5—;- zlaft .‘l‘ ' "5 A

”.0?"

:A 1.2.1;

I.
t
5
V
o

a»:

4‘ ' ".rlf'tufll"

w 63-01”!

‘ 1 u_U—'.
L '. 32:31... wag?!-
v. ‘ .~ v ' 1 ‘ ' ‘I
"’.."..a|-‘.7:--" ‘ ‘
",.§:X:LI£‘.{l-§:{nk

.a
. -.--...

.._.

45'?»

Ol'.
~¢

l4
fl.‘ [2
"t. 3. .

m-f

p:
n
. lift: :

o
l. I '1‘
,. n .
A I

oa-v

1??
0,1, "t
. :Eififlfi‘fii} a!" ‘ Si
M- ~=‘4‘i‘1 ‘1 9. uun- war":
.it“1iéiii£1 3 - (ntfi’mfi’!
$33 *‘éfifle*?' ’13 w
. 5'

ul'

1: uric

"5!. 3’59
' :‘M‘ 1|
. :~

,1

-9

.n.

 

,1,1

33
\

.. r
M» "0“

A ...,—

3
.—
-y—n

l
l

x ..-..

RSBRAHITYU

llllllllllllllllllllllllllllllll

3 1293 017102

Illlllllllll

 

 

 

This is to certify that the

thesis entitled
THE MERLE BEACH SCHOOL SITE (ZOCL 275):
AN ASSESSMENT OF THE ARCHAEOLOGY
OF RURAL MID-MICHIGAN EDUCATION
DURING THE NINETEENTH CENTURY,
AND METHODS FOR ITS EXAMINATION
presented by

Valerie J. Hartzer
has been accepted towards fulfillment

of the requirements for

Master of Arts degree in Anthropology

 

 

 
    

Major professor

Joseph L. Chartkoff
Date February 4, 1998

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY

 

Unlverslty

Mlchlgan State

 

 

PLACE IN RETURN Box
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE I DATE DUE

DATE DUE

 

Mamie!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M Mpg-“1‘

THE MERLE BEACH SCHOOL SITE (ZOCL 275):

AN ASSESSMENT OF THE ARCHAEOLOGY
OF RURAL MID-MICHIGAN EDUCATION
DURING THE NINETEENTH CENTURY,
AND METHODS FOR ITS EXAMINATION

By

Valerie J. Hartzer

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Anthropology

1 998

ABSTRACT
THE MERLE BEACH SCHOOL SITE (ZOCL 275):
AN ASSESSMENT OF THE ARCHAEOLOGY
OF RURAL MID-MICHIGAN EDUCATION
DURING THE NINETEENTH CENTURY,
AND METHODS FOR ITS EXAMINATION
By

Valerie J. Hartzer

This thesis examines a rural schoolhouse, the Merle Beach School Site,
located in Olive Township, Clinton County, Michigan, which was in use from 1863
to 1966. The thesis has two foci. The first involves the examination of the rural
schoolhouse as a type of archaeological site, one which to date has been largely
unexamined in Michigan archaeology. Another part of this focus involves the
examination of the significance of the rural school as a community institution and
the evolution of its functions over time. The second focus assesses the
comparative effectiveness of alternate procedures used in site sampling to

provide relatively accurate representative data on this type of archaeological site.

This is dedicated to my two greatest loves and supporters
Paul and Selena

ACKNOWLEDGEMENTS

Many, many people and organizations helped greatly with the preparation
of this paper. Thanks goes out to the Michigan State University Archives, the
Bently Historical Collection, and the State of Michigan Archives, all of which were
consulted for various stages of this paper. Special thanks to Terry Shaffer, the
current landowner of the Merle Beach school for allowing me to excavate on his
property, and providing me with his knowledge of the site.

Special thanks go out to my family, especially to my grandparents Frank
and Eleanor Dietz, whose financial support helped to fund my research, and to
my mother, Rosemary Hartzer, for her support and encouragement.

Special thanks also go to my Close friends Laura Godin and Robert
Henderson for their support and tolerance with me for ignoring them, putting
dates on hold, and cutting things short in order to have more time for my thesis.

Thanks also to my committee members, Dr. Joseph Chartkoff, Dr. \Mlliam
Lovis, and Dr. Kenneth Lewis, for their support, encouragement, and shoving
when needed.

And finally, thanks to all of the people who helped me get through one of
the most emotionally trying times in my life, including Dr. Sundari Rajan, Dr.

Carey Krause, Dr. Phillip Parker, and Saundra Waundless, M.A.

TABLE OF CONTENTS

List of Tables ............................................................................................ vii
List of Figures .......................................................................................... viii
Introduction ................................................................................................. 1
Chapter 1: History ....................................................................................... 8
1.0. Introduction .................................................................................... 8
1.1. The Nature of the One-Room School ............................................. 9
1.2. The Federal Role in Shaping Local Rural Education .................... 11
1.3. The Philosophy Guiding the Development of Local Schools,
Especially in Rural Michigan ..................................................... 13
1.4. Establishment of the Merle Beach School .................................... 14
1.5. Operation of the Merle Beach School .......................................... 16
1.6. The Fate of the Merle Beach School ............................................ 19
1.7. Types of Potential Archaeological Data in the Context of a
One-Room Schoolhouse .......................................................... 21
1.8. Conclusions .................................................................................. 28
Chapter 2: Methodology ........................................................................... 30
2.0. Introduction .................................................................................. 30
2.1. Sampling Research Design .......................................................... 31
2.2. Sample Collection Units and Sample Unit Selection .................... 33
2.2.1. Phase 1: Non-invasive techniques ................................. 41
2.2.2. Phase 2: Invasive techniques............. ............................ 42
2.3. Archival Research .............................. ‘ .......................................... 42

V

2.4. Data Analysis Methods ................................................................ 44

2.5. Conclusions .................................................................................. 45

Chapter 3: Data Analysis and Interpretations ........................................... 46

3.0. Introduction .................................................................................. 46

3.1. Spatial Organization: Surface Collection Data ............................. 46
3.2. Interpretations of the Spatial Organization Of the Surface

Collection Data ......................................................................... 48

3.3. Spatial Organization: Posthole Data ............................................ 50

3.4. Interpretations of the Spatial Organization of the Posthole Data..53

3.5. Spatial Organization: Five Foot by Five Foot Unit Data ............... 54
3.6. Interpretations of the Five Foot by Five Foot Unit Data ................ 57
3.7. Analysis and Comparison of the Sampling Methods .................... 59
3.81 Additional Non-Statistically Based Interpretations ........................ 65
Conclusions .............................................................................................. 67
1. Conclusions of Compared Sampling Strategies .............................. 67

2. Results of the Patterning and Interpretations About Schoolhouse
Use ........................................................................................... 69
3. Summary of Conclusions and Recommendations ........................... 72
Appendix A: Tables .................................................................................. 76
Appendix 8: Maps and Graphs ............................................................... 120
Bibliography ............................................................................................ 166

vi

LIST OF TABLES

Table 1: School Expenses for the Olive Township Schools 1861-1930
(In dollars, taken from the Annual School Supervisor's Reports) ............. 17

Table 2: Attendance Records for District Three, Olive Township,

Clinton County, Michigan (Taken from the Annual School Supervisor’s
Reports) .................................................................................................... 17
Table 3: Table of Basic Statistics for Artifact Distributions ........................ 60

Table 4: Table of Basic Statistics for the Transformed Artifact Densities .63

Table 5: Surface Collection Data Set ........................................................ 76
Table 6: Raw Data for Posthole Collection ............................................... 80
Table 7: Raw Data for 5x5 Ft. Units .......................................................... 84
Table 8: Coal, Concrete, and FCR Weights for 5 Ft. Units ....................... 96

Table 9: Pearson Correlation Matrix for the Surface Collection Data ....... 97

Table 10: Pearson Correlation Matrix for the Posthole Data .................... 99
Table 11: Pearson Correlation Matrix for 5 Ft. Unit Data ........................ 102
Table 12: K-Means Cluster Analysis Table for Posthole Sample ............ 105

Table 13: K—Means Cluster Analysis Table for 5 Ft. Units — 3 Clusters.110

Table 14: K-Means Cluster Analysis Table for 5 Ft. Units — 5 Clusters.114

vii

LIST OF FIGURES

Figure 1: Sketch of Girls Playing on a Rough Stride Swing ...................... 24
Map 1: Map of Michigan, Showing Lansing and the Location of the

Merle Beach School Site ........................................................................ 120
Map 2: Merle Beach Schoolhouse Site Map ........................................... 121
Maps 3-16: Surface Collection Contour Maps ........................................ 122
Maps 17-42: Posthole Contour Maps ..................................................... 129
Maps 43-72: 5 Ft. Unit Contour Maps ..................................................... 142
Graphs 1-2: Frequency Distribution for Surface Collection ..................... 157
Graphs 3-4: Frequency Distribution for Posthole Tests .......................... 158
Graphs 5—8: Frequency Distribution for 5’x5’ Test Units ......................... 159
Graphs 9-12: Surface Frequency Distribution Box Plots ........................ 161
Graphs 13-16: Posthole Frequency Distribution Box Plots ..................... 162
Graphs 17-24: 5 Ft. Units Frequency Distribution Box Plots .................. 163
Graphs 25-27: Transformed Frequency Distribution Box Plots .............. 165

viii

INTRODUCTION

The purpose of this thesis is two-fold. One is to examine the rural
schoolhouse of the Nineteenth Century, its role in American rural society, its
function in the community, and the spatial patterning of its archaeological
remains, through which we can investigate these roles and functions in one
particular schoolhouse. The second aim is to study the effectiveness of a number
Of different sampling methods and examine the relative contributions of alternate
sampling methods on a small, discrete site, such as a small rural schoolhouse.

For much of the United States’ early history the vast majority of the
population has lived in rural areas. In 1790, 96.6% of the US. population lived in
rural areas. By 1880, 77.4% of the US. population still lived in rural areas, and
by 1900, 66.9% of the US. population was still living in rural areas, accounting
for the majority of the total population, despite increasing urbanization and
industrialization (Foght 1910: 5). VVIth such a large proportion of the population
living in rural areas, rural education was extremely important, since that is how
the majority of children was educated. Even as late as 1910, half of all US.
schoolchildren, 12 million in all, attended rural schools (Foght 1910: 2). With the
number of rural schools, their vast importance to the life of the people cannot be
underestimated. There has been some historical study of rural education and
individual rural schools, but comprehensive studies incorporating archaeological
evidence as well as historical evidence are virtually nonexistent. Archaeology is
beginning to look at institutions and their social functions, such as the standard

forms of manorial architecture and the power they conveyed (Johnson 1992:

1

45-55), and total control and integration of social life in Lowell, Massachusetts, by
the Boott Mills corporation (Beaudry 1989: 19-32). The investigation of rural
schoolhouses, with their important role in American society, would be a
significant contribution to archaeology.

By using a combination of different archaeological sampling methods to
extract a wide range of data from across a large section of the site of the Merle
Beach school (20CL 275), as a means of delineating activity areas at and around
the schoolhouse, an archaeological picture of daily activities at a schoolhouse
can be drawn. By adding to the archaeological record through the use of archival
records of the school district, its schoolhouse and local history, a more complete
picture of the schoolhouse, and its place in the community can be drawn than
has before been possible through either history or archaeology alone.

Nineteenth and early Twentieth Century schoolhouses have undergone
little in the way of formal archaeological investigation throughout Michigan. There
have only been two formal excavations of schoolhouses within the state
(John Halsey, personal communication 1997). No site report was completed for
one of the schoolhouses, as nothing but a building trench with evidence of a
mortar-topped stone foundation and a few slate pencils were uncovered (Mark
Branstner, personal communication 1997). The other schoolhouse is the subject
of the present thesis, the Merle Beach school, which had a rich collection of
artifacts, including construction debris and a small amount of domestic debris,
such as pieces of dishware, as well as buttons and other small personal objects.
This makes the Merle Beach school site an especially rich site and a uniquely
valuable research opportunity.

In addition to being a unique research opportunity for the study of an

important, as yet little-studied, facet of history, the site under study was small and

reasonably well documented, making it an ideal laboratory for the study of a
number of different sampling strategies. By performing different sample test
strategies, the results of each test could be compared to look at the relative
effectiveness of each type of test. Also, documented information about the school
and its use could allow for a controlled assessment of the productivity of each
testing method. Archaeologists have long been looking at different sampling
methods, but seldom does such a good site, one that is small enough to
encourage the use of multiple types of sampling, present itself.

The first goal of the study is to ask what the role was of the schoolhouse in
rural America during the Nineteenth Century. What functions did the schoolhouse
serve in the community in which it was constructed? What in particular was the
role of the school, education, and the schoolhouse in Michigan during this time
period? How does Merle Beach school fit into the patterns of rural schoolhouses
through out the state?

Education was highly valued by even the earliest Colonial settlers to North
America. In 1642, the General Court of the Massachusetts Bay Colony enacted a
law that required all parents and masters (craftsmen with apprentices, or those
with indentured servants) to provide an education in reading, writing and civil law
through either home instruction or a common school. This act further required
that in every township containing fifty families a public school was to be
established and that, when the population reached one hundred families, a
grammar school capable of educating children to be university-ready be formed
(Hosford 1870: 23-24). Schooling was compulsory under this system (Hosford
1870:24y

The common school education of each child was anoriginal
condition of settlement, a fundamental principle of the social

4

compact, as between parents and children, masters and servants,
under the guardianship of the State.

The usual manner of school management in most areas of the country
was for rural education to be administered directly by the local community. In the
beginning this would have been by towns or villages, which were broken into
school districts, and then by township, as the country was surveyed; many states
had switched to this method of school management by 1882. As the Nineteenth
Century progressed, and more people began to move into urban areas, urban
schools began to become too large to accommodate students in a one room
structure. The concept of the“ graded school came about during the mid-
Nineteenth Century (Kiddle, et al 1877: 3-4), in which children were separated by
their ages and moved through the studies of one grade of course work a year.
This concept was particularly valuable to the urban schools, as children could be
broken into grades and each grade placed in its own room, accommodating the
much larger number of students. In addition, the graded school of the City was
often likened to the wonders of industry and the new factories and conveyor
belts, and was heralded as the modern, progressive way to educate children. In
contrast to the urban graded schools, students in the rural one-room school
progressed at their own pace through the eight primary grades. Usually, several
small groups of children at about the same level as each other would be seated
together in several clusters around the same common room, more or less broken
into grades. By the Close of the Nineteenth Century many educators began to
see the rural school as an outmoded icon of the past.

As rural enrollments began to decline in the 18705 and the concept of the
graded school began to become popular, rural school districts could not maintain
the costs of upkeep for a large number of scattered schoolhouses, each with its

own teacher, equipment, library, school building, and out-buildings. This led to

the formation of consolidated schools (Foght 1910: 306). Consolidation meant
that small, poor districts could pool their resources. They could construct multi-
room schools and offer graded schools for less than the upkeep and hiring of a
teacher had cost each one-room school. The consolidated schools were
frequently placed in the center of the school districts that they served and were
farther than the district schools they replaced. Fortunately, savings among the
consolidated school districts most often resulted in enough extra money that
children living far from the new schools could be conveyed by wagons, later
buses, at no additional cost. This new practice reduced the long walks that many
children needed to make in order to go to and from school. In some cases high
school courses could also be taught, allowing more rural children the chance of
both high school education and the background needed to attend college (Foght
1910:322)

Rural schoolhouses in the Nineteenth Century were frequently the center
of a great deal of community activity, with dances, socials, plays, and public
assemblies of all kinds such as town officers meetings and school board
meetings. Often the schoolhouse was the first public building constructed in a
settlement and was used for multiple purposes, until additional public buildings
could be constructed (Smith 1880: 7-9). As a whole, education was highly prized
in American society of the Nineteenth Century. It was seen as the means to an
enlightened citizenry (Culter and Stone 1913: 159) and a great deal of emphasis
was put upon improving education and providing instruction in as many branches
of learning, including classical studies, as could be accomplished on the budget
of the rural school district (Smith 1880: 28).

Rural schools typically employed either two school teachers for two short,

several month terms, or one school teacher for afour to six month term of school

each year. Rural schools accommodated students of all grades, first through
eighth, with one teacher providing instruction at all grade levels. In order to
ensure quality education the best teacher a district could afford was a high
priority for at least one session a year. A sufficient schoolhouse with some type
of blackboard and enough desks to fit the number of students that regularly
attended as well as some basic equipment were the other expenses needed to
maintain a rural school. Because the equipment and personnel were fairly
minimal in a rural school, even relatively poor areas could frequently afford to
provide an adequate, or at least a minimal, school. Also, as the expense of a
schoolhouse was necessary for even a basic school, it could be reused for other
community social events and activities and be of further use to the community
reducing the expense of building other buildings if need be (Smith1880: 7-9). For
example, in Olive Township, where the Merle Beach school is located, church
services were held at schoolhouses around the township until the Olive church
was built in 1864 (Anonymous A 1880: 496). One of the schools in that same
township, the Muskrat schoolhouse (which later became Merle Beach school),
was also used to house early meetings of the local Grange, a rural social society,
until a proper Grange hall could be built (Anonymous A 1880: 496).

In this study a rural schoolhouse, Merle Beach School, located near
DeWItt, Michigan, was chosen to be investigated archaeologically. Historical
records were drawn upon to derive a model of what a typical rural schoolhouse
and grounds would contain, what buildings would be commonly found on such a
site, where they would commonly be placed, and what activities occurred at a
rural schoolhouse.

After generating a rough model of a rural school, a strategy for excavating

a rural school site was devised, including the excavation of several different size

sampling units. Both systematic and random means of sample unit selection
were utilized.

Following the excavation of the site, the artifacts were analyzed through a
number of statistical tests in order to examine the relative effectiveness
(SChiffer1978) of the different sampling units and unit selection strategies.

Also, an examination was made of what this excavation can tell archaeologists
about the nature of the archaeological remains of school sites and how they can
be more effectively excavated.

The next chapter is an examination of the rural school, its history, and the

history of the Merle Beach School.

Chapter 1
HISTORY

1. 0. Introduction

In order to examine the Merle Beach schoolhouse as an example of a
Nineteenth Century rural schoolhouse, it must first be placed in the context of its
times. First, a brief examination of the nature of the rural schoolhouse is provided
to establish a framework for its investigation, followed by an examination of the
Federal laws and the ideology governing Federal regulations and funding of
education, which establishes the role the federal government played. The impact
of federal regulations and funding on schools at the local level, during the
Nineteenth Century, was also examined for this study. Examining the Federal
govemment’s role in early school history is fundamental to understanding why
schools were organized as they were, why they were funded locally, and why
they were not standardized. Another aspect which places the Merle Beach
schoolhouse within the context of its times is the philosophy which guided
education during the Nineteenth Century, and the impact that this philosophy had
on the structure of rural education and its management within the local one-room
schoolhouse.

Shifting the focus of the analysis of schools to the local level, the Merle
Beach school must be examined within the context of the organization of the
county, township, and school district within which it was formed. This task
requires the examination of the community as it formed, grew, and declined, and
how these phases impacted the schoolhouse and its changing role within the

community. Another important aspect of examining the Merle Beach school is to

look at the schoolhouse itself and the types of activities which would have

occurred there.

