad similarities found among Puebloan groups
across the Southwest (Wilshusen and Ortman 1999: 371-373). Since the work of
the Dolores Archaeological Program (DAP), however, the Pueblo I period in the
Mesa Verde region has come to be seen in less simplistic terms. Recent Pueblo I
research has revealed and focused on subjects such as local cultural diversity,
large-scale population movement, village formation and dissolution (Wilshusen

and Ortman 1999), as well as agent-driven changes in ritual practices and
sociopolitical strategies (Schachner 2001; Wilshusen 1991).
Perhaps the most notable aspect of Pueblo I culture history is the initial
appearance of villages in the Mesa Verde region. Villages are conventionally
defined as settlements of 50 or more contiguous or near-contiguous surface
structures and associated pitstructures (Wilshusen 1999b: 210). The occupation of
Pueblo I villages appears to have been fairly short-lived, with an average
occupation span of 25-40 years (Wilshusen 1999b: 210; Wilshusen and Ortman
1999: 374). These villages tend to occur in clusters and are generally surrounded
by more dispersed clusters of hamlet sites, which can range in size from a few
surface rooms with a single associated pitstructure to 25 surface rooms associated
with 3-4 pitstructures (Schachner 2001: 173).
Although a majority of Pueblo I people may have lived in villages by the
end of the period (Wilshusen 1999b: 210), these villages did not develop with
temporal or spatial uniformity throughout the central Mesa Verde region. The
earliest cluster of villages appeared in southeast Utah in the mid-700's and seem to
have been abandoned by A.D. 800 (Schachner 2001: 174; Wilshusen and Ortman
1999: 374). Populations then shifted eastward in a cluster of villages that stretched
from present-day Durango, Colorado to the southern portion of Mesa Verde
proper. These villages were occupied from approximately A.D. 800-840

(Schachner 2001: 174; Wilshusen and Ortman 1999: 374). Then from A.D. 840-
880, villages clustered in an extensive area from Mesa Verde proper north to the
area around present-day Dolores, Colorado and westward into Utah (Schachner
2001: 174; Wilshusen 1999b: 210-211; Wilshusen and Ortman 1999: 374). This
area that provided residence for an estimated 22 households from A.D. 760-800,
was home to approximately 496 households between A.D 860 and 880 (Wilshusen
1991). Such dramatic demographic shifts are argued to have been associated with
changes in sociopolitical organization across the Puebloan Southwest (Feinman et
al. 2000: 459).
There are at least 21 villages associated with this latest episode of Mesa
Verde region village formation. The Dolores village clusters were an important
focus of DAP research, and have consequently received the most attention in the
archaeological literature (e.g., Orcutt et al. 1990; Schlanger 1992; Wilshusen
1991). The Dolores area was largely abandoned for habitation purposes by A.D.
900, with populations possibly moving to northwest New Mexico, then following
a more dispersed settlement system oriented around great kivas (Wilshusen and
Wilson 1995; Wilshusen et al. 1999). Archaeologists working in the Mesa Verde
region (Wilshusen and Ortman 1999: 382) have recently emphasized that the
Pueblo I emigration from the Mesa Verde region in the late ninth century is

equally important to understanding the prehistory of the northern Southwest as the
more familiar 13th century Puebloan migration.
Wilshusen and Ortman (1999: 374-376) note similarities among Pueblo I
villages from the A.D. 840-880 Dolores cluster. For instance, all of these villages
are comprised of groups of multiple roomblocks, ranging in length from 22m to
170m. These roomblocks are comprised of several household suites, with rooms
that functioned primarily as domestic activity areas associated with smaller, more
specialized rooms that likely served as storage areas for agricultural foodstuffs.
Depending on size, each roomblock could have provided habitation for four to 22
households. Population estimates from these villages greatly exceed numbers that
could have resulted from in situ population growth. The dramatic population
fluctuations associated with village formation and shifting locations of village
clusters can be explained by the migration of large groups of people within and
beyond the Mesa Verde region during the Pueblo I period (Wilshusen and Ortman
Although such meaningful similarities exist, differences among Pueblo I
villages are perhaps more useful in attempts to understand the social context of this
period. Schachner (2001) focuses on the diversity of architectural forms in which
Pueblo I ritual took place. These settings include great kivas (kivas with a
diameter greater than 10 meters), "oversized" pitstructures, and "regular"

pitstructures (Schachner 2001:172-178) that are associated with both ritual and
domestic functions (Varien and Lightfoot 1989). Schachner points out that in the
A.D. 790-840 period community ritual took place in great kivas. These great kivas
fell into disuse when the residential sites with which they were associated were
abandoned. From A.D. 840 to A.D. 880, community ritual appears to shift rapidly
to oversized pitstructures, coincidentally with the shift of village clusters to the
Dolores area. Schachner argues that the smaller size and location of these
oversized pitstructures within u-shaped roomblocks allowed certain segments of
the population to control ritual participation. This control of ritual participation is
considered by Schachner to have been an intentional attempt by individuals to
increase their social standing and ability to control land, factors associated with
ritual in modem Pueblo society. Schachner asserts that these attempts at control
and manipulation of community ritual failed, since oversized pitstructures did not
continue to be a locus of local community ritual, nor did they come into use in the
later settlement clusters of northwest New Mexico (Schachner 2001:184).
As mentioned briefly above, Wilshusen and Ortman (1999: 382-392) cite
the coalescence of historically distinct ethnic groups to explain differences
between Pueblo I villages in the Dolores area. These authors compare excavated
villages on opposite sides of the Dolores River valley. In particular, they note
differences in village demographics, construction history, ceramic assemblages,

and architectural layouts among these villages, with the Dolores River creating the
boundary for these distinctions. Wilshusen and Ortman (1999: 380) suggest that
the ethnically distinct groups responsible for these differences may have come
from the upper San Juan River area, the Durango area, and Elk Ridge, Utah.
Having briefly noted some similarities and differences among Pueblo I
villages, the question of why such episodes of migration, aggregation, and
abandonment occurred during the Pueblo I period in the Mesa Verde region still
remains. These processes have been addressed to varying degrees in the
archaeological literature (see Cameron 1995). Although they have received
unequal attention, it is important to make a few points in regard to these processes.
First of all, the processes of migration, aggregation, and abandonment are all
inextricably related and should be considered with this in mind (Adler et al. 1996;
Cameron 1995). Additionally, it must be realized that a universal or even pan-
Mesa Verde region model is not likely to be able to explain these processes as they
occurred at different times and places. For example, the factors that affected the
settlement patterns of the Pueblo I period in the Mesa Verde region may not have
had the same influence on the migration and aggregation that characterized the
Pueblo HI period. Finally, as Kohler and Sebastian (1996: 599) emphasize writing
with specific attention to aggregation, any plausible model for such complex
processes must certainly consider multiple factors.