1.1. The Nature of the One-Room School

The place to begin when considering the excavation of a one-room
schoolhouse is the nature of the one-room school and the activities held there. In
rural communities, the one-room school was frequently the first civic building
erected, and sometimes the only one a very small town would have (Anonymous
A 1880; Foght 1910: 148-151). School sessions were held for between three and
nine months a year, usually in two sessions. Typically, school terms were held
from late summer to early spring, and a young female teacher, usually a local
teenager with a state certificate, would be hired for the beginning of the year
(Anonymous C 1861 -1 930). An adult male teacher would usually take over when
the snows started and the larger, older boys came to school. School classes
were generally held from eight o’clock in the morning until four o’clock in the
afternoon, with two fifteen minute recess periods and an hour break for dinner
(that is, lunch). During these periods children played group games and organized
games or played with whatever equipment the school had, usually a set of
swings or other inexpensive equipment that was simple (Culter and Stone 1913:
112-124).

All ages of children were taught together in one room at these schools.
Children were grouped together by grade and each grade worked its way though
a series of exercises in a particular subject. Students were advanced as they
progressed and held back if they did not do well on tests and assignments. As a
result, one student might complete two grades in one year and another might

take two years to complete a grade. Each student in each grade in each subject

10

was required to recite, a process in which the class stood in front of the school
and answered the teacher’s questions on the presented material. Students also
were required to present long oral papers and to recite speeches of famous
people, placing a great deal of emphasis on public speaking (Culter and Stone
1913: 75-100; Mueller 1926: 179—194).

Schools were simple buildings with one large classroom and two small
coatrooms, one for the boys and one for the girls. The classroom held a set of
desks for students of all ages and sizes, a large table or desk for the teacher,
and a stove. A black-painted panel of wood or a slate chalkboard was nailed to
one wall, and there might be pictures or maps as school funds allowed. The
surrounding out-buildings were few and simple, usually consisting of two
outhouses, one for each sex, and a shed to house fuel for the stove. Some
schools had a nearby cottage for the teacher, or a stable for the teacher’s horse
if the teacher rode a long way to the school (Otvvell 1923; Foght 1910: 122, 124,
126-1 32).

School was only held from Monday to Friday. The building was frequently
used after hours for social events such as school socials, school board and town
meetings, children’s plays and activities, and local farm Club meetings. On
Sundays the schoolhouse was frequently used for prayer meetings or church
services when the town had no church. Touring preachers or religious revivalists
might preach at the schoolhouse on Sundays (Anonymous 1880).

All in all, the rural one-room schoolhouse was an all-purpose building with
simple fittings that served many vital roles in the community which it served. lt
differed slightly in every community where such a school was built, and its
functions differed slightly from community to community, ever flexible, ever

changing with time.

11

1.2. The Federal Role in Shaping Local Rural Education

An initial consideration of the Federal role in rural education will be an
investigation of the attitudes and perceived role of education by the federal
government, and how these attitudes shaped educational policy in Michigan.
Examining the Federal govemment’s role in early school history is fundamental
to understanding why schools were organized as they were, why they were
funded locally, and why they were not standardized. Education played an
important role in the community and in the social life of individuals. Education
was intended to socialize children for their adult lives, teach them the necessary
skills they would need as adults, and convey a knowledge of the world at large.
With the many functions the school filled, it played an important role in children’s
lives.

As Oramel Hosford states in the 1873 Atlas of Michigan, “as early as
1787, an ordinance was passed by Congress for the government of the North
Western Territory, in which it was declared that ‘schools and the means of
education shall forever be encouraged.” The federal objective for education in its
early legislation was to encourage the education of children to promote the
development of an enlightened citizenry. At the turn of the Nineteenth Century
the federal government became much more forceful in seeing that at least a
minimal education be given to all children.

The territory of Michigan was organized in 1805 and the first school law
enacted in 1809 by the territorial legislature (Anonymous D 1880: 7). This law
called for the division of the most heavily settled areas into school districts, and
established an annual tax to support schools (Putnam 1904: 4). In 1817 a new

school law was passed establishing the University of Michigan. That law gave the

12

university faculty complete control of overseeing schools at all levels, but proved
to be impractical in that it prevented taxpayers from having any control over the
university faculty, which were appointed by the territorial government (Putnam
1904: 6). In 1821 a new education law was passed which put the university under
the control of an elected board of trustees (Putnam 1904: 8). This law, in effect,
limited the control of the University over local schools, nullifying the 1817 school
law. It was in effect until the passage of the 1827 school law.

In 1823 and 1827 the United States government changed the
administrative structure of the territory by establishing the position of Governor
and board of advisors elected by the people. Governmental reorganization also
resulted in the division of the territories into townships, and provided for local
township control of municipal services (Putnam 1904: 11).

As a result of this reorganization, a new education law was passed in
1827 which turned management of schools over to the township government. It
provided for the division of townships into school districts, and established
minimum requirements for the operation of a school (Anonymous D 1880: 11-12).
These modifications gave local residents greater control over their area schools
by allowing them to elect their own school board officers and run their schools in
a way that met their needs best, both financially and for the greatest benefit of
their community. Each township could then subdivide itself into school districts.
Each district could hire its own teacher with the direct consent of the local
parents, rather than being forced to comply with either state and federal
regulators or interference from collegiate governing boards, like the University of

Michigan, as under the old regulations of 1823.

13

1.3. The Philosophy Guiding the Development of Local Schools, Especially in
Rural Michigan.

One of the ideals held by educators across the country, at the turn of the ‘
Twentieth Century, was that schools contribute greatly to the public life of the

community. As Horace Seerley points out (1910: 4-5):

the people need to be awake to every interest that contributes to
the welfare of the whole community, and, since happiness,
usefulness, and prosperity depend very largely upon intelligence,
morals, and culture, it becomes a matter of self-preservation to
have good schools. In the United States public school education is
commonly left to the local community, and experience has proved
this plan is wise and good if the people recognize the value and the
importance of good education to society as a whole and to the
individuals in particular. '

Educators across the country also believed that a good school could and should
be the center of community activity and that, in so doing, foster democracy:
“Democracy requires that people meet to discuss problems of government and
community welfare, and choose officers from among their number to carry out
their will” (Mueller 1926: 314-15).

Another ideal that turn of the century educators were striving to realize
was the use of school gardens to teach practical farming to the children, as well
as lessons in cooking and canning and other farm skills (Foght 1910:236-253).

In Michigan achieving the goals of good education and using the school as
a basis of community life was something of a challenge for the rural schools.
Rural schools frequently had classes when a teacher could be found and enough
money to pay a teacher was available, so school years were variable in length
from year to year as the fortunes of the area changed. Children in rural schools
also had variable attendance as they were frequently needed for tasks at home.

Sometimes older children left school to work full time on the farm. Children in

14

rural Michigan frequently attended from two to four months of school a year
(Anonymous C 1861-1930).

The early schoolhouse served as a meeting room until other buildings,
such as town halls, grange halls, and churches, could be built. Spelling bees,
school socials and literary society meetings were also held in schools in some
areas. However, rural schools in Michigan varied widely, with some schools
being used for school activities only.

In Michigan, like other “northern states, from the fact that the school terms
extend over that part of the year in which gardens do not grow not much has
been accomplished in school gardens in our rural schools” (Culter and Stone
1913:175); little use was made of school gardens or the teaching of farming to
rural children. Many Michigan schools, like Merle Beach with its walnut and
hickory trees along the edges of its plot, had only a few trees along their fence
lines, and school yards too small to plant gardens or much in the way of trees or
flowers. The greatest contributing factor to a lack of school gardens in Michigan,
however, was the short growing season, which falls almost entirely outside of the
months that school was taught. In the case of Merle Beach, the land is unsuitable
for gardening because it is poorly drained and may have been designated as the

school land for that very reason.

1. 4. Establishment Of the Merle Beach School

Clinton County was established by an act of the Territorial Legislature on
March 2, 1831 (Anonymous A 1880: 334). The first settler in the county was
George Campeau, who established the first trading post in Maple Rapids in
1826, before the official establishment of the county (County History 1880: 338).
Another trading post was established in 1832 by Hiram Benedict (Daboll 1906:

15

445). In 1833 yet another trading station was erected near DeWitt by David Scott
(Daboll 1906: 445). Settlers soon arrived in the area and by 1857 the first railroad
line going to St. John’s was established. Originally the county was all one
township, that of DeWItt, and was broken up in 1837 into four townships
(Anonymous A 1880: 334-335). Today, by contrast, Clinton County is divided into
16 townships.

Olive Township was detached from DeWItt Township in March of 1841
(Anonymous A 1880: 494-5). It was first settled in 1837 by Ephram Merrihew,
who settled with the families of his children and other relatives, in the town of
Olive. Settlers soon began populating the area and in 1853 the DeVVItt Road was
opened as a stage coach and mail route. The first tavern in the area was opened
by Myron Wolcott in 1853. It was called The Half-Way House because of its
position at the halfway point between DeWitt and St. Johns (Anonymous A 1880,
492)

On October 12, 1841, School Districts Two, Three, Four, and Five were
established (Anonymous A 1880: 496). School District number Three was served
by the Merle Beach School. The school was originally referred to as the Muskrat
School because of its proximity to Muskrat Lake. It was subsequently renamed in
1898 after the Merle Beach Hotel and Resort, which sat on the shores of Muskrat
Lake (Anonymous 8 1980: 510).

The first schoolhouse, the Muskrat School, sat across the road from the
present day location of the Merle Beach school and was a crude log building.

A new clapboard sided schoolhouse was erected at the present location of the
Merle Beach school in 1863 (Terry Shaffer, personal communication 1995) and is

the school building which still stands at the site.

16

1. 5. Operation of the Merle Beach School

The Merle Beach School was a fairly typical rural school of the late
Nineteenth and early Twentieth Centuries. It often struggled financially, with
school taught for as many months as could be afforded. Upkeep was provided for
the school building as money could be found to improve it.

The following chart shows the total annual expenses for each school in
Olive Township, as well as the annual expenses for hiring teachers, and
expenses for the annual upkeep of each school and its grounds (Anonymous C
1861-1930). The data Show that in 1861 District number Three, the district with
the Merle Beach School, had the highest expenditures for teachers of any
district. In 1863 Merle Beach spent the second highest amount for teachers of
any district in the township. By 1886 the amount spent on teachers had slipped to
fourth out of five districts.

This trend continued in 1873 and 1874, but in 1880 Merle Beach had risen
to the third highest expenditure for teachers of the five districts and by 1900 was
paying the second highest amount for teachers in the township. By 1920 Merle
Beach was paying the highest amount for teachers of any school in the township.
By 1930 Merle Beach was once more paying the second highest amount for its
teachers. The total amount spent on school expenses was about commensurate
with how much was being spent on teachers: If Merle Beach was paying the
second highest wages, then the district was spending the second highest amount

overall to support schooling. (See table one for details.)

17

Table 1:
School Expenses for the Olive Township Schools 1861-1930
(In dollars, taken from the Annual School Supervisor’s Reports)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L 1861 1863 1866 1868 1873 1874 1880 1900 1920 1930

#1 total expenses 99.23 69.31 95.28 79.25 13251 187.1 296.43 274.71 1054 1696

annurt paid teachers 24 47 65 45.5 102 116 133.37 196 765 1035

buid'ng repair 5 3 6.08 17.49

#2 total expenses 84.55 18278 151.71 504.68 290.32 26239 334.23 1020 1667

smart paid teachers 52 32 146.77 136 161 192 165.3 255 675 900

buid'ng rent 2 2.55 6.8 300 15.9

#3 total expenses 79.5 87.66 188.7 191.8 242.8 344.81 355.55 1137 1568

amomt paid teachers 75.5 19.5 51 127 127.5 145.88 252 200 810 990

buid'ng repa'r 10 3.99 0.6 48.93 5 27.18

#4 total expenses 61.5 345.77 329.45 813.57 319.63 319 316.4 1110 1856

smart paid teachers 57 61.5 112.6 142 187 216 164 199.75 560 1035

buid‘ng repair 7 235 200 3.95 460.04

#5 total expenses 125.25 171.25 323.89 338.14 242 270 1967.3

annmt paid teachers 97.87 144 234.5 216 164 23.63 990

building repa'r 0.5 54.97 J
I

Table 2:

Attendance Records for District Three, Olive Township, Clinton County, Michigan
(Taken from the Annual School Supervisor’s Reports)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of Number of Number of % children who in

months of children Iiu'ng children attending district who
Year school held in the district school attended MB school
1861 7 41 41 100%
1863 6 57 40 71.90%
1866 4 46 42 91.30%
1868 6 61 40 65.57%
1873 6 62 48 77.42%
1874 7.5 57 51 89.47%
1880 9 95 86 90.52%
1900 9 17 11 64.70%
1920 9 , 35 20 57.14%

 

 

 

18

The Merle Beach school, according to the map in the 1861 school
supervisor’s annual report, was located in District Three of Olive Township in
Clinton County. The report shows that, for that year, forty-one children attended
the school, which was held for seven months. Table two shows the trends in
attendance in District Three of Olive Township, Merle Beach school.

The data from the school supervisor's reports (1863-1930) show the
attendance patterns at the Merle Beach school. In 1863 fifty-seven Children
between the ages of five and twenty were located in the district, but only forty-
nine attended school that year. In 1866 forty-two out of forty-six school-aged
children attended. From 1868 to 1874 school attendance was relatively stable at
between forty and fifty Children, despite an increase in the number of school age
children in the district, which rose from forty-six in 1866 to sixty-two in 1873. In
1868 school attendance dropped sharply. That year, of the sixty-one school age
children in the district, only forty attended school at Merle Beach. Of the six
months that school was in session, the average child during these years
attended school for between two and four months.

After 1880 the number of children of school age living within the school
district declined. In that year, of the ninety-five school age children living within
the district, eighty-six attended the school. By 1900 there were only seventeen
children of school age living in the district, with eleven children attending the
school. By 1920, the number of children within the district had risen again to
thirty-five school age children, with twenty attending Merle Beach school, and in
1930 there were twenty -eight children of school age living in the district, with
only thirteen of them attending the school. Clearly as time progressed, fewer and
fewer children attended the school. Falling enrollments may represent the

increasing urbanization of the United States as the Twentieth Century

19

progressed. Decreases in rural school enrollments may also be reflective of a
trend toward smaller families as farm machinery became cheaper and more
readily available, requiring fewer people to operate a farm. These hypotheses
require further study in order to determine the exact reasons for ever-decreasing
rural school enrollments in the early Twentieth Century, but turn the investigation

in the direction of the fate of the Merle Beach School in particular.

1. 6. The Fate of the Merle Beach School

In 1898 a resort/hotel was built on Muskrat Lake and named the Merle
Beach Resort for the Marl, Merle to the locals, a natural fertilizer substance,
found on the lake’s shore. According to the sChool supervisor's reports from
1880-1900, the number of school age children in the district decreased drastically
from fifty-seven children to only seventeen. This is a surprising trend, as the
resort should have provided a large influx of local jobs, which should have
allowed more people to start families. On the other hand, the increased number
of jobs from the resort may have attracted either young unmarried people from
outside of the community or within it, and neither group tended to have many
children. Enrollments declined as the number of school age children declined
during these years. The enrollment at Merle Beach school fell from eighty-six in
1880 to eleven in 1900.

From 1900 to 1920 the number of school age children rose from
seventeen to thirty-five, so the increased jobs may have had some impact. The
number of school age children enrolled at Merle Beach school increased at this
time from eleven to twenty, almost double. The number of school age children
decreased between 1920 and 1930 from thirty-five to twenty-eight as the

Depression began. School enrollments at Merle Beach decreased during this

20

time as well, going from twenty to thirteen. In the 1920’s the Merle Beach Hotel
and Resort became largely a public dance hall and began to wane as a resort
(Anonymous B 1980: 511). The shift in function of the resort may have reduced
the number of jobs at the resort and consequently may be reflected in the falling
number of school age children.

The schoolhouses in Olive Township all suffered the same sort of
enrollment attrition rate that Merle Beach suffered. Under consolidation laws
passed in 1964, the schools began consolidation with the St. John’s school
district (Anonymous B: 509). Merle Beach school closed its doors in June of 1966
(Terry Shaffer, personal communication 1990).

The mral one-room school had become increasingly common as settlers
moved westward across the newly expanding western territories of the growing
United States. They allowed children living in sparsely settled rural areas to get a
least the basics of reading, writing and arithmetic. The schoolhouse also served
as a meeting hall and town center for the fledgling communities. As the
Nineteenth Century progressed cities became larger. The sheer numbers of
Children to be educated became too much for the one-room school model and
children were broken into in graded schools, divided by their ages. Educational
reformers saw this as a better way to ensure a broader and more formal
education for children. By the turn of the century, rural schools were forced to
compete and were often required to meet minimal standards for curriculum by
law. Rural schools increased the number of days that school was held,
introduced an increased number of subjects and introduced new courses in

taming, home economics for the girls, and mrpentry for the boys.

21

But more and more people were being attracted to the cities, and rural
numbers dwindled. Fewer and fewer children attended each rural one-room
school and then they began to combine at the county level and consolidate.

By comparison to the other rural schools of the Nineteenth and early
Twentieth Centuries, Merle Beach was a very typical school with increasingly
better teachers and curriculum, and longer periods that school was in session.
Merle Beach school stayed in use far longer than most one-room schools of its
day and was not consolidated until the mid-19605.

Turning from the specific history Of Merle Beach school, examining the
archaeological record, what types of data are likely to be evident on the site of a
rural schoolhouse from the Nineteenth Century? What activities could be
reconstructed? What types of social data are the archaeological remains of a

schoolhouse likely to provide?

1. 7. Types of Potential Archaeological Data in the Context of a One-Room
Schoolhouse

When considering the Merle Beach schoolhouse site, we need to consider
the types of data that are likely to be found when archaeologically investigating a
school from the Nineteenth Century. The first consideration must be what types
of artifactual remains are likely to be common. Are these artifacts likely to be in
high or low frequency? Accounts of typical one-room schools suggest that these
schools were usually sparsely fumished and had few small pieces of equipment
that would have been used and discarded. This account of the furnishings of a
schoolroom by a teacher in 1913 (Culter and Stone 1913: 203) is a good
example of a very well equipped rural school during the late Nineteenth and early

Twentieth Centuries:

22

There were seven fine pictures on the walls, a large case of maps,
a globe, a dictionary, liquid and dry measures, a cupboard for seat-
work material and tools, a bookcase with three hundred volumes,
an organ, a teacher’s table, two chairs, and a blackened stove.

Also included were a number of student desks, a wall of blackboards, a water
pail, and assorted pencils, pens, ink, paper and coal to fuel the stove (Otwell
1923: 88). Most rural schools were not that elaborate. Things such as paper,
pens, ink, and school books were usually purchased by each student’s parents
for their child. The sparseness of the inventories of Nineteenth Century
schoolhouses would seem to indicate a low probability of potential artifacts. Rural
children were not likely to be from wealthy homes so the number of objects that
they would have carried around with them would have been very few. Typically
buttons, pencil pieces, slate pencil pieces, and perhaps a penny would have
been common small objects likely to have been lost and not retrieved, entering
into the archaeological record (Schiffer 1976: 32-33).

The grounds of schoolhouses in the Nineteenth Century usually contained
two or three outbuildings, usually one outhouse for each sex and a building for
storage of yard and maintenance equipment and stove fuel (Otwell 1923: 88).
Some schoolhouses also had gardens for demonstrating and teaching agriculture
(Foght 1910: 179-203). A better glimpse of what a typical Michigan schoolhouse
and its grounds would include can be found in Card (1919: 41-42), which gives
the following description for a “Standard School”:

Yard and Outbuildings

1. Ample ground of at least one-half acre.

2. Some trees and shrubs.

3. Good approaches to the house.

4. Two well kept, widely separated outhouses.
5. Convenient fuel house.

23

The Schoolhouse
1. House well built, in good repair and painted.
2. Good foundation.
3. Well lighted.
4. Attractive interior decorations.
5. Good blackboards, some suitable for small children.
6. Heated with room heater and ventilator in corner, or

basement furnace which brings clean air through the
furnace and removes foul air from room.
7. Hardwood floors

Furnishings and Supplies

1. Desks suitable for children of all ages, properly placed.

2. Good teacher’s desk.

3. Good book cases.

4. A good collection of juvenile books suitable as aids to
school work as well as for general reading.

5. Set of good maps, a globe, dictionary, thermometer,
sanitary water supply.

It should be emphasized that many schools were below “Standard School”
requirements and many were above those requirements. One of the implications
of the description of the “Standard School” is that the minimum number of
expected buildings is four, with two privies, most likely located toward the edges
of the property, and a coal or woodshed close to the schoolhouse proper. This
should necessitate some sort of test excavations near the edges of the site and
around the schoolhouse proper, if it can still be located. I
Schoolhouses were frequently remodeled as their needs changed, rather
than being torn down and replaced with new buildings, so some building debris
such as rusted nails and plaster pieces would be likely to have become artifacts.
Rural schoolhouses would also have had a portion of their land set aside and
used as a playground for the children during recess and lunchtime. Playgrounds
often had equipment such as the giant’s stride, or stride swing (see Figure 1),

swings, vaulting horses, teeter totters and play field areas (Culter and Stone

24

1913:112-124). These areas would have been likely to receive a lot of heavy foot
traffic, and small objects would most likely be either carried to the edges of the
play areas, or broken from being stepped on. Artifacts would rarely be found in

the playground areas, and large open areas of low artifact density might be

indicative of such playground areas.

Figure 1:
Sketch of Girls Playing on a Rough Stride Swing

 

 

 

 

25

Different artifact categories would be potentially sensitive to a number of
different social organizational categories, such as wealth, religion, ethnicity, and
gender. In the case of a rural schoolhouse several social categories would not '
really be possible to detect. .

Race would be difficult to determine as most of the children attending a
rural school would have the same types of school clothes, books, pens, pencils,
and small toys, with little to differentiate pupils of different racial backgrounds.
Unlike areas where slavery had existed, Michigan was home to a fairly sizeable
free black population which would not necessarily have used radically different
goods of a poorer quality, as they would have on a Southern plantation area
(Otto 1977: 105-107). As schoolhouses do nOt have a large number of discarded
household goods (Schiffer 1976: 30-31), it would be very difficult to determine
any differences in race among the students.

Ethnicity would be difficult to judge as well, because the most distinctive
markers would be styles of dress, hair, languages spoken at home, and the types
of food the children brought for their lunches (Mueller 1926: 195-230, 234-245).
Clothing would not be expected to be discarded at school, hair styles leave no
remains unless there are human remains (which would not be expected to occur
at a schoolhouse), and food remains would be easily perishable, except possibly
in a privy context. Privies, being small, would be generally hard to locate.

Food remains taken from privy contexts would only be likely to give
percentages of particular types of food remains, and butchering methods and
some general food preparation data. Food waste data might, in turn, lead to a
percentage of individuals using the site with the same food preferences, from
which ethnicity might be possible to tease out, but this is largely speculative (Otto
1977: 104-105).

26

Religion would also be difficult to differentiate archaeologically at a
schoolhouse, as the only markers that would survive would be metal ornaments,
such as crosses, crucifixes, and stars of David. Had any of these religious
symbols been evident in the archaeological remains, it would be difficult to
assess whether these remains related to the schoolhouse’s function as a school
or as a community building, frequently housing church activities (Anonymous A
1880: 496). A mixture of different religious symbols, such as a star of David, a
crucifix, and a plain cross would have indicated loss during school activities, as
these would not be found together if they were the result of a ceremony of a
single religious sect (Schiffer 1976). Small Children would likely have either not
worn such ornaments of faith to school or would have taken very good care of
them, especially if they were made of a precious metal, such as gold or silver.
Most likely these artifacts would have been curated, or cared for very carefully,
as they had spiritual significance (Schiffer 1976:39-40).

Social class indicators might be possible to discem archaeologically at a
schoolhouse. They may consist of such things as silver buttons, instead of horn,
shell, or plastic. China doll’s heads, which were used on expensive dolls, would
be another indicator of a child from a high social class family. Other social class
sensitive artifacts would include imported china with intact and traceable maker’s
marks, and gold or silver jewelry. Unfortunately, as many schools were used for
church and community activities, fancy clothing articles could be the result of
such activities, rather than an indicator that students came from wealthier homes
(Foght 1917: 157).

Many social class sensitive artifacts would be unusual on a school site,
such as fancy, imported ceramics, as there were very few domestic activities

occurring on such sites, and large scale ceramic discard would be unlikely

27

(Schiffer 1976). Fancy jewelry would likely be curated items (Schiffer 1976:
39-40).

Gender could be detected from the toys and toy pieces in an
archaeological context. Broken doll heads or jacks might indicate where girls
played, while pieces of toy soldiers, remnants of baseballs, baseball gloves, or
marbles would be more likely to indicate where boys played (Culter and Stone
1913: 6-7, 22-23, 116-124). In many aspects boy’s and girl’s facilities, such as
cloak rooms, play areas, and privies, were highly segregated, and in schools with
home economics or wood shop Classes these activities were also gender
segregated (Foght 1910: 236-253).

Age differences on a schoolhouse site is potentially more difficult to
determine. archaeologically. The largest markers of a change from childhood to
teen/adult status would have been the wearing of long dresses for women and
long trousers for men (Culter and Stone 1913; Fought 1910). Differences of
dress would be difficult to view archaeologically, as artifacts such as lost buttons
would have few differences between longer and shorter dresses and trousers.

Teenagers would engage in different activities, such as girls stitching and
knitting in their spare time rather than playing jump rope or with jacks. Lost items,
such as pins and sewing needles, would then be indicators of older girls;
however, such small artifacts would be difficult to find archaeologically, and might
not preserve well. Boy’s activities most likely would not change radically, as boys
of all ages tend to be active in organized team games (Culter and Stone 1913:
113-117). Not until the 19505 or so do artifacts such as tampons and sanitary
maxi pads become widely available for use. These, however, would only be
recoverable in a privy context. This would require that privies were in use at a

site as late as the 1950’s and that the pads and tampons had not decomposed,

28

to be recovered. Such a find would indeed be an indicator that teenage girls had
used a schoolhouse, but it is doubtful finding this type of evidence would be
recovered archaeologically.

During the Nineteenth Century it was common for teenagers to stay home
on the farm and begin their adult life of chores, so few older children went to
school (Anonymous C 1961 -1930).

Some artifacts could also be useful in dating strata at a schoolhouse site,
items such as coins, square cut nails (which were in use predominately in the
Nineteenth Century), china with maker’s marks which were used for only a
discrete period of time, and to a lesser degree slate pencils which tended to be
used during the Nineteenth Century, but may have held on longer in poorer areas

where paper was difficult to come by.

1. 8. Conclusions

By examining the history of school legislation and attitudes toward
education, a good deal can be learned about the conditions under which
schoolhouses came to be formed. Local communities became the primary forces
supporting schools. For better or for worse, local fortunes and financial
fluctuations had substantive impact upon the quality of teachers, school facilities,
and the range of subjects taught in a rural schoolhouse. Frequently, small
schools in relatively poor rural areas, such as Merle Beach, had short school
sessions. The main focus of studies was on basic subjects such as reading and
writing, arithmetic, history, and sometimes geography.

Sites, such as historic schoolhouses, present a challenge for
archaeologists. Schoolhouses were small sites with heavy foot traffic, which

would result in artifactual materials being deposited in sparse quantities, at the

29

edges of the highest foot traffic areas. These sites were used for short periods,
and the children attending rural schools would likely have had little in the way of
playthings and belongings to lose or discard regularly; artifacts would likely be
disperse and few in number. Small rural schools, generally did not have teachers
living on the premises, nor did they generally have formal cooking activities,
which would have resulted in the discarding of glass and tableware in sizable
quantities (Schiffer 1978: 30-33). This is in stark contrast to house sites, where
domestic activities and high rates of discard of domestic refuse would have been
quite common (Carillo 1977: 73-79).

Despite the Challenges that schoolhouses present to the archaeologist,
they can still be sources of information on site use, artifacts can indicate the
specific activities that occurred on a site, and such sites can provide some
general information on the Children that attended the school, such as their
gender, age, and socioeconomic status.

With the ground work of placing Merle Beach school in a historical context
in place, with some of the potential archaeological expectations enumerated, the
next step in evaluating the site in an archaeological framework was to develop a
methodology and a sampling strategy which would be likely to reveal activity

patterning.

Chapter 2
METHODOLOGY

2. 0. Introduction

A methodology chapter focuses on the types of research strategies
employed in a study, as well as the ways in which the researcher goes about
collecting the data and why she Chose the ways she went about collecting this
data. VIfithin the present chapter the rationale behind sampling and why a
sampling approach was used in the excavation design at the Merle Beach
schoolhouse site is presented and examined. In addition, several different types
of sampling research designs are examined along with the reasons for applying
some aspects of a particular sampling research design, for example, using a
particular size of sample unit or choosing a particular sample unit selection
process, while discarding others. The use of archival research, and its nature and
limitations, are also examined. Finally, the data analysis methods employed are
discussed and examined.

For archaeology, methodology can be considered to involve the means by
which archaeologists go about conducting their field and laboratory research, the
choices they make in why they are doing what they do, and their reasons for
making those choices. Methodology involves how archaeologists decide what
types of units to use, where to place them, and how deep to excavate them.
Methodology also involves how archaeologists decide how to wash, treat, catalog
artifacts, and in fact decide how to store, handle, and conserve artifacts after
their excavation. In the case of historical archaeology, methodology also includes

what types of archival research are considered appropriate. It helps to determine

3O

31

which archival sources to use, and to decide whether interviews from former
property holders, or participants at the site when the site was in use, are
appropriate.

In sum, methodology covers all of the different decision making aspects of
how archaeologists go about excavating a site, handling the artifacts they find,
and initially interpreting the artifacts they find. Methodology both shapes the
research design and enables the researcher to discard unfeasible research
designs or designs that would be too costly before the money for the research is
spent. Careful attention to research design can prevent costly mistakes in the

field.

2.1. Sampling Research Design

In arriving at the choice of the final sampling research design that was
employed in the study of the Merle Beach School site, a review of the types of
structures and features that would potentially be at the site, their sizes and
properties, and what sort of testing might reveal these features were considered.
Then a review of the sampling literature was undertaken, and a series of
potential subsurface sample unit sizes and sample selection strategies were
examined. Then several types of sample collection units were chosen, along with
several different ways of selecting sample units. Of the selected strategies, it was
hoped, at least one would reveal some sort of patterning across the site, as well
as reveal which sampling method had been more successful at revealing
patterning. The following section outlines the reasoning and literature that were
employed in this study.

When Chenhall (1979: 3) points out that “archaeologists have long
recognized that inferences about human behavior are in faCt based upon small

and sometimes inadequate samples,” he raises the very valid point that no

32

archaeologist can hope to completely excavate a site. Archaeology is a field in
which excavation is always a form of sampling the total potential universe of a
buried site. This is for two reasons. First, the cost of excavation is great enough
that an archaeologist can almost never excavate a large percentage of any site.
In addition, the act of excavation destroys the data base, so that so that all of the
data in a site can never be fully recovered.

Archaeologists have long been looking for better ways to investigate a
site in a systematic, probabilistic way without resorting to large scale
excavations. The earliest attempts at a probabilistic method of sampling the
archeological record include Carl Lloyd’s 1937 study of Ackmen pueblo (Mueller
1974: 1); Kroeber (1916) and Spier (1917) in their Zuni area surveys (Chenhall
1979: 3); and the VIru Valley survey (Ford and WIlley 1949, as cited by Chenhall
1979: 3). Probabilistic sampling “guides the selection of cases to be investigated,
and does so in such a manner that biases are minimized” (Redman 1973: 63), so
that rather than investigators relying just on their judgment or biases, a
percentage of the total area is Chosen for study in some more objective manner.
Binford (1964: 427) adds “sampling does not mean the mere substitution of a
partial coverage for a total coverage. It is the science of controlling and
measuring the reliability of information through the theory of probability.”
Probabilistic sampling has been successfully carried out on areas as large as
whole river valleys, such as the Chaco Canyon survey (Judge 1972, as cited by
Judge, Ebert and Hitchcock 1979: 92-123) and also as small as sampling areas
within a site such the sampling program at Ft. Johnson (South and WIdmar
1977)

The Merle Beach school site presented some interesting challenges as a
site for sampling. It is a small site, measuring only 165 feet by 125 feet with the

schoolhouse itself extant, as well as a piece of playground equipment and a flag

33

pole remaining on the site. Since the site is small, and much of it may have been
used as a playground, artifact distributions may be spotty and heavily dispersed
by heavy foot traffic. With the seantiness of potential artifactual remains, some
types of data collection methods would be better than others for recovering
widely distributed, low frequency artifacts on relatively small sites. With large
portions of a site having potentially very low numbers of artifacts, which would be
scattered broadly over the site, knowing where to intensively excavate would be
almost impossible, and the time and financial resources that it would require
make large scale, intensive excavation a highly ineffective investigation
technique for this type of site. It would seem that some sort of a systematic,
probabilistic sampling strategy would be a more sensible approach. Sampling is
usually faster, covers the whole area in some way so that patterns can be
determined in number and type of artifacts uncovered, and can be used to point

out areas where additional, more concentrated excavation might be worthwhile.

2. 2. Sample Collection Units and Sample Unit Selection

One way of sampling a site requires no subsurface excavation: surface
collection. Surface collection can be done on nearly all sites, assuming that the
ground cover is relatively sparse. As Lightfoot (1986: 485) states, “where
optimal conditions prevail, the most common discovery procedure is the
pedestrian surface survey, a technique that involves the systematic inspection of
the ground surface at a given level of intensity,” or rather where a patch of
ground is walked over systematically by a particular number of field workers per
unit area.

While intensive surface collection is most frequently performed in areas of
dry, sparse vegetation, it can be performed, though with much less reliable

results, in areas with short cut grass or low lying ground cover. Ideally, where the

34

ground has a fair amount of vegetation, the best way to surface collect is to
scrape the grass layer off and then surface collect, an example of which is the
1968 study by Binford at the Hatchery West site in Illinois (Flannery 1976: 52-54).
Clearly, the technique has more benefits in sparsely vegetated climates, but
since no site can be entirely determined by its surface remains, the technique
can still be used, with significantly reduced results, in areas with uniform, low
level ground cover.

One benefit of intensive surface collection is that a fairly small site can be
covered rather completely in a short period of time. Another advantage is that the
site can be covered completely if the site is rather small, which is even better
than trying to draw a representative sample from the area of the site. The
technique is, however, flexible enough that a very large site can be sampled and
those units selected can be intensively surface collected, as in the case of the
studies done at CayOni'J Tepesi and Girik-i-Haciyan, Turkey, performed by
Charles Redman and Patty Jo Watson in 1968 (Flannery 1976: 54). Additionally,
intensive surface collection does not require a highly trained field crew or special
field crew training, above and beyond showing the crew members the types of
artifacts that they are likely to find. Intensive surface collection has a great deal of
flexibility and can be done with fairly minimal effort and cost.

Intensive surface collection, despite its many benefits, has some
significant drawbacks. First, it is only most effective when the land surface is bare
or nearly so, or where a significant amount of disturbance has already occurred
and peeling the vegetation off of the surface will not unduly disturb the spatial
placement of artifacts. Secondly, intensive surface collection is, for larger sites,
time intensive and requires a fairly large field crew. While not expensive in terms
of needing large outlays of money for expensive equipment and large work crews

on average and small sites, it can be labor intensive performing it on larger sites.

35

Also, modern artifacts tend to build up on the surface of undisturbed sites and
can obscure the patterning of older deposition or give the mistaken impression
that there was only modern use of a site. The final drawback is that soil erosion
patterns, animal soil disturbances, and human digging, construction and mound
building activities can make artifact patterning misleading, but this is true of all
archaeological recovery techniques. The technique of intensive surface collection
clearly has a number of advantages over subsurface excavation techniques, as
well as some drawbacks.

Subsurface testing refers to methods that all involve excavation of some
kind. Subsurface sampling is invasive and destructive by its nature, for once a
structure or site is excavated the primary data is lost for future archaeologists.
There are many different subsurface sample collection units that can be
employedduring sampling. Those that will be examined here include shovel
testing, core boring/posthole testing, and test pitting.

The first subsurface sample collection unit to be examined is shovel
testing, in which a shovel is used to dig into the soil and the soil is either
observed for artifacts or is sifted through a screen to detect artifacts. Usually
these tests are done in lines or transects at some standard interval. Shovel
testing results in a wide, shallow test hole about a foot in diameter. This is a
method that has primarily been used for “regional survey in places where the
ground surface is covered by vegetation, prohibiting its direct examination” (Shott
19852457)

The advantages of shovel testing are many. It is quick, requiring only two
people to quickly stake out several rows with measuring tapes and chaining pins.
The samples can then be excavated, one row to a field member with only a
shovel, a Sifter screen and a few artifact bags. Shovel testing needs little in the

way of special equipment, requiring only a few sifter screens, shovels, measuring

36

tapes and stakes or chaining pins. Special training of field crew members is not
required as no special equipment is used. Shovel testing can be done quite
effectively with a very small field crew of two to four and still cover a reasonable
area. It an also be effective where there is significant ground cover such as tall
grass, dense forest, or low-lying shrub growth. Despite the many advantages of
shovel testing, there are also many drawbacks inherent in the technique.

While shovel-testing has been in use in archaeology since at least 1976,
it has been under great scrutiny and criticism since the mid 1980’s. Shovel-
testing has been argued to be poor at locating sites (Shott 1985 and 1989). It has
never really been used to define intrasite patterning, since the samples are wide
and shallow shovel -tests do not necessarily accurately detect natural
stratigraphy. After examining a wide body of the literature debating the viability of
shovel-teSting (Lightfoot 1989; Nance and Ball 1989; Shott1985, 1989), it
appeared that shovel-testing would not be an adequately sensitive testing
method and seemed better suited for presence/absence data than predicting
concentrations of artifacts that would indicate features and structures on the
Merle Beach school site. It therefore was not employed in this study.

Another subsurface sample collection unit that is often employed is core
sampling. Core sampling involves boring into the soil with an auger, a posthole
digger or other boring tool to remove a column of soil. The soil is then observed
to determine the stratigraphic layers and then, typically, screened for artifacts.
This is done with either a small hand boring device with a diameter a few inches
across or with a larger mechanical boring device with a six inch or larger
diameter.

The advantages of using core boring techniques are several. First, the
cores obtained provide “minimally disturbed section[s] of subsurface materials”

(Stein 1986: 505), which means that cores can be used to distinguish between

37

culturally relevant strata and natural strata; next, cores can, if spread regularly
over the site surface, indicate the horizontal extent of any cultural deposits (Stein
1986: 513). Also, cores can be used to determine the subsurface depths and
outlines Of features (Stein 1986: 513). In addition, core samples can be analyzed
using the same techniques as column samples, so very small materials, such as
seeds and pollen, can be collected as well as larger artifacts. Finally, cores can,
depending on the type of boring equipment used, be used to sample sites of
nearly any depth, and of almost any soil composition and in almost any type of
environment (Stein 1986: 511- 512). As Schuldenrein points out (1991: 132),
coring is very versatile, and can be used as a site discovery method, as a
method of testing a known site, and as a means of extracting specific
stratigraphic data from areas or features within a site.

Coring also has a number of limitations. Firstly, observations are limited by
the type of coring tool that is selected. With a small bore coring tool, such as a
1% “ diameter hand auger, the data is limited to observing stratigraphy, in a very
narrow area, which may be useful for studying features but which will be harder
to use for site discovery. A 3” or 4” diameter mechanical drill or auger can yield
significantly larger samples which can reveal artifacts, a more in-depth analysis
of stratigraphy, or be used as a column sample (Stein 1986: 512-13;
Schuldenrein 1991: 134). Some types of augers also force some of the soil into
their hollow handles, and these portions are lost for research. Posthole diggers
have also been employed as augers, but they have a number of drawbacks
including “possible contamination because sediments are churned by the device
rather than extracted in consolidated form,” and posthole diggers can only reach
sediment depths of “50-1250m below the surface,” making them inappropriate for

deep cultural deposits, or where deep deposits are expected (Stein 1986: 517).

38

Posthole diggers used in place of core boring devices, however, have a
number of benefits to their use. One benefit is that they have a larger bore size
than a hand auger, so they can provide a better view of soil profiles. While they
do not bring up completely intact soil stratum, carefully used to penetrate a
stratum at a time, they can achieve a more intact profile than a shovel test.
Posthole tests can also probe deeper than shovel tests, although not as deep as
a hand or power auger. Posthole testing can also be accomplished economically,
and a hand posthole digger can be used in soils difficult for power driven
equipment to handle, or in tight areas, or in areas with difficult topography (e.g.,
the sides of steep hills), where power driven equipment is not practical to use
(Stein 1986: 517). Posthole diggers also require less specialized training of field
crews than power driven equipment, which makes it easier to find labor.
Mechanical coring devices can also be expensive, with the most advanced
costing from $500-$1000 dollars per day to rent both the machine and operator,
with each of these cores taking one to three days to obtain (Stein 1986: 516). A
large number of sites have been sampled using posthole tests from as early as
1947 (South and VVIdmer 1977: 128).

Then there is the sample collection unit of the square test pit. Test pitting
“can provide information on the composition and the stratification of the site,
locate areas of activity or especially rich deposits within it” (Hestor et al.
1975:.71-2). Test pitting can be used as a final step in the survey of an area,
used as a method of excavation or used as a preliminary step in excavating a
large area. Test pits are always square and tend to be uniformly one meter
square, if the site is excavated metrically, and between three and five feet
square, if excavated in English units, as some historic sites are excavated. The
reason for the relative standardization and small range of variability in test pit

size is that “squares much smaller would squeeze out the archaeologists, and

39

larger units are overly destructive (and too time-consuming)” (Thomas 1991:
108)

David Hurst Thomas (1991: 108) notes, in test pits, “archaeologists must
maintain control of the finds the x axis (front to back), the y axis (side to side),
and the z axis (top to bottom). This is why archeologists dig square holes.” In
other words, square test pits allow archaeologists to place an artifact in a
bounded three dimensional space, which allows the researchers to observe the
patterns in distribution across test pits in order to detect floors and activity areas.
Artifacts may be located in time, relative to each other, by observing their location
in the z axis, assuming undisturbed stratigraphy.

The advantages of test pits include the fact that they provide a larger
exposure of the site than a shovel test or posthole test can. Secondly, posthole
tests can allow for much greater stratigraphic control, by allowing the
archeologist to peel layers off in an area large enough to maneuver, but not too
large. Thirdly, test pits can be strung together to observe a large area of a site or
to focus on a feature. One disadvantage is that test pits, because Of their
relatively small size when compared with larger area exposures, cannot show
large features such as structures, except for small portions of such features,
unless several test units are strung together. Another disadvantage is that test
pits take more time and cost more than either shovel testing or posthole testing.

After examining all of the different sample collection units outlined above,
looking at their strengths and weaknesses, I decided upon a research design that
combined surface collecting and two different subsurface sample collection units
for the excavation and analysis of the Merle Beach school site. For the surface
collection I decided to divide the entire site surface into a number of large units to

intensively surface collect, after the site was Closely mowed. I decided to follow

40

the surface collection with a series of posthole samples selected through a
uniform systematic arrangement of lines, or transects, across the site.

Systematic unit selection was used for the posthole samples for a number
of reasons. A systematic unit arrangement allows for uniform subsurface testing
of a site; however, the regular placement of units may cause regularly placed
features to be missed if the feature regularity causes features to fall between the
test units (Plog 1976: 139-143). In an effort to lessen this common pitfall, l
staggered half of the transects, in effect creating not square arrangements of
posthole tests but diamond shaped arrangements. The other reason for using a
systematic arrangement of postholes is that, in some instances, systematically
chosen units can be better at observing site structure than randomly chosen units
(Plog 1976: 141-143).

Then I elected to use a random selection to choose a number of five foot
by five foot square units. A random selection method was chosen, as little was
known archaeologically about schoolhouse sites, so stratification of the area was
impossible. Also, five foot by five foot test pits are rather large sized units to
excavate, and only a limited number of them could be excavated, so without a
way of stratifying the site, random selection of test pits seemed the least biased
method of test pit selection.

Surface collection was chosen to cover as much of the total surface area
of the site as possible and to see whether surface remains had any correlation to
patterns found through subsurface testing. Posthole testing was chosen as a
compromise between mechanical augering and shovel testing, since posthole
testing allows for more control in observing and preserving the natural
stratigraphy during excavation than shovel testing, and is less costly and requires

less expertise of field crew than mechanical augering.

41

Five foot by five foot test units were selected to provide a number of wide
exposure excavations in order to observe site patterning of artifacts and to
provide a larger excavation unit that could then be compared to the posthole
tests, in order to see if the posthole tests revealed similar patterning. The design
was selected with the idea that all of the tests could be compared to see if one
testing method revealed similar patterns that were borne out by another testing
method, in an attempt to compare the effectiveness of the different sampling
methods. The research design for the sampling program as a whole was largely
based upon South and Widmar (1977: 128-131) and Plog (1976: 137-144). A full
formal description of the sampling research design that was chosen and the
number of units used follows.

The Merle Beach School Site measures 165 feet north to south by 125
feet east to west, and has been divided up into 25 foot by 25 foot square units. A
15 foot wide strip running the length of the site from east to west along the
southern site boundary was excluded to make the units easier to handle. This
section of the site was chosen for exclusion because it is not known to have
played an active part of the site’s use historically and because it contains the
remains of a pet belonging to the owner. The site also contains the schoolhouse,
which is still standing, along with a piece of remaining playground equipment and
several trees, most of which are located around the perimeter of the site (see

map 2 page x of appendix B).

2.2.1. Phase 1: Non-invasive techniques
Surface collecting. Each 25’ x 25’ square was walked and the artifacts
collected. This yielded a sample of 100 percent of the total site area.

42

2.2.2. Phase 2: Invasive techniques

Part 1: Posthole testing. One posthole core 12 inches in depth was drilled
every 12.5 feet in transects 12.5 feet apart in an east-west direction across the
site. The posthole tests were offset by 12.5 feet in half of the transects in order to
minimize symmetrical deposits that could be missed by a completely regular
arrangement of posthole placement. This yielded 68 tests, or .0078 percent of
the total site surface.

Part 2: A series of nine 5 foot by 5 foot pits was excavated to a depth of
1.5 feet. Each test pit was randomly Chosen within its 25 foot square (excluding
large features, such as the schoolhouse and the flagpole). This yielded a 1.09
percent sample of the total site area.

Overall, 100 percent of the site received some sort of testing.

Approximately, 1.10 percent of the site received subsurface testing of some sort.

2. 3. Archival Research

Archival research is a bit different in terms of methodology and requires
different considerations than those applied to field and lab methods. First, the
types of data are very different. Archives are usually repositories of old records,
diaries, histories, and other documents. Usually, they are built up through the
donations of private families, contributing their old family diaries, letters, and
scrap books. Histories are usually of local towns or counties and were often
commissions for community anniversaries or other big events. Old records are
frequently donated to archives when local town halls, county halls of records, or
state bureaus run out of space for the storage of old records. Because nearly all
of the materials in an archive is donated material, it is frequently a fragmented

collection of different things.

43

In archaeology this hodgepodge of records presents a number of
challenges. Frequently, direct records of the site in question, .such as deeds of
sale, wills of previous landowners, diaries of previous landowners or users, or
account books of site users/owners, may exist; then again, none of these may
have been created or kept, or documents may have been lost or burned during a
court move or building fire. For public types of buildings and activities, there are
usually state, county, or municipal records, but these may have been kept
sporadically, not filled out completely, or available for only short periods of the
time that a site was occupied. There may also be little that documents a site
directly, but there may be a great deal of indirect references, requiring a good
deal of intense research in order to establish accurate connections. Finally, with
archival sources, the viewpoints of the individuals and the time periods they were
writing in may need to be taken into consideration, and the information obtained
from sources culled through carefully. Additionally, since many things, like local
histories, may have been used to highlight the most wealthy citizens in the area
or may have been written as a public civic pride project, they may be inaccurate
or not match other sources. Careful selection of archival sources and the
selections that are taken from them, therefore, becomes very important.

In assessing what sorts of archival research to undertake for this study, a
number of things had to be taken into consideration. As far as is known, none of
the teachers who taught at Merle Beach school left any memoirs that had been
entered into any of the local archives. However, a great deal could still be found
out about the school, since the state required annual school supervisors, reports
starting in 1860 and ending in1931 (Anonymous C). These reports covered most
of the time period that the school was in use. County histories, both those written
during the Nineteenth Century (Anonymous A) and one written fairly recently by

the local county historical society (Anonymous 8), were able to give background

44

about the area, how the schoolhouse was used, and the types of events and
activities it was used for. They also gave the social and local business profile for
the area around the schoolhouse, providing a glimpse at the rise and fall of the

community.

2. 4. Data Analysis Methods

Not only is the sampling research design used in artifact collection
important, but so is how the material is handled upon collection and afterward. All
of the artifacts from each surface collection unit were bagged and marked with
the date collected, the type of unit,*and the initials of the collector. As each level
of each sample unit was excavated, it was screened through V4” mesh screen.
The resulting artifacts were then bagged and labeled with the date, provenience,
and excavator’s initials. The bags were then given a field number and were
counted and logged in the field catalog. Artifacts were then washed a bag at a
time, each bag being noted in the washing log. The artifacts were then counted
once again and the final counts were recorded. For the items where enormous
amounts of materials were collected, such as unburned coal or brick, these items
were counted, weighed, and discarded. The final data tables can be found on
page y of the appendix.

A series of statistical tests was then run on the raw data. These tests
included a three dimensional contour map for each artifact type for each data set,
a Pearson correlation matrix for each data set, a three cluster K-means cluster
analysis for the posthole and the five foot unit data sets, and a five cluster K-
means cluster analysis for the five foot unit data. The three data sets were
compared to test sampling effectiveness using exploratory data analyses

methods, employing box and whisker diagrams.

45

2. 5. Conclusions

The focus of study at the Merle Beach school site was two fold: to both
archaeologically examine a Nineteenth Century rural school site and examine the
effectiveness of a number of different sample strategies used in archaeology on
this type of site. To accomplish both of these ends, the sample research design
outlined in this chapter and the analysis methods employed were designed to
address both issues. The use of a series of different sampling strategies,
followed by a statistical comparison of their results, was meant to allow the
relative effectiveness of each strategy to be determined. The sample research
design was also chosen to increase the likelihood of uncovering the activities that
occurred at the site and the areas in which these activities occurred. The sample
research design that was employed was also Chosen to cover as much of the
surface area of the site as possible in order to increase the probability that any
subsurface features would be discovered. The study as a whole strives to
understand the activities that occurred at a rural school in use during the
Nineteenth Century, taking the Merle Beach school as its example, and to
develop methods by which many small, poorly understood sites, such as
schoolhouses, can be studied archaeologically in productive, and cost effective

ways.

Chapter 3
DATA ANALYSIS AND INTERPRETATIONS

3. 0. Introduction

A series of different sampling research designs, each incorporating a
different sized sample collection unit (with these units chosen in a different
manner) were excavated at the Merle Beach schoolhouse. Then each sample
research design was evaluated utilizing statistical analyses to determine its
effectiveness at locating artifacts and activity patterns across the site. The three
different sampling research designs were then compared statistically, to
determine that sampling research design which was most effective at

determining artifact patterning across the site.

3.1. Spatial Organization: Surface Collection Data

The surface collection consisted of 30 units of 25’ by 25’, 625 sq. ft. each,
which were walked by a field crew member who collected all objects found within
each area, after the site had been mowed. Artifacts were found in only 15, or
half, of the 30 collection units. The highest percentages of artifacts (those units
containing five percent or higher of the total artifacts) were from units 24, 25 and
30, which were Clustered around the schoolhouse in the northeast comer of the
site (see map 2, p. 121). The highest percentage of the total artifacts within this
test was unit 30, which accounted for 56.12% of the total artifacts recovered in

the surface collection.

46

47

Several statistical analyses were then performed on the surface collection
data. The first stage involved constructing a contour map of each artifact variable
plotted against the xy—coordinates of the surface collection units, with the intent of
displaying the density of artifacts as they were spread out spatially across the
site area (maps 3-6, pp. 122-128), similar to the technique employed at the Mask
site (Whallon 1984: 249-253). The second stage was to assemble a frequency
distribution and two graphs (graphs 1, p. 157) of the distribution, in order to
observe the shape of the distribution curve and any sharp changes or outliers.
This procedure revealed that the bulk of the units had artifact frequencies
clustered around zero, with a few outliers, indicating that most units had a small
number of artifacts. Finally, a Pearson’s correlation matrix was run to determine
any relationships between the artifacts that may not have become evident
through visual comparison of the multiple contour maps. This is also similar to
Whallon’s treatment of the Mask site (Whallon 1984: 250-251 ).

The resultant contour maps revealed clustered patches of a few artifacts
in one or two areas of the site. The spatial patterning revealed for the artifact
categories of bottle glass, metal, flat glass, square headed nails, wire nails, and
unburned coal were all nearly identical, with very high concentrations of artifacts
around and just to the west of the schoolhouse. Plaster and shingle were all
aggregated in one area centering in the northeastem-most area of the site
behind the school house. Wood has a small clustering along the east fence row
directly behind the schoolhouse, not far from several trees along the fence row. A
small cluster of bone was located in the far southeast corner of the site, along the
east fence row. Mixed small artifacts also tended to Cluster along the fence row,
between the Clusters of bone and wood.

The artifact category Other Mixed Ceramics has a rather unusual contour

map, with two sizable Clusters. One is directly in the region around the

48

schoolhouse, similar to the distribution of metal, and the other occurs toward the
south end of the site about twenty feet from its southern edge. The contour map
distributions for slate display the highest concentrations in what would have been
the playground area. Concrete also occurs in the same spatial patterning as
slate, with one discrete cluster in the center of the playground area. Plastic has
the most unusual contour map, with two small clusters. The largest cluster of
plastic pieces occurred in the northeast corner of the schoolhouse, in the same
area as the highest concentrations of shingle and plaster.

The Pearson’s correlation matrix resulting from the surface collection
showed strong positive correlations between metal, wire nails, square headed
nails, and unburned coal (r values of 1.0). Weaker associations were found
between flat glass, square and wire nails, and unburned coal, as well as between
other mixed ceramics and square and wire nails. These correlations reinforce the
results of the contour mapping, with many of the same‘pattems emerging in the

surface collection data.

3. 2. Interpretations of the Spatial Organization of the Surface Collection Date
That the spatial patterning for bottle glass, metal, flat glass, square
headed nails, and unburned coal were all very similar, with concentrations in the
same area right around and just to the west of the schoolhouse, would suggest a
similarity of depositional patterns of activities at the schoolhouse. This is probably
linked to a combination of two factors. One of them involves the reconstruction
activities of the landowner that occurred shortly before the excavation of the site
(Terry Shaffer, personal communication 1991 ). Secondly, the close scattering of
artifacts such as metal and unburned coal, which would have been used when

the schoolhouse was in use, suggests that small amounts of commonly used

49

materials fell directly around the schoolhouse, possibly as a result of regular
sweeping and maintenance by teachers and cleaning personnel.

As all of the plaster and shingle were clustered into one area behind the
schoolhouse, it may possibly indicate that plasterboard and shingles were briefly
stored in this area, during a relatively recent remodeling phase of the
schoolhouse. Since all of the wood remains come from the fence row, near still
extant trees, the surface wood probably comes from modern shed tree bark and
has no cultural significance. With bone remains located in only one small area
close to the fence, the bone may represent the residue of modern small animal
deaths or, more likely, the meals of local farm cats or other small animals. Given
the distance from the schoolhouse, the bone most likely is not directly related to
activities at the schoolhouse. The other mixed small artifacts Clustered between
the bone and wood deposits are probably small bits of modern trash that have
blown onto the site, especially since they are near the fence row and don’t
appear to be objects that would be directly related to the school, such as bits of
rubber.

As the artifact category Other Mixed Ceramics Clustered in two different
regions of the site, this may either indicate different types of ceramic being used
at two different times in the site’s history, or that ceramics were discarded in two
areas of the site. Some of the ceramics on the south side of the site are most
likely drainage tiles, based on the dark brown glazed interior, their lighter glazed
brown exterior, and the red terracotta paste of the sherds. Since the school did
put in a septic field in the southwest corner of the site in the 1950s, the clay drain
tile pieces may date from this period (Terry Shaffer, personal communication,
1990).

The slate was Clustered in the middle of what would have been the

playground area, which could indicate that children either took their slates

50

outside and used them to play games with, or that modern children, just before
the close of the school, used found slate pieces as toys of some sort. It is highly
unusual to find small artifacts, such as slate pieces, lying in an area that was
open to heavy foot traffic. The pieces of slate may simply not be indicative of a
specific organized activity but may simply have landed where they happened to
fall from a broken slate, or a teacher sweeping the floor, rather than indicating a
particular use of slate in that particular area.

Concrete had the same patterning as slate. As small chips of cement may
have flaked off of the cement porch in the last years of the school, it is again
possible that the children at the school may have picked them up and used them
as toys, or that cement was mixed there during reconstruction of the school.
Finally, plastic has its highest concentrations in the same area as shingle and
plaster, reinforcing the possible interpretation that building materials were placed

here briefly while the school was in a phase of reconstruction.

3. 3. Spatial Organization: Posthole Data

The sample chosen for posthole testing consisted of a series of 70
postholes. The locations of the postholes were plotted along east-west transects
spaced 12.5 feet apart, with tests placed every 12.5 feet along each transect.
Two of the selected postholes could not be excavated as they fell inside of the
foundation of the schoolhouse. The transects were numbered from south to
north. All of the postholes were excavated in layers by visual soil changes, and
were screened. All of the resultant artifacts were bagged and counted. Data
Table 2 (Appendix) shows the level counts for all posthole test pits. Of the 68 pits
excavated, only 13 yielded artifacts, with the highest percentage of artifacts
coming from 10 of the test pits. Each of these pits had one or more levels which

accounted for 4 or more percent of the total artifacts collected during the

51

excavations. The highest percentage of the total artifacts recovered in this phase
came from the units N1077.5 E1000IB level (4.97%), N1077.5 E1125lA level
(13.82%), N1090 E1112.5IA level (11.23%), and N1102.5 E1125IA level
(11.02%). These units cluster in two areas, with one around the rear back corner
of the schoolhouse, and another cluster in the front of the schoolhouse (map 2, p.
121)

As with the surface collection materials, the posthole tests were first
analyzed by plotting a series of contour maps for each artifact (maps 17-42,

p. 129-141). Graphs of the frequency distribution (graph 3, p. 158) were
observed for the shape and nature of the distribution, which revealed that most of
the units clustered around zero with a few outliers, indicating small quantities of
artifacts in most units.

Unlike the surface collection artifacts, which displayed a large number of
Similar patterns of spatial distribution, the artifacts obtained from the posthole
tests had few artifacts that shared similar contour patterns. Chert flakes had one
cluster that overlapped the square headed nails, located on the south side of the
school, but also had a cluster further to the west in the middle of the playground
area. Burned coal, unburned coal, screws, and walnut Shells all clustered in the
northwest comer of the site, closest to the road. This location happened to be
near what appeared to be a nut tree, so the walnuts are probably modern and
have no bearing on the interpretation of the site as a whole. Bottle glass and flat
glass and the other mixed artifacts occur in the same area of the site, in the
northeast corner along the fence row. About ten feet to the west a cluster of
glazed white ware occurs in a similar pattern to the cluster of barbed wire.
Pennies and wood clustered in the same pattern to the east and slightly south of
the schoolhouse. Metal, brick, shotgun shell casings, and plaster all clustered in

a similar manner along the east fence row behind the schoolhouse.

52

Chalk and charcoal appear to cluster together at the far north of the site
along the north fence row. Slate and concrete also appear to cluster together just
to the south of the school.

A Pearson’s correlation matrix of the posthole data did little to clarify the
patterning observed through the contour maps. The analysis of the Pearson
correlation analysis run on the posthole tests showed very few strong
correlations. There were moderate correlations (between .66 and .70) between
plaster and screws, brick and shotgun shell casings, and between bottle glass
and flat glass. The only fairly strong association (.944) was between slate and
chalk, which would make sense for a school site, as they are related in their use
within the school room. The Pearson’s correlation results again largely reinforce
the spatial patterns observed by the contour maps, with the exception of the
slate/chalk correlation, which did not appear in the comparison of the contour
maps of these artifacts. Also, screws did not appear to Cluster with brick, shotgun
shell casings, brick, and plaster in the comparison of the contour maps. The
Pearson’s correlations may have been off, since the use of the correlation matrix
is highly unreliable if the relationship between the variables is non-linear and if
the distribution is not symmetrical and can yield inaccurate results (Speth and
Johnson 1976: 36). Also the data set included a number of zero cells which can
“drastically alter the magnitude of a correlation coefficient, and even transform a
statistimlly significant negative correlation into a positive value” (Hesse 1971, as
cited by Speth and Johnson 1976: 38). With the inconclusive nature of the spatial
patterning data for the posthole test data, another method of statistical analysis
was performed, called the K-means test.

The K-means test is "a non-hierarchical divisive cluster analysis which
attempts to minimize the intracluster variances while maximizing the intercluster

distances” (Ammerman and Kintigh 1982: 39). In other words, it attempts to

53

break a population into clusters without first separating the population into strictly
nested stages, so it can frequently show patterns in raw data from the spatial
clustering that other methods may obscure or overlook. In short, the K-means
test looks for clusters of artifacts whose means all cluster around some common
point or centroid (Johnson and Johnson 1975: 286).

The K-means test was performed with three clusters in an attempt to
tease out some relationships between the artifacts. The first cluster consisted of
all of the artifacts, the second Clustered brick and bottle glass, and the third
cluster paired unburned and burned coal together. The results of the K-means
test simply confirmed the results of the Pearson’s correlation and the contour
maps. The K-means test’s failure to find additional artifact associations points to
a general lack of large scale artifact spatial clustering for the posthole tests.
Probably, “the tests yielded too few and too dispersed artifacts to be able to give

any really meaningful patterning.

3. 4. Interpretations of the Spatial Organization of the Posthole Data

As metal, brick, shotgun shell casings, and plaster are primarily building
materials, and occur in similar clusters, this may have represented a dump for
them when the school was remodeled at one point, but with so few materials
clustering together, this is highly conjectural at best. As most of the artifacts that
clustered together in the posthole data appear to do so primarily along the fence
rows, it is possible that the clusters all represent trash thrown out from the school
or from farmers in the field adjacent to the schoolhouse. There are also several
artifacts which have distinctive contour maps all their own, such as mortar, which
occur independently of the other artifacts. However, the posthole tests did result
in very small amounts of material being found within each test, and the lack of

total material may be part of the reason for the spotty artifactual patterning.

54

3. 5. Spatial Organization: Five Foot by Five Foot Unit Data

A series of eight units, five feet by five feet square, was intensively
excavated to one sterile level, and a ninth unit was partially excavated until
severe rains and flooding forced a halt to the excavation. The dirt from each level
of each unit was sifted through %” mesh screening. Again, levels were
determined by visible changes in the color of the soil. The highest percentage of
the total artifacts in all of the test units occurred in N1160 E1005 (15.01%),
N1110 E1120 (24.54%), N1145 E1080 (38.47%), N1125 E1025 (6.90 %), and
N1055 E1070 (5.27%). These units are located at the front northwesternmost
comer, 25’ north of the schoolhouse, fifty feet in front of the schoolhouse, directly
behind the schoolhouse, and in the middle of the playground (map 2, p. 121).

In comparing the highest concentration of artifacts within each test and
mapping them together (see map 2), the highest artifact concentrations seem to
cluster around the schoolhouse near the porch and directly behind it. There are
also dense concentrations of artifacts toward the north edge of the property, one
of which revealed a building trench, possibly the coal house. There is also,
unexpectedly, one cluster of artifacts in what would have been part Of the
playground.

As with both the surface collection units and the posthole tests, a series of
contour maps was generated for each of the artifact classes excavated from the
five foot by five foot units (maps 43-72, pp. 142-156). A frequency distribution
was also constructed and a graph of it (graph 5, p. 159) observed to determine
its shape and nature, with all of its values clustering around zero with a few
outliers, indicating a low density of artifacts within each test unit. When the
contour maps were compared, five groups of artifacts emerged with very similar
contour maps. Fire cracked rock, chert flakes, wood, square nails and brick all

had the same general spatial pattern, with three areas of high concentration, on

55

the far south side of the site, the northeastern corner of the site, and the far west
side of the site, with the highest concentrations located at the northeast comer.
Unexpectedly, stone tool fragments had a completely different spatial distribution
than fire cracked rock or chert flakes, and had a completely different pattern than
any of the historic artifacts.

A second group of artifacts that had similar contour maps consisted of
wire nails, metal, unburned coal, bottle glass, concrete and bone. This accounts
for the majority of the historic artifacts. These cluster in the northeast of the site,
to the north of the schoolhouse, sweeping right next to it and on the far west
edge of the site, directly next to the road. The footing trench of an outbuilding
which was filled with unburned coal was recovered to the north of the
schoolhouse.

A third group of artifacts that have highly similar contour maps include
sheet metal, barbed wire and charcoal. The clusters of these artifacts occurred
largely in the northwest corner of the site, closest to the road. In the case of
barbed wire specifically, there is also a concentration in the far southeast corner
nearest to the road.

The fourth group of artifacts with highly similar contour maps was
comprised of other mixed ceramics and mollusk shell. These artifacts clustered in
a wide band across the west and south sides and another wide band along the
north-northeast side close to the schoolhouse.

The fifth and final group of artifacts with highly similar contour maps
consists of slate and mortar. These concentrations are primarily located in the
southwest corner of the site, near the road, with a second cluster on the east side
of the site. For slate they are in the same general pattern but in a more narrow

band a little further from the road and further from the east side of the site as

56

well. These artifacts may have been pushed to the edges of the site from one
particular phase of reconstruction of the schoolhouse.

Finally there is a whole Class of artifacts that have distributional contour
map patterns that appear to behave differently than any other artifact groups.
These artifacts include flat glass, glazed white ware ceramic, pearl ware and
transfer printed ceramics, other miscellaneous artifacts, plastic, peach pits, and
walnut shells. The walnut shells are clustered in only one discrete place on the
site, in the far northwest corner in a unit that sits under a nut tree. They are
almost certainly the modern annual fall from the tree. The glazed white ware
ceramics and the pearl and transfer printed ware ceramics had very different
spatial patterns from each other and from the other historic artifacts. The peach
pits had very specific clustering in primarily two areas, the south margin of the
site and the west margin of the site, and a small pocket at the northeast comer of
the site.

After the comparison of the contour maps, a Pearson’s correlation matrix
was performed and the number of correlations was very high, obscuring any
artifact associations. Almost every artifact had a correlation with several artifacts
that displayed an r value of .6 or higher, indicating a moderate to high level of
correlation. Again, Pearson’s correlations can only be reliable with linear
relationships between pairs of variables and with symmetrical distributions,
neither of which case held in this instance (Speth and Johnson 198: 36) Also,
there was one contour map which showed all zero values, which can grossly
distort a Pearson’s correlation (Speth and Johnson 1976: 38). In an attempt to
reduce the noise and pick out useful artifact associations, a K-means test was
performed.

A K-means test was performed in two separate runs, the first with three

Clusters and the second with five clusters. The first Cluster in the first run included

57

fire cracked rock, chert flakes, stone tool fragments, square headed nails,
shotgun shell casings, metal, and barbed wire. This cluster cut across several of
the clusters observed by the inspection of the contour maps. The second cluster
from the first run was composed of just sheet metal, and the third cluster included
just wire nails. In an attempt to get a less confusing and more meaningful
statistical analysis, a second run of the K-means test with five clusters was
performed. The first cluster of the second test included fire cracked rock, stone
tool fragments, square headed nails, shotgun shell casings, and barbed wire.
The second cluster was comprised of only sheet metal; the third cluster, only wire
nails; the fourth cluster, only metal; and the fifth cluster, only stone tool flakes.
These new Clusters were no better than the original K-means test and did not
contribute meaningfully to the delineation of useful artifact associations. The
failure of both the Pearson’s correlation and the K-means test to provide
significant artifact correlations and clusters suggests that the artifact Classes
possibly exhibit independent clustering behavior. This indicates that there were
few specialized activity areas with related artifacts for each use. It points to a
multi-use site with a great deal of flexibility, which would be expected for a
schoolhouse that was frequently was used for community activities.

The difficulty in examining artifact associations for the five foot by five foot
unit tests may be that the large number of different types of artifacts, many in
relatively small numbers, throws off attempts to view associations and clustering

between different artifact types.

3. 6. interpretations of the Five Foot by Five Foot Unit Data
From the materials found in the one pit located in the middle of the
playground, it would appear that these remains represent a small trash deposit.

The artifacts scattered around the schoolhouse are mostly building materials and

58

appear to be trash deposits from remodeling episodes. The unit with a high
concentration of artifacts, located directly next to the road (map 2, p. 121) could
represent either a trash pit from the schoolhouse or refuse from the stage coach
route which ran past the schoolhouse. The postholes with the highest artifact
concentrations are all in close proximity to the five foot by five foot units, with
high artifact concentrations, and are clustered close to the schoolhouse. This
clustering close to the schoolhouse would be expected as small objects, like
buttons, would tend to be lost as children came back to the schoolhouse from
recess and lunch time games. It would also be explainable that a teacher or
janitor sweeping the schoolhouse floor would sweep small items out of the door
and into the grass immediately around the porch.

The coal-filled trench may represent an outbuilding which may have been
the school’s coal house. The wire nails probably indicate that the structure was
from after the turn of the century, when wire nails came into exclusive use
(Nelson 1968). Bottle glass and concrete are also possibly indicative of
remodeling of the schoolhouse at the same time that the shed was erected.

From the location of the clusters of metal and barbed wire, most likely the
metal and barbed wire pieces were being deposited there by farmers, and this is
refuse tossed to the margins of the farmer’s fields, not primarily school material.
As the mollusk shell was primarily buttons and button pieces, it is not surprising
that they would occur close to the school and all along the playground area, as
that is where they were lost as children ran inside from recess or other activities.

The very different patterning of the glazed white ware, pearl ware, and
transfer printed ceramics suggests it is possible that these were remnants from
the schoolhouse’s auxiliary uses as a meeting house rather than as part of its

regular use as a school. As the scattering of peach pits is very discrete and

59

seems almost systematic; these areas may have been where the modern

children ate their lunches and tossed their peach pits.

3. 7. Analysis and Comparison of the Sampling Methods

The first step in analyzing and comparing the three data collection
methods was to establish data frequency tables and inspect the line graphs that
resulted from plotting the data (graphs 1-8, pp. 157-160). Two data frequency
tables were constructed for each of the two data sets (the surface collection data
and the posthole test data). For each data set, one data frequency graph plots
the number of units having artifacts, and the resulting artifact frequency was
divided in grouped intervals, Le. 06, 7-12, etc. The other data frequency graph
for each data set plots the number of artifact classes against the frequency of the
artifact’s distribution, again grouped in intervals. The five foot by five foot unit
data set is treated in the same manner, except that four data distribution graphs
were constructed, two of each type. The second set of data distributions removes
the unburned coal, which occurs in such abundance that is can shift the mean
and standard deviation, making any summary statistics of limited utility.

The second step in analyzing the different data collection methods was to
perform an exploratory analysis of the data frequency distributions. First the
basic statistics of mean, median, and standard deviation were computed for both
the frequency and the artifact variables for each of the distributions as shown in

table 3.

60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 3:

Basic Statistics for Artifact Distributions
Type of testing MEAN MEDIAN STANDARD DEV.
SurfaceCollection 1 .
NO. of artifacts 72.5 72.5 41.413
frequency of artifacts 0.75 0 2.447
Suface Collection 2
No. of artifacts 62 62 35.496
frequency of artifacts 0.75 0 2.245
Post Holes 1
No. of artifacts 83 83 47.329
frequency of artifacts 1 .35 0 3.066
Post Holes 2
No. of artifacts 40.7 41 23.191
frequency of artifacts 2.75 0 6.98
Five Foot Units 1
No. of artifacts 1358 1367 841 .25
frequency of artifacts 0.563 0 1 .504
Five Foot Units 2
No. of artifacts 1244.26 1239 706.989
frequency of artifacts 1 .632 0 6.157
Five Foot Units 3
No. of artifacts 249 249 140.683
frecmency of artifacts 1 .579 0 3.58
Five Foot Units 4
No. of artifacts 359 359 202.583
frequency of artifacts 0.474 0 0.772

 

 

 

 

 

61

In choosing which statistical method to use in viewing the spread of a
distribution and its attributes, a statistic that is more resistant to outliers can be
helpful in recognizing the actual patterning of the values within the distribution. A
good statistic for this is the median, which is the value above which half of the
values in a data set occur and below which half occur. The mean and standard
deviation, on the other hand are “in particular nonresistant summaries of
location and spread” (Hartwig and Dearing 1979: 19). Examining the chart of
basic statistics, above, in the case of the first and third set of data for the five foot
by five foot units, the mean for artifact frequency goes from .563 to 1.579 when a
very large value was removed to make the distribution more representative of the
whole. The median, in contrast, remained zero.

One median-based statistical display is the box and whisker diagram. The
box and Whisker diagram uses the median as its point of reference and splits the
distribution as a whole into two parts using a vertical line to indicate the median.
The sides of the “box” in the diagram are made up of the upper and lower
“hinges.” The lower hinge “is the point above which three-fourths and below
which one-fourth of the values lie and the upper hinge is that point above which
lie one fourth of the values and below which lie the other three quarters” (Hartwig
and Dean'ng 1979: 21). The “whiskers“ on the plot represent plus and minus one
intervals of the same size as the numeric distance between the two hinges. Any
values beyond the whiskers represent outliers, and if far away from the median,
they represent distant outliers.

In examining the first box and whisker diagram of the artifact frequency for
the surface collection data (graph 9, p. 161), the distribution of values is found to
be largely contained to zero with very little spread and one outlier at 10, meaning
that the frequencies of the artifacts Clustered around zero with one outlier. In

examining the box and whisker diagram for the numbers of artifacts, which is the

62

listing of the intervals of numbers of artifacts, a distribution from zero to the
highest number of artifacts with equal whiskers, and equal sized box sides,
means that the number of artifact intervals is a normal distribution, with equally
sized intervals.

The second box and whisker diagram for the surface collection data
examined the artifact frequency (graph 10, p. 161). Again the box is centered
narrowly around 0-1 with an outlier around 2 or 3 and two distant outliers around
10, meaning that the majority of artifact classes have very few members except
for three which had very large values. Again, examining the second box and
whisker diagram for the numbers of artifacts for the surface collection data, the
distribution has a box centered at 83 with equal sized boxes and whiskers,
meaning that the artifact number intervals are equal in size (graphs 11-12,
p.161)

All of the box and whisker diagrams for the postholes (graphs 13-16,

p. 162) and the five foot by five foot units (graphs 17-24, pp. 163-164) display the
same pattern as the diagrams for the surface collection data. All of the data
distributions are heavily skewed toward zero and the numeric region immediately
around zero.

With data frequency distributions which are heavily skewed in one
direction, and are non-normal, not producing a bell-shaped curve or half bell-
shaped curve when plotted, there is little statistically that can be done to compare
them. In such an event, descriptive statistics and an exploratory data analysis
approach can be a useful way to compare the data distributions for trends,
similarities, and differences.

In order to be able to properly compare the frequency distributions, since
each sample was of a different size and examined a different amount of soil, it

was necessary to derive a common unit of measure and apply it to each of the

63

distributions. A first set of transformed data frequency distributions employed a
volume-based factor of the average number of artifacts per cubic inch of soil.
These were generated by dividing the number of artifacts in each sample unit by
the number of cubic inches in that sample unit, and then taking each of the
artifact frequency measurements for each unit, summing and dividing the total by
the number of units containing artifacts. Then each original data frequency value
was multiplied by the average number of artifacts per cubic inch for that type of
sample. The surface collection data, however, could not be viewed in this
manner as they occurred on the surface, so this set of transformed distributions
was abandoned.

A second set of transformed data distribution tables was constructed in
the same manner as the first except that the number of artifacts per square inch
of the test units was used. Then each of the values within that distribution was
multiplied by the artifact per square inch value to obtain the values in table q. A
series of basic statistics including the maximum value, the minimum value, the
mean of that test’s distribution, and the standard deviation were calculated for

each test’s distribution table. The results are presented in the table below:

 

 

 

 

 

 

Table 4:
Basic Statistics for the Transformed Artifact Densities
Frequ 1 Frequ 2 Frequ 3
No. of cases 20 20 19
Minimun Value 0 0 0
Maximm Value 0.001 0.38 1
Mean 0 0.043 0.112
Standard Deviati 0 0.105 0.313

 

 

 

 

 

The values are very small for all of the distributions. For the first distribution, the

maximum value was .001, which is very low, a minimum value of 0, a mean of 0,

 

64

and a standard deviation of 0. For the second distribution, the maximum value
was .380, which is also very low, a minimum value of 0, a mean of .043, and a
standard deviation of .105. For the third distribution, the maximum value was 1,
which is again very low, a minimum value of 0, a mean of .1112, and a standard
deviation of .313. All of these values are extremely low, and most of them are
also non- integer decimal numbers. Most statistical tests are designed for integer
number values and are not reliable for non-integer and zero values. This would
indicate the utility of an exploratory data analysis framework within which to
examine the nature of the data sets.

Consulting the box and whisker diagram for the distribution for the surface
collection (graph 25, p. 165), the diagram shows a lower hinge of zero, an upper
hinge a fraction above zero, and a whisker just barely above the upper hinge.
This is a very tight spread of values around zero. There is one distant outlier at
.001. The largest value is very far away from the majority of the values, is a
distant outlier, and is well below one.

The box and whisker diagram for the posthole test distribution (graph 26,
p. 165) is similar to that of the first distribution. There is a lower hinge at 0 and an
upper hinge at about .005. Again the main values of the distribution are very
closely grouped together. There are two distant outliers, one at .3, and the other
at .380, the latter of which is the maximum value for the distribution. Again, even
the distant outliers are well below zero and the majority of the distribution is very
closely clustered around zero.

The box and whisker diagram for the five foot test square distribution
(graph 27, p. 165) is similar to both of the other distributions. The lower hinge is
at zero and the upper hinge is a fraction above zero; there is an outlier just above
the upper hinge and a distant outlier at one. This distribution has one as its

greatest outlier, with all other values tightly clustered around zero.

65

The resultant artifact frequencies had values which were so small that
statistical tests of comparable test effectiveness, such as the difference of means
test, could not be applied. It is therefore inconclusive as to which testing strategy

was most effective since their efficacy could not comparatively measured.

3. 8. Additional Non-Statistically Based Interpretations

The surface collection almost exclusively resulted in the collection of
modern trash and building materials, such as asphalt shingle pieces. It was in
general not very useful for the prediction of high concentrations of subsurface
artifacts. In fact, for the three surface collection units with the highest artifact
concentrations, the postholes in these areas had low artifact concentrations. This
probably reflects different activities occurring in the post-school house period and
to the accUmulation of small waste items that have blown onto the site, such as
Cigarette wrappers.

In only two cases, N1115 E1100 and N1165 E1005, did posthole tests
with high artifact concentrations coincide with five foot by five foot units that had
high artifact concentrations. In short, the posthole tests did not appear to be a
good predictor of high artifact concentration areas, which would have predicted
features and out-buildings in any meaningful way. Overall, the five foot by five
foot units gave the best view of subsurface features and concentrations of
artifacts, but that would be predictable with any larger unit, since the larger the
unit, the more data and artifacts one would naturally predict finding. However, the
way the research design was setup, with each test method applied
independently, it is unknown what would have been recovered if the five foot by
five foot units had been located near the postholes with the highest artifact
concentrations, rather than independently. It is similarly unknown what the

difference in the results would have been if each of the locations of the sampling

66

tests had been located based on the highest concentrations of the units from the
previous tests; for example, locating the postholes in rows near the surface

collection areas with the highest artifact concentrations, then locating the five by
five foot units by the postholes with the highest artifact concentrations. However,
since it appeared that the posthole and surface collection artifact concentrations
didn’t seem to overlap very significantly, it is doubtful that a strategy where each
phase is dependent on the previous phase would not have been that successful

in locating significant features.

CONCLUSIONS

The thesis has two foci: the first is to archeologically analyze the Merle
Beach School Site as an example of a rural Nineteenth Century schoolhouse,
and the second is to evaluate the types of sampling methods that would be best
used to excavate such a site. As there is very little literature written on
archaeological investigations of schoolhouses in Michigan, little is known about
the exact nature of the data to be found when excavating a schoolhouse, and
there is still a great deal left to discover regarding the best techniques to employ

in their investigation.

1. Conclusions of Compared Sampling Strategies

From examining the data provided by the different sampling methods,
several things become apparent. Firstly, while surface collection does allow for
the widest coverage of a given site, it frequently results in the recovery of modern
trash blown or later deposited on the site after the site’s previous usage. On sites
such as the Merle Beach School site, with a dense cover of vegetation, small
objects can be obscured, so coverage is often not truly complete and not entirely
representative of the materials which may be buried at the site. Therefore,
surface collection should be used with care and the results should be closely
examined. Surface collection is frequently not a good indicator of the nature of
the artifact deposition and patterning within a site.

Posthole tests have the advantages of being smaller and better at

preserving vertical stratigraphic layering, and are also fast and easy to perform,

67

68

just like shovel testing, but they are also not necessarily good predictors of areas
with features and high artifact concentrations. In the case Of the Merle Beach
School excavation the posthole tests with the highest artifact concentrations did
not point to the area where a five foot by five foot pit located the trench of a small
outbuilding, probably a coal shed (Terry Shaffer, personal communication 1991).
However, since five by five foot units were not located near all of the areas with
postholes with high artifact concentrations, it is impossible to speculate whether
the postholes with high concentrations would have revealed additional features,
and perhaps later examination of the site in these still unexamined areas would
answer this definitively.

The five foot by five foot unit excavations, not surprisingly, seemed to
provide the greatest amount of data. One located behind the schoolhouse
seemed to be a refuse pit of building materials from one of the renovations of the
schoolhouse. Another five foot by five foot unit, located to the north of the
schoolhouse, revealed a trench from a small outbuilding, probably the school’s
coal house. It is hardly surprising that larger excavations would reveal larger,
more complex site patterning. So would exposing, layer by layer, the entire
surface area of a site. Doing so is usually impossible, however, as it is simply too
costly or would require heavy earth moving equipment which is often expensive
to maintain and often difficult to obtain and operate.

From the excavations at the Merle Beach School, a series of more
effective and productive methods can be constructed. Surface collection provided
little in the way of artifacts representative of the time period that the site was in
use. Surface collection can therefore be excluded. Rather than using auger
holes, postholes, or shovel tests which do not provide adequate volume or
statigraphic control as excavation units , a series of smalltest pits would be more

appropriate. Small units such as two foot by two foot test pits, or fifty centimeter

69

by fifty centimeter test pits in a much higher frequency, say a hundred, drawn
from a stratified random sample would be much more useful and time efficient.
Then, instead of the larger five foot by five foot pits, larger surface area
exposures could be excavated in areas where large artifact groupings or features
were indicated by the test pits. This would allow for a greater understanding of
the site layout, allowing for whole features, such as the entire building from which
the trench segment was found, to be excavated. This would give an all around

better picture of the site and site activities.

2. Results of the Patterning and Interpretations About Schoolhouse Use

From the excavations at the Merle Beach School site, it appears that there
was at least one phase of significant remodeling, evidenced by the large amount
of building materials found in the five foot by five foot unit located behind the
schoolhouse. Probably several smaller episodes of remodeling occurred as well
during the school’s use life, such as repainting, tightening loose or damaged
shingles and boards, painting, and other general maintenance. There is also
evidence of a small outbuilding located north of the schoolhouse. It had a three
foot deep foundation trench which was largely filled with coal and old metal
remains. It appears to have been the coal house.

There was very little in the way of domestic refuse such as tableware and
bottle glass in the deposit. Although both were present in small amounts, there
were very few sherds of each. The sherds were very small in size and from a
large number of ware types. The paucity of sherds indicates that domestic
activities were not a regular occurrence at the Merle Beach School. It is most
probable that the teachers lived someplace other than the schoolhouse. It also
appears that large scale cooking “domestic science” classes, common at the turn

of the century, were not being taught at the Merle Beach School, or the amount

70

of domestic wares would be greater (Foght 1910: 236-253). This interpretation is
not entirely conclusive, however, because neither the boy’s nor the girl’s privies
was discovered in the excavation. Those pits may have been where such
domestic refuse would have been deposited.

Faunal remains were very scant, with only small fragments of animal
bones even from the large five foot by five foot units. Probably any food
consumed onsite, such as the lunches of the teacher and school children, was
likely to have been of more processed foods (perhaps slices of meat rather than
larger animal parts, for example, chicken legs or rib sections). It is also possible
that any domestic food remains, such as lunch box waste, was removed from the
site daily by each student, or that it was dumped into the privies or into other
trash pits not yet located.

A significant amount of the artifacts that were recovered during the
schoolhouse excavation consisted of lumps of burned and unburned coal. Coal
would have been stored and used in abundance to heat the school after the
conversion from wood fuel around 1920 (Terry Shaffer, personal
communication), but it was found only in the area of the coal filled trench,
confirming the building’s uses as a storehouse for coal. Also, the fact that no
other coal was found indicates that it wasn’t used for other site level activity.

A general lack of artifacts in the wide open space outside of the immediate
surroundings of the schoolhouse would indicate that this was the playground
area. A play yard would be expected to be virtually artifact free. Since it was a
zone of high foot traffic with many children running around, any artifacts would
likely be pushed to the very fringes of the area. There are also a number of
children’s games, like tag, jumping rope, or playing ball games, which use few or
no pieces of equipment and hence would result in no loss of these objects and

no deposition into the archaeological record. In the cases of jumping rope and

71

ball, the equipment used would be used over and over until it could no longer be
used and would then end as intentional discard, not something that would be
found on a playground.

The only personal objects that were found that related to the children were
a few buttons made of mollusk shell and plastic and one 1946 penny. These lost
objects were probably found around the playground as vigorous games and play
would be likely to put stress on Clothing buttons, causing them to fall off and
swinging on the stride swing would be likely to cause loose items, like change,
fall out of the pockets, becoming lost. The general lack of small portable toys,
such as marbles and jacks, could be related to the location and social class of
the school district. Small portable toys like jacks and marble, because they are
small and highly portable, would be expected to appear in the archaeological
record, as their small size makes it easy for them to fall out of a pocket or
disappear in tall grass. Merle Beach School was located in a relatively poor rural
area with risky fortunes, which rose and fell several times during its uselife. Poor
rural Children are not likely to own much in the way of toys and are likely to curate
closely those they do possess. Such students, therefore, are either not likely to
bring toys to school, or they are likely to keep a Close eye on them, losing them
very infrequently. Additionally, poor rural children are likely to use natural Objects
or found objects such as sticks, pebbles, leaves, and nutshells as playthings in
the absence of formal toys.

Finally there were a few school-specific objects found during the
excavation, including a number of pieces of slate pencil, smooth slate, chalk, and
the broken-off erasers from pencils. These indicate the use of the site as a
school and cover potentially the entire time period of the school’s use, with the
slate pencils being the oldest and the pencil ends the newest. These were found

Closest to the schoolhouse and were possibly swept out the door during cleaning.

72

Alternatively, they may represent concentrations of certain outdoor activities
close to the school buildings.

Overall, the artifact patterning did show marked patterns of activity,
identifying a coal shed, a lack of artifacts indicating a playground, a number of
artifacts indicating the site’s use as a schoolhouse, and a lack of domestic
activity. The lack of toys indicated that the school site was a rather poor rural

school.

3. Summary of Conclusions and Recommendations

The data obtained from the sampling program at the Merle Beach School
site illustrates some important weaknesses as well as strengths of some Of the
sampling methods employed, and points to new directions for research
methodology. The surface collection and posthole samples yielded too small a
number of artifacts to be able to allow for statistically relevant analyses of the
data. The posthole samples yielded samples too small volumetrically and allowed
too low a level of visibility of the stratigraphy. The five foot by five foot
excavations allowed for an adequate volume of soil to be excavated, yielded a
significant number of artifacts, and afforded good visibility of the soil’s
stratigraphy. However, the five foot by five foot units were time and labor
intensive, and there weren’t enough of them in this sample to give a
comprehensive view of site structure and usage patterns.

The inadequacies of the sampling methods employed at the Merle Beach
School Site suggests a new direction for sampling small sites. Using a test pitting
strategy employing a large number of small test pits, such as two foot by two foot
or fifty centimeter by fifty centimeter test pits, would yield a larger and more
representative data set. A hundred test units could be chosen randomly from a

set of zones dividing the site. This large number of test pits could be excavated

73

relatively rapidly, would allow for goOd stratigraphic control, and would allow for a
much wider area to be Observed. In this project, nearly 60 two foot by two foot
test pits could have been excavated in the time used to excavate the nine five
foot by five foot units that were completed. One hundred two foot by two foot
units could be excavated in the time used for all three of the types of excavations
completed for this study.

Following the small test pits, larger shallow area exposure type
excavations could be undertaken, especially focussing on areas where features,
such as the partially exposed building trench, are uncovered. These large
exposures, as they would be shallow scrapings, could be done fairly easily as
well, and would yield a significantly greater area to be exposed. This would allow
for the potential discovery of whole building remains and the exposure of whole
living floors. Such an approach would enable the researcher to look at the
patterning across the site as a whole in a really broad sense, unmatched by test
pitting or trenching. In addition to a change in sample size and the frequency of
units used to sample an area of a rural school site, some areas would provide
better data than others.

As a result of the archival, historical, and archaeological examination, a
great deal can be deduced about the way rural schoolhouses were laid out. Rural
schoolhouses were laid out with extreme regularity. Plans describing them
appear in nearly every publication on the rural one-room school (Foght 1910:
121-124, Culter and Stone 1913: 19-34, Smith 1880). From these accounts and
the archaeological data, a template of a typical rural schoolhouse can be
established. The rural schoolhouse site consists of a schoolhouse, two
outhouses (one toward each edge of the property), a woodshed situated not far

from the schoolhouse, and possibly a small stable for the teacher’s horse.

74

Based upon this template, several new approaches can be taken in future
archaeological study of a rural school site. Since rural school sites are very
regular, future sampling strategies can make use of this knowledge. Excavations
can be intensified in certain specific areas, such as along the fence rows, where
the privies would most likely have been dug and re-dug over the years, as well as
in the open spaces directly adjacent to the schoolhouse, to locate the stable or
fuel house common to all of these sites. This allows archeologists to stratify their
sampling areas and focus on the areas of schoolhouse sites that would be most
likely to reveal the structure and lay out of school buildings. .

The results of the Merle Beach School study do reveal a great deal about
the structure of the rural schoolhouse. This can be used to contrast the rural
school with the urban school. For example, the North Larch Street school in
Lansing, which was excavated by Mark Branstner, revealed the remains of a
multi-room structure building foundation, thought to be two stories high, Clearly a
much different type of site than a rural school building (Mark Branstner, personal
communication 1997).

An urban school, in addition to being considerably larger, would be likely
to produce very different artifacts than a rural schoolhouse. There would likely be
a number of broken toy pieces discarded deliberately, as a school with several
grades as well as a gymnasium would likely have toys and sports equipment that
a small rural school would not have the facilities to house. Urban children might
come from a wider spectrum of socio-economic backgrounds and different
neighborhoods, so there might very well be a wider spectrum of items that were
lost in play, such as fancy marbles, expensive buttons, china doll heads, perhaps
a piece or two from a ceramic tea set, or lost jacks. There might also be more
noticeable differentiation in what children ate, possibly indicating either socio-

economic or ethnic differences, from the refuse deposited in most likely privy pits.

75

More formal and systematic disposal of debris and waste is more likely to occur
in urban areas which are frequently covered by sanitation codes than in rural
areas which were less formal about sanitation during the Nineteenth Century
(Foght 1910: 282-83).

The new sampling strategies suggested by the shortcomings of the data
collected at Merle Beach School lead in a bold new direction for small site
excavation. Such new sampling strategies can contribute greatly to the field,

leading archaeologists to new levels of understanding of cultural processes.

APPENDIX A
TABLES

76

 

$09324 .92.

 

ooo>>

 

ocom

 

Em: .oa P

commficmc w

tam F

osmmE

 

 

.6922 on V

850

 

22280

 

EDEN—n—

 

 

.om r

8.822

 

8.82 9.52

 

282 Seem

 

85E: rm

mam—w

 

593.3 _.

288,90 EEO

 

mmm_0 6.“.

 

 

mmma 250m

 

80053538 9.

 

 

.922

 

 

mr

 

mr

 

E.

 

we

 

NV

 

or

 

 

 

gamma

”cozmcmfimo ED

 

 

8m 98 888:8 888 um 28E

77

 

0:.

mm

98:5 :90»

 

82$

 

ocom

 

:8 w

22 e

tow w

ozmmE

 

 

EEO

 

oumuocoo

 

.88 289

 

92me

 

mm

or

993w

 

£62 952

 

‘—

m__mz 506$

 

92m

 

:58 Na F

o_EEoo 550

 

E.

3

.330 um.“—

 

 

‘—

mwma mEom

 

60053588 #1

 

 

.8 r

Em.)—

 

 

 

290k

on

 

KN

 

mm

 

mm

 

vm

 

FN

 

mr

 

g

“cozmcgmmo .ED

 

 

8m 23 882.8 88:3 ”m 289

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N

w

 

(EN—3
mos

 

$959.
meow

 

 

(3.99.

moo r

 

(\mno—

m.~mow

 

(BN3

mNmE.

r

 

 

(Boo? .

 

 

m.~mop

m\m.Nv 3
ovo—

 

_.
(EN? 2.
over

 

mBNnow
ovow

 

$050..
over

 

 

(Boar
mNNo—

 

 

u_m>o.<oo:o_:o>oi

 

:28230 0.05.9.9". 3. 250 Bay. no 03w...

79

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N V o h N #0 N w P ”N N
.mtoE n 235E N
w h v om
v
n
_.
r
F
n F
e
m w=wz «:0 Eu... w
WINE". .06.? 0:06
_. F
I N
03.503 $952 ($803 mEN: 9(3er (EN: . mac: <50: mane some mace .
omow omor oao— mfihor mKNo— mfihor mfihov WES .8152: 0:2 mshow U.mv>e.<eocm_co>o.__n_

 

 

 

 

 

 

 

5:02.00 Beg-mom .2 San 3am ”m 055...

 

 

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8:02.00 2053.". 3.. Eco 3am No 03w...

 

 

 

v o F n o Fm FF n F N Nm a F
F N F n
F
av F 0N
F mF
F
F 3
F F F
F F a
N F N N
m N
v F m .mooEancamon.
F
F F
F F
F N £82 50 En: w
may“. _ooF ocoum
F
F F
mENoF <\mNc F (\mNnOF mBNF F (\mNF F $2.2 . (\omoF <\oooF m\m.NF FF (\mNF FF OBKmo F .
mKNFF mHNFF mFFF mNoFF mNoFF mNoFF mNoFF mNoFF oaoF omoF omoF H_0>ol_\moca_F6239".

 

8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N F F 3 «N F N F. F 2 F
F
F
F
N
F
F m
F
F .uF on n F
F F N F N
F
F
F
F
I F
3.88 <B.~8F <39: mSNFoF $32: <85. 98: $8: 938 <59 <82 .
3.: 8: 3.: 8: 3: 3h: 3m: QR: 3w: 3m: 2w: ”8898:2555

 

 

 

:o_Foo__oO 20:80.". .8 Sun. 261 no oEak

82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N NF m m n F N FN FF FF v FN
86885.8. F
N
F m
n v
uoEzn F n v
v
N F. F
o N m
o F n
F F F N N F N VF
F
F
F £32 «:0 2a: m
may". .00... 986
III I
mBNF FF (EN F FF (\mNmoF <\m.NmoF mBoF F <\mNF F . (\ooF F 92.3 (\omoF mFmNoF (FocoF .
moFF moFF mcFF moFF mNmFF mNmFF thmFF mNmFF mNmFF mNmFF mNmFF H_mv>w._\oo=mz.._.n..>o.n_

 

:28230 205231 .8 Eco 3am no Soak

83

5:80:00 20:63. .2 END 26m no OBOE

 

.30... 9.90

 

ocom

 

came—n.

 

.650

 

3:ch

 

22250

 

362d

 

va

seem

 

wQEU Efin.

 

Econ.

 

25w UOEzm

 

22w

 

$08ch

 

=ocw xm:=o—2_

 

noo>>

 

:25 55mg

 

.>>OFE>> OONSO

 

$30 .2”.

 

 

8&0 Exam

 

.mooUoEammomi

 

.eeoeeecaeeelm

 

25> ooEwm

 

.322

 

@5me =95

 

328w

 

£32 0.53

 

262 So 96:

 

 

lame“. .68. ESE

 

www.cra—

 

OMOQMa-FMCOP

F

mo“.

 

 

 

£26...

 

OFmNF FF
mmFF

 

3%

”_0>01_\8:0_C0>O._l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

on
m Em: F mcozsn N
N N
F F

F

F
F F
N e2 We 3 N
u N

F u n
F F

F .F .F

3 V V

$82: m 29 2 89 m 3.on 22°... mZBmoF m mFoF z mBmoF m 2.3 2 33°F M23 2 I

 

325 .c nxm .8 Sun. Bum K 03w...

 

 

 

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wcotzn N

 

 

 

 

 

 

 

 

F

 

EtmE_..m.§>F HN

 

 

 

~‘v-w—r‘—

 

 

:‘V’Y‘

 

 

 

 

 

 

 

F

 

F

 

F.

 

 

F
m
owoF m 2.2 2 £80...

[I

 

NFF m mBooF m 05°F 2

_N\F Z mboF m onoF Z

 

 

m\<\omoF m 05°F 2

< owoFm ehoF Z

 

3E2 .c mxm .8 Eco 26m K 03a...

 

 

86

 

«ED .0

308:2 =m Fo £28.

 

 

 

 

£3 E __ .o F ” .8 m HMN; {BEEN MS: :26ch mF

 

55:0

 

22280

 

223.8

 

xecm

 

ea 58¢

 

Eros.

 

mam—w

 

Econ. 22w

 

F

.8820

 

:95 333:2).

 

.mom sznF Nun

 

 

.mo .vF.wN\F Fwd—0E5 ono

 

uoEsn .30 0.023 NF ”.8 FndNFF voF.o_o:>>omF”Nnm

 

too>>
:95 55mg
2EEeo $50

 

FED .mchmLFwocmstmon.

 

n

.>>2_c>> UONEO

 

820 F2".

 

 

 

3&0 050m

 

.mooUoEammom

 

3005358.“.

 

25> ooemm

 

.222 .095

 

.8 as: mm

.822

 

£30 :25

 

mzwz o:

 

£32 50. m

 

m Eu. _oo._. 2on

 

8wa

 

N

 

 

N\F>> (\mocF m ooFF Z

 

N\Fm (\mooF m owFF Z

 

.tszooF m ooFF Z

 

mac: .c mxm .8 850 3am K Sank

 

mo".—

87

 

 

Foo
F

 

 

3

 

 

 

 

 

 

 

 

BEES N

 

 

um

 

 

 

 

:05 NP;

.moa N

 

2: 5% F

 

53:3 3 ~—

 

 

 

Nr

--

 

‘7“‘1

 

 

 

 

 

 

 

 

 

‘—

 

 

 

 

N

 

o

 

 

 

m2 no: u one 2

tam 3: m onow Z

 

moor m oo: 2 29°...

 

 

mmoow mom: 2

mch .c mxm .8 Sun. in! K 02.3%

 

m
m\<\moo_. m co: 2

 

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘—

 

 

mizmcem uoNEm xma F

2: 555 F

 

 

 

mw

 

 

 

 

 

 

 

 

 

 

('ON

55 95.98. N

    

£62 So- w

 

 

mami _.oo._. 22w

 

 

 

 

 

 

 

K .
no: N one 2 £20... .35 .335 no: m one. 2

 

m
N:Z m mo: w 002 Z

 

2...: .t mxm he San. 3am K 05a...

89

 

 

 

 

 

0. 55 w

 

mmm.

O:

3.22; f
:5... __ .o F” 8.2058203 ‘

ED 5 3082.2 =w “.0 23°

 

.mEEm r

ocom

 

ennui

 

 

h$50

 

55:0

 

wacky—.00

 

.EmmE

 

xocm

 

mza zomwm

 

Etc—2

 

822:: n

22w

 

‘—

=o=0m 92m

 

.82ch

 

‘—

=wcw xw:=o_>_

 

w

voo>>

 

__05w «:55

 

EBom E; a

28200 550

 

“En. 8.285033:me

 

.>>3E>> umNEG

 

on
:35 9. 6K

 

mmmn Scum”:

non Nioocw 5N

320 .2".
3&0 050m

 

 

.mooBEsmmun.

 

 

25> 895m

 

WES

 

 

2825 N ucoEmEo F

8 :8 0 E320 F 6:; o F

 

.oa _.

is: .85

 

mo :8 o 2568 ficozan L: 503.. F

.322

 

 

 

gag—€53. :3“ :

 

.8 nutofi K

:0. F

tozm om.vE~. 5me

£30 :95
m=mz 9:
m=wZ 50. m

 

mmmi .8 h 285

 

$me

 

v

 

 

 

N>>>owwpmo$rz

 

(our—m 2:2

 

tam on: wow: 2

 

N:.>><ON: movCZ

 

to";

 

mac: .c mxm .8 Sun— a”. K 03w...

90

 

non

 

.mEEa o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

KN
. _‘
w r
w
.
8:3 P cos—5 F u
m u
mENaEn P w
a... «£5an 5?". 2m. 0 F
m Km
2 _‘ SF
565 N .o xom: v _..E cam: .26.... 225:3 m an
m N
om
N
v
K
5
mm
m

 

‘—

 

 

 

m\< 02:. m mv: Z

 

 

(owe m mv: Z tzw owow m mv: Z

 

V
cm: w or: 2 £38.

 

325 .c mxm 58 Sun. 3a! K 03a...

 

 

91

 

ES .6. 3005.2. =m be $98.

 

 

_EEm .Em 852.? N

 

 

.o a: F .265. ”N

 

cozsn to: 6:55 t

 

=2 :53 N

 

mucm .880 .88

 

26:0»

 

x320

 

.mn. mN.oN @ 3 new

 

 

 

 

 

.3. mN. @ N

:25

 

9E :88

 

5:22

 

22w

 

=ocom 32m

 

 

 

=93 =95 .Em r

95:3 v

.32ch
=ch xm==os_

 

 

22.; 18¢ a ”2

 

 

.oq .ozozfiou :: a

 

.mom 3.20:3 P M:

voo>>
=25 Sc_m>>
2Efioo .050

 

 

E: uwmmohmE: ”3

.En. .8mcm.....w2m>>tmon_
.>>2E>> umNEO

 

VN

 

NF

:85 «:86 86:3 o :20 320 N

890 3.".
3&0 050m

 

EoOuaEammoa ..

 

.8q mm

mm?

ESE—5:38”; ..

 

25> toEmm

 

 

 

355. 895

 

 

m=ac 5on or 682 .maoENécaEd cENum

.205.

 

 

 

 

 

 

2:80 :95
m=wZ 25>
332 Solw

 

98". .08. oco—m

 

 

m

mome

 

.3. Km; omi m

mun.

 

 

NFMB owow m mv: Z

 

 

05 one m mv: Z

 

mower m mv: Z

 

 

3E: .c nxm

.8 Sun. 3am K 03a...

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v SKN . 068E< =0 .o 0.0.3
w .256 .E0 .8 K
.m 0.. 0 0.26:0» N a
0.2. Hogan :0: v .0662. 60.2 P 2
8
vN K..mn_ oé .0 K”:
m
0
EL};
K
2
06:2. 2 0m 06:; N60 0 UK
ON 6 26:60! .5 v
n
:
c2
K:
a
.23. .8. .0
_.
0:2, .0 :0508 F KN
0F
5
n

 

8.30 mNoK m mN: Z

 

 

 

.tzm 08—. N no: 2

 

 

one m 9.: 2 £80.. N>>>\o one u mv: Z

 

.mo 3.

8:5 .c mxm .8 800 30m K 030..

 

 

 

 

93

 

 

.0EE0 .Em P

.mEEm .50 r

 
 

«ED .0 808...; =0 .0 £20

 

 

 

.2... 020.0 .0 F .05: 9.0:. N

 

 

KN

 

.wo =0Em v

MN

 

 

 

 

 

 

 

 

 

 

 

 

 

0.8.0... t

 

 

u
.006 w
.006 V

 

 

an

 

 

=0Em v

 

 

 

.2. .0 8 2200.68 655.8 P

 

.m 0.. man... .20... F

 

 

mam c8. .2. F

 

 

m—

.3

 

 

 

 

 

mNoK m mNK ..Z 0.80...

 

 

m mNoF w mN: Z

 

.tzm 08.. m mN.. K2

 

N22 8.3 mNoK m mN: Z

8...... .c mxm .8 Sun. 26m K 02.5%

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

c852 N\K”K tom x.nK .08.:2. KHN
60. m U0. n
.8 N
K
K o
m
m .006 K.
K .006 m
N mK
K
K
K N
9.0: .0355 8.5.03 K
K N K
m . N
_..02.o\m\< OKoKmmmcKZ OE OKoKMmmoKZ mOKoKMmmoKZ mZOKcmemoKZ (OKoKmmmoKZ

 

 

9...: .c 96 .8 800 2.0”. K 0.90...

 

 

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.3 2. as: .8 98...... =0 .o 208K
3. 0Com
0. n 2.3.
an .26
.0 .65
no 0.0.9.00
m .050?—
.. 0 .25
m m a... 58.
0 5.22.
.0 . 220
0 65.. 220
8 . 38.0.6
.. :20 .0232
R m 082.
.8 :20 5502.
mm 0.50.00 .050
m-. «Em .0chw..r.w0.0>>t00n_
B. .3222, 0306
~00 a $06 .2.
.0. o. 9.0.0 2.80
.0 3885:0803
008 .0 585358..
cm 2.2. Beam
.. . .222 .85
E 35.2
. 0528 :20
0.. . £02 25>
00 m 2.02 .800
K. 30.“. .00... 0:2
0 .. meat.
0.. .. mo.
3...: =0 28...... .20. 2.... m A03.2 was. i

 

0:23 .c mxm .0.“ mama 3mm K 030...

 

 

m.

<ON: we: 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmnd Xmomowmmvrvz
mN o:0.:.~\.>>000o_‘ mva.
V 0v «\.>>0000Fm_mv:z
mumd mmwd v mm.vm o 000. w mi. 2
mNd mmmé mndw mm.‘ m 000. w mv: Z
mNO mNd mnmfiv m mmow m mm: 2
.mn._ E .§> mm... c_ is mn._ E <5 .mn.._ E .S. .84 E .3
.65 mo". 229.60 :80 8:50 :80 853:: mocm_cm>o.n.

 

 

 

9E3 .E m .5.— mEQ0>> 10”. new 0.0.9.00 .300 N0 033.

 

 

97

50.0
0' F.0-
0h0.0-
000;

0:050 KMIPO 0.10200 10.05.... m-_02=._0. 0-=<Z>> 042200

N000-
000.0-
«00.0-
000.0-
000. F

Ema 5.8260 8350 m... .o. x...m_2 50.00.50 comawa .0 223.

#000-
000.0-
30.0-
050.0-
N000-
000. v

30.0
000.0-
~000-
000.0
~000-
500.0-
000. _.

N30-
050.0
N30-
.000
30.0-
N30-
>000
000...

000.?
000.0-
#000-
53.0
N000-
30.0-
03.0
N30-
000...

000..
000.0-
«00.0-
:«0
N000-
V000-
N30
N30-
000,—.
000.?

#000-
000.0-
.000-
050.0-
N000-
000. v

500.0-
Nv0.0-
#000-
#000-
000. P

whim mmomko 040.5“.

000.0
:00-
000.0-
000.0
000.0-
000.0-
30.0
.000-
000.0
000.0
000.0-
000%

«00.0
No «.0
0000-
N020
000.0-
000.0-
000.0
000.0-
00.0
.000
000.0-
0020
0002.

000...
000.0-
30.0-
» 2.10
N000-
#000-
03.0
Nvod-
000.9
000.9
#000-
000.0
_.00.0
000.?

-_<hms_

.2000:
000>>
wZOm
02.050
«ETC-0
0.10200
mbwja
m-_02_10
00_<Z>>
0..__<200
whim
mmomPO
040.5...
-_<._.m2

98

000.? 000.0- 30.0-
000... 0020
000%

4<000: 000.5 MZOm

Ema 5.82.00 momtzm m... .o. 5.65. 5.5950 538... .0 2an

4<000:
000>>
m200
0_._.0<4n_
KMIPO
._.~.._0200
«2.ij
w402=._0
04_<2>>
043.200
mh<40
«mom-F0
040.5".
4<._.m_2

99

5N0.0-
000.0
000.0
5000-
0N..0
05 ...0
N5N0
500.0-
500.0-
:00-
0N00-
000.0-
500.0-
50 F0
MVNO
000.0-
000.0
000. P

040-500 4<000 4<002: 1.>>1<m 44.5w... 0(0410 >>m100

000.0-
0N00-
0N00-
000 .0
VN00
000 .0
0V00-
0N00-
0N00-
0 .. 5.0
0000
00¢.0
0v00
N000-
0V00-
5V00-
000.0-
NN00
000. F

N000-
0.0.0
0500
505.0
500.0
N5N0
:00
50.0-
5V00-
000.0-
000.0-
0020
0.0.0
00N0
500.0
050.0
005.0
000.0
0FNO
000...

..000-
5N0.0-
5N0.0-
5N0.0-
050.0

000.0-
530-
5N0.0-
5N0.0-
900.0-
#000-
:00-
5N0.0-
0N00

:00-
0v00-
#000-
..000-
..000-
N 30-
000.?

..000-
5N0.0-
5N0.0-
5N0.0-
0NF.0
000.0
..5V0
5N0.0-
5N0.0-
500.0-
000.0-
:00-
5N0.0-
v00 .0
300-
V000
0520
00N0
..000-
5N ...0
.000
000...

3. ...0

0.0.0-
0500-
050.0-
5N0.0-
NN00-
#550

0.00-
0.00-
5500-
0N00-
0N00-
03.0-
5 5.0-
0N00-
0N00-
5000-
N500

0N00-
5v00-
5N0.0-
5N0.0-
000.5

0500-
030-
0.0.0-
0.0.0-
5N0.0-
N050

0N00-
0.00-
0.0.0-
5.0.0-
0000

0N00-
050.0-
550.0-
0N00-
0N00-
5000-
N500

0 500

..m ...0

5N0.0-
5N0.0-
0.00-
000.5

50.0-
5N0.0-
5N0.0-
5N0.0-
050.0
000.0-
5.00-
5N0.0-
5N0.0-
500.0-
#000-
0V00
5N0.0-
0N00
30.0-
000
000.0
30.0-
000.0-
NN00
50.0
00.0-
5N0.0-
5N0.0-
000. ..

04.<Z>>

0000-
N000-
N000-
N000-
055.0
0N00
05 P0
N000-
N000-
0000-
005.0
0V00-
N000-
000 .0
#050
000.0
0v00
00.0
0000-
N050
000.0
0¢N0
N000-
N000-
0¢N0
000...

..00 .0
5N0.0-
5N0.0-
5N0.0-
0NPO
000.0-
NON0
5N0.0-
5N0.0-
500.0-
000.0
3.00-
5N0.0-
00.0-
v0.0-
000-
500.0
0N F0
..000-
N000-
00.0-
.000
5N0.0-
5N0.0-
00.0-
50.0
000.5

030-
#30-
300-
#30-
550.0-
3.20
0 ..00
300-
5N00
03.0-
.00-
..000
#30-
00N0
50.0-
000.0-
030
N00-
500.0-
5.00-
550.0-
00.0
30.0-
00.0-
5500-
0000
550.0-
000.5

042200 0mv.<4n. 10".

£8 m_o£mon_ m... ..o. x532 5.6.9.00 598.”. 0.95.5

1.3.-10$.
1MI50
101.50
v.4<....0
...10200
150.141
v.0.1m
110221
>22ma
M5500
m...<40
1<...0
4m10405.
000>>
4m104<>>
w1<>>>>0
040.54”.
040.500
4<000
4<002:
1.>>1<m
4<._.m.2
0(0410
>>m100
04_<Z>>
04.<200
0mx<41
10“.

100

#500
0N00-
0N0.0-
0N0.0-
0.. P0
00..0
000...

v0.10

00.0-
0.00-
00.0-
00.0-
5N0.0-
NN00-
0No0-
000.?

1....02_<1 >sz1 m...<40m 0.540

0.00-
000-
00.0-
00.0-
5N0.0-
NN00-
0N00-
00.0-
000.—

0N0.0-
500-
5.0.0-
0020

00.0-
0N00-
000.0-
5.00-
5.00-
0004

NN00-
0N00-
0N00-
0N00-
v000-
0V00

000.0-
0N0.0-
0N00-
NN00-
000. F

0N00-
0N00-
0N00-
30.0

:00-
000.0-
00.0-
0N00-
0N00-
00.0

0N00-
000. —.

1<IO

00.0- 00.0-
00.0- 50.0-
000- 50.0-
00.0- 5 0.0-
5N00- 0000
NN00- 500.0
0N00- 00N0
0.0.0- 5 .00-
000- 5 .00-
5 ..00- 00.0-
0N00- 00.0
0N00- 0N00-
000.? 5 ..00-
000%
4910402 000))

0N00-
0N00-
0N00-
0N00-
:00-
0000-
N00-
0N00-
0N00-
0N00-
00.0-
500.0-
0N00-
5N0.0-
000. w

4NI04< m1<>>>> 040.541

0N00-
0N00-
0N00-
0N00-
0000-
N0 F0

000

0N00-
0N00-
0N00-
0 0.0-
000.0-
0N00-
50N0

0'00-
000. F

San. 90581 9: .8 5.55. 5:39.00 5281 0332.

.000-
000.0
000.0
500.0-
500.0
020
V000
500.0-
500.0-
:00-
050.0
5N0.0-
500.0-
00N0
000.0-
0 0.0
000. w

1<...10.2
10.1.0
19.1.0
v.4<I0
510200
11.0(41
10.10
1IOZ.<1
>22m1
m..-<40m
m...<40
1<I0
4w...0-.Os.
DOO>>
4MI04<>>
w1<>>>>0
0401.41
0401.00
4(000
4<0023
1.>>1<m
4<...m_2
0<04I0
>>m100
04.<Z>>
042200
09.51
101

101

000.? 00.0- 0 .00- 0 .00- 00.0-
000... 000.? 000- 5N0.0-
000.w 0.0.0- 5N0.0-
000.w 5N0.0-
000;.

1<...1OS. 10......0 1011.0 X4<IO ...10200 1...0<41

£3 90581 9: .8 5.5.2 5.5.2.00 5281 5333.

0N00-
NNo0-
NN00-
NNo0-
000.0-
000. P

1<...10.2
107:0
101-r0
0.4.410
...10200
1...0<41
v.0.10
110221
>2201
0.2400
0.540
13.0
407040....
DOO>>
40104<>>
w1<>>>>0
0401.41
0401.00
4<000
4<0023
1.>>1<0
44:02
0(0410
>>m100
04_<Z>>
04.<200
002(41
101

102

V20
0 w v.0-
V500
0V00
V..5.0
0V00
0N0-
0N00
03.0-
00...0
v0.0-
00N0-
000-
000.0-
N..N0-
0N00-
00N0-
5..N0-
..0 ...0-
000.0
000..

4.1000 4(0003

050.0
005.0
5N0.0-
000.?
000.0
0006
00N0
VNN.0-
53.0-
50N0-
0NPO-
5V...0-
0N00
V000
..000-
000
0N ...0
00.0
000.0
N000
VV00
000.?

0N00- 050.0-
5000- 00V0-
V050 5 ..00
V000- 05 V0-
00 P0- 0Nw0-
V000- 05 ...0-
0000- 5 FVO-
300- 0020-
0N00- 0000-
05 ...0- 00 P0-
5 .00- ..000-
0500 V500
V0 90- 0NNO-
V000 NN00-
V000 050.0
VV...0- 0NNO-
00.0 500.0
00V.0 V000
000.0 VN00
500.0- 00...0-
50N0- V000-
0000- 05 F0-
000... 500.0
000.?
01:50 4<1.05.I0

2mm. :5 .10 .o. 5.5.). 5.5.9.00 5281 N Z mam 1.

000.0
050.0
5Nw0
0000
0000
000.0
0V00
00N0-
000.0
0000-
V.. P0
00'0-
0V00
000.0
.000-
V000
000.0
000.0
0N00
000.0
000.0
V000
000.0-
5V..0-
000.?

4<1.02 0(010 04.<Z>> 04.4200 0(111.0 00251

000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000.0
000. P

0N50
00N0
000.0
000.0
005.0
000.0
03.0..
000.0-
000.0
000.0
050.0
00N0-
500.0
5V00
00N0-
505.0
0NFO-
00V.0
0.0.0
V000
0N00
000.0
5NNO-
00 F0-
0050
000.0
000...

N00
0V0
0V0
000.0
00 P0
0000
00...0-
050.0-
000.0
00N0
000.0
0N00
5.50
005.0
0000
50.0
V520
N000
NV00
555.0
V30-
0000
5N..0
000.0
0050
000.0
.500
000..

0000-
000.0-
V000
V90-
0000-
V0...0-
VOV0-
N000-
0N00
:00
0V00
N..N.0
000.0
NV...0
02.0
0520
V000
V000
000.0
000.0
0N0-
00...0-
500.0
..0N0
000.0
000.0
05N0
V000
000...

0VVO
000.0
0.00-
0 ..00
53.0
0.00
005.0
00V0
0..N0-
00N0-
5..N0-
0020-
..0V0
00.0
50...0-
V500
000.0-
N0...0-
P050-
00V0
09.0-
NNO .0
VPN.0-
00N0-
0NVO
000.0
0VPO
50...0
00N0-
000...

005.0
..VVO
5V0.0
005.0
NON0
005.0
NN00
0090-
N000-
50N0-
V000-
0w P0
5V50
:00
03.0
NN50
0N0
.020
V0 v.0
005.0
00.0-
005.0
0V...0
0N00
V050
000.0
..N00
0000
000.0-
0000
000.F

101

0200
0.1.0431
1011.0
v.4<.-.0
1.10200
11.0(41
v.0.10
1.11001
1<1.10.2
. 0h<40
2011.50
1<IO
401040.).
0002.
0I04<>>
10011.0
01<>>1
1<>>>>0
0401.41
0401.00
4<000
4<000.4
01.>>0
4.690.210
44.1.0.2
04010
04223
04.<200
0<11h0
00v.<41
101

103

NV10-
0N00-
0N00
00N0-
0V10
00N0-
000.0-
0010-
0550
000.9

01.<40 2011.50

N500-
0V00
0NVO
0N10-
3N0-
0N...0-
000.0-
0NNO-
000.0
350
000...

0010- 0N00
V500- 000.0
000.0 000.0
0V...0- 0N00
0VNO- 000 .0
0V10- 0N00
NON0- 0010
N010- 00.0-
0010- VV10
0000- 000.0
0F 10- 0N10
000.9 VNNO-
000.?
1(10 4010405.

0.00 ...5 ..10 .01 5.10.2 00:99.00 .0931 H: 0.00...

00.0
NP50
0010
V000
00V0
V000
5010
0010-
V..10
N010-
VV10
000.0-
00.0
000...

0002. 0104<>> 1001.50 01<>>1 1<>>>>0

000.0-
00.0-
0N50

N000-
0010-
N000-
NN00-
0VNO-
0010-
N00-
5010-
..000

0010-
000.0

000...

0000
N000
000.0
0.00
00V0
00.0
0.10
5N0.0-
0..N.0
VV00-
VVNO
0NNO-
0N00
500.0
0010-
000.?

0510
NPNO-
0.5.0
0N...0
000.0-
0N10
FON0-
50N0-
0010-
050.0-
0010-
000.0
000.0
00N0
N000
0N00
000.?

5010
0V00
r050
V000
0V...0-
V000
00V0-
000.0-
0000
05V0
000.0
FN00
VON0
500.0
0N00
000.0
5V00
000.P

N010
0N...0
000.0
000.0
500.0-
000.0
N00-
0000-
000.0
0000
000.0
N000-
VN00
0000
000.0-
0VVO
000.0-
090.0
000..

0401.41 0401.00

5V00
0N50
0N10
N000
V000
N000
0N10
00N0-
0010
VV10-
0N10
0010-
NV00
500.0
00.0-
000.0
0010
PV00
000.0
000...

0200
0..r0<41
1011.0
v.4<10
1.10200
11.0<41
v.0.10
1.11001
1.1.102
01.50
2011.50
16.10
4010402
000>>
0104<>>
10011.0
01<>>1
1<>>>>0
0401.41
0401.00
4(000
4(0003
01.>>0
4<1-0.210
4<1.0.2
0(010
04_<2>>
042200
0.111.50
001(41
101

104

000.?

0200 0.1.051 1011.0

0.00 ...5 0..-.0 .0. x...m_>. 5.6.9.00 50.001 ”I. 0.00 1.

000 .0
000...

FV10
000.0-
000...

050.0
V050
0N00-
000. P

v.4<10 ...10200 1.051

0N5.0
V? 10
..N10
0000
000;

050.0
V050
0N00-
000. P
0000
000. F

N000
0000
0V00-
V010
N..10-
V010
000.?

v.0.10 1.11001 1<...102

00N0-
0010
V000-
0NNO-
0.0.0-
0NNO-
0.00
000...

0000-
VN00
N 5V0
VV10-
00N0-
VV10-
000.0-
VON0-
000.—

0200
0F0<41
1011.0
¥4<10
...10200
11.051
v.0.10
#11001
1.2102
0..-<40
201.5(40
1(10
4010402
000>>
0104<>>
10011.0
01<>>1
1<>>>>0
0401.41
0401.00
4<000
4.4000...
01.>>0
44.1.0210
4<1.02
0<010
04_<2>>
04.<200
0<111.0
00151
101

105
Table 12: K-Means Cluster Analysis Table for Posthole Sample

Summary Statistic»: for Three Cluster Analysis

VARIABLE BETWEEN 68 DF WITHIN 88 DF F-RATIO PROB
FCR 0.092 2 12.438 65 0.24 0.787
FLAKES 0.43 2 2.438 65 5.735 0.005
SCNAILS 0.061 2 14.984 64 0.13 0.879
WNAILS 0.009 2 2.857 64 0.096 0.99
SCREW 0.485 2 0.5 64 31 .045 0
SHLCAS 0.485 2 0.5 65 31 .544 0
METAL 0.43 2 2.438 65 5.735 0.005
BARWIR 0.008 2 2.859 65 0.094 0.91
UNCOAL 11.391 2 279.359 65 1.325 0.273
BCOAL 1383.721 2 171.75 65 261.839 0
BOTGLS 5.898 2 80.984 65 2.367 0.12
FLTGLS 0.283 2 94.938 65 0.097 0.98
GWWARE 0.045 2 18.234 65 0.08 0.923
WALSHEL 0.059 2 25 65 0.076 0.926
WOOD 0.111 2 99.109 65 0.036 0.964
MOLSHEL 0.004 2 3.938 65 0.03 0.97
CHAR 22.948 2 30.246 64 24.279 0
SLATE 0.295 2 363.234 65 0.026 0.974
BSLATE 0.309 2 63.5 65 0.158 0.854
PENNY 0.001 2 0.984 65 0.03 0.97
PAINCHP 0.001 2 0.984 65 0.03 0.97
BRICK 6428.592 2 917.938 65 227.607 0
PLASTR 0.457 2 1 .484 65 10.002 0
CONCRT 0.361 2 11.109 65 1.057 0.353
CHALK 1 .941 2 2 65 31 .544 0
OTHER 0.001 2 0.984 65 0.03 0.97
MORTAR 2.824 2 525.234 65 0.175 0.84

 

106
Table 12: K-Means Cluster Analysis Table for Posthole Sample

CLUSTER NUMBER: 1
Members Statistics

CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM STDEV.

1 0.23 FCR 0 0.16 2 0.44
2 0.23 F LAKES O 0.03 1 0.17
3 0.23 SCNAILS 0 0.13 3 0.49
4 0.23 WNAILS 0 0.05 1 0.21
5 0.23 SCREW 0 0 0 0
6 0.19 SHLCAS 0 0 0 0
7 0.23 METAL 0 0.03 1 0.17
8 0.23 BARWIR 0 0.05 1 0.21
9 0.23 UNCOAL 0 0.7 14 2.09

10 0.23 BCOAL 0 0.31 9 1 .25

11 0.23 BOTGLS 0 0.27 8 1.11

12 0.23 FLTGLS 0 0.34 7 1 .21

13 0.46 GWWARE 0 0.1 1 4 0.53

14 0.23 WALSHEL 0 0.13 4 0.63

15 0.23 WOOD 0 0.17 10 1.24

16 0.45 MOLSHEL 0 0.03 2 0.25

17 0.3 CHAR 0 0.06 2 0.3

18 0.23 SLATE 0 0.36 19 2.38

19 , 0.23 BSLATE 0 0.13 8 0.99

20 0.28 PENNY 0 0.02 1 0.12

21 0.23 PAINCHP 0 0.02 1 0.12

22 0.23 BRICK 0 0.47 27 3.35

23 0.25 PLASTR 0 0.02 1 0.12

24 0.26 CONCRT 0 0.08 3 0.41

25 0.23 CHALK 0 0 0 0

26 0.23 OTHER 0 0.02 1 0.12

27 0.42 MORTAR 0 0.36 23 2.85

28 0.22

29 0.23

30 4.37

31 0.23

32 0.46

34 0.23

35 0.23

36 0.69

37 3.64

38 5.57

39 0.28

40 0.38

41 0.23

42 0.71

44 0.23

flfl
Table 12: K-Means Cluster Analysis Table for Posthole Sample

0.42
0.23
0.23
1.81
0.28
1.16
0.67
0.29
0.23
0.29
0.23
0.23
2.72
0.62
1.62
1.44
0.31
0.23
0.23
0.43
0.32
1.98

8388982880030200805555

 

108
Table 12: K-Means Cluster Analysis Table for Posthole Sample

CLUSTER NUMBER: 2
Members Statistics
CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

33 1 .97 F CR
43 1.97 FLAKES
SCNAILS
WNAILS
SCREW
SHLCAS
METAL
BARWIR
UNCOAL
BCOAL
BOTGLS
FLTGLS
GWWARE
WALSHEL
WOOD
MOLSHEL
CHAR
SLATE
BSLATE
PENNY
PAINCHP
BRICK
PLASTR
CONCRT
CHALK
OTHER
MORTAR

O
O

. PP .
000000000000
00

C
0

3000000000

UiOOUIOOOOOOOOOOOUI-‘OOO0101000010

O
D

00000
000A08000000000400004400040

000008000OOOOOOO‘OOOOOOOOOO

A
A

109
Table 12: K-Means Cluster Analysis Table for Posthole Sample

CLUSTER NUMBER: 3
Members Statistics
CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

53 1 .37 FCR
62 1 .37 FLAKES
SCNAILS
WNAILS
SCREW
SHLCAS
METAL
BARWIR
UNCOAL
BCOAL
BOTGLS
FLTGLS
GWWARE
WALSHEL
WOOD
MOLSHEL
CHAR
SLATE
BSLATE
PENNY
PAINCHP
BRICK
PLASTR
CONCRT
CHALK
OTHER
MORTAR

0
0

OVQOOOU’IOOOO
.0 .
0000000000000000

on

9°
000000
0
0

0.0

00
.0.
0

0
00-‘001000
0
0|

00M0-‘000-‘4V00000d8000040000
.0 .
00-s00000

000000000000000008000000000

110
Table 13: K-Means Cluster Analysis Table for 5ft. Units -- 3 Clusters

Summary Statistics for Three Cluster Analysis

VARIABLE BETWEEN SS DF WITHIN SS DF F-RATIO PROB
FCR 102.127 2 63.429
FLAKES 5.46 2 13.429 6 4.83 0.056
STFRAG 0.508 2 7.714 6 1.22 0.359
SCNAILS 554.032 2 1002.857 6 0.198 0.826
WNAILS 133.143 2 190.857 6 1 .657 0.267
SHCAS 0 2 0 6 2.093 0.204
METAL 578.286 2 41.714 6 . .
SHMETAL 891.175 2 35.714 6 41.589 0
BWIRE 183.841 2 5.714 6 74.859 0
UBCOAL 4564732571 2 417.429 6 96.517 0
BCOAL 62.794 2 1227.429 6 32806.086 0
BOTGLS 10571.175 2 839.714 6 0.153 0.861
FLTGLS 231 .365 2 20204.857 6 37.767 0
GWWARE 376.889 2 2694 6 0.034 0.966
PWARE 128 2 0 6 0.42 0.675
OTHCER 303.143 2 52.857 6 . .
WALSHEL 1 57804.698 2 16.857 6 17.205 0.003
WOOD 1 93.143 2 24.857 6 28083.887 0
MOLSHEL . 37.841 2 5.714 6 23.31 0.001
CHAR 4296.127 2 37.429 6 19.867 0.002
SLATPEN 0.127 2 3.429 6 344.346 0
SLATE 14 2 56 6 0.111 0.897
MORTAR 2.032 2 40.857 6 0.75 0.512
PECHPT 0.508 2 3.714 6 0.149 0.865
BRICK 4.889 2 32 6 0.41 0.681
PLASTR 72 2 0 6 0.458 0.653
CONCRT 342.286 2 607.714 6 . .
CHALK 854.222 2 0 6 1 .69 0.262
OTHER 106 2 96 6 . .
PLASTIC 13.841 2 9.714 6 3.313 0.107
BONE 62.508 2 3.714 6 4.275 0.07
6 50.487 0

111
Table 13: K-Means Cluster Analysis Table for 5ft. Units -- 3 Clusters

CLUSTER NUMBER: 1
Members Statistics

CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

1 6.29 FOR 0 3.29 7 3.01
2 7.9 FLAKES 0 0.71 4 1.39
3 4.94 STFRAG 0 0.43 3 1 .05
4 5.26 SCNAILS 0 5.86 35 11.97
6 3.96 WNAILS 0 3.86 13 5.22
7 26.26 SHCAS 0 0 0 0
9 5.94 METAL 0 1.57 7 2.44
SHMETAL O 1 .57 6 2.26

BWIRE 0 0.57 2 0.9

UBCOAL 0 6.71 21 7.72

BCOAL 0 8.29 38 13.24

BOTGLS 0 7.57 33 10.95

FLTGLS 0 26.86 158 53.73

GWWARE 0 9 57 19.62

PWARE 0 0 0 0

OTHCER 0 1 .86 8 2.75

WALSHEL 0 1 .14 4 1 .55

WOOD 0 1 .14 5 1 .88

MOLSHEL 0 0.57 2 0.9

CHAR 0 1 .71 7 2.31

SLATPEN 0 0.29 2 0.7

SLATE 0 3 8 2.83

MORTAR 0 1.14 7 2.42

PECHPT 0 0.57 2 0.73

BRICK 0 2 6 2.14

PLASTR 0 0 0 0

CONCRT 0 5.43 27 9.32

CHALK 0 0 0 0

OTHER 0 3 10 3.7

PLASTIC 0 1.57 3 1.18

BONE 0 0.57 2 0.73

Table 13: K-Means Cluster Analysis Table for 5ft. Units -- 3 Clusters

CLUSTER NUMBER: 2
Members

CASE DISTANCE
8 0 FCR
FLAKES
STFRAG
SCNAILS
WNAILS
SHCAS
METAL
SHMETAL
BWIRE
UBCOAL
BCOAL
BOTGLS
FLTGLS
GWWARE
PWARE
OTHCER
WALSHEL
WOOD
MOLSHEL
CHAR
SLATPEN
SLATE
MORTAR
PECHPT
BRICK
PLASTR
CONCRT
CHALK
OTHER
PLASTIC
BONE

14
3

0
31
15
0
27
0

1
2273
9

1 17
40
17
3
20
28

—l
05

(JON
00A4#0000000\l

Statistics

VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

14 14 0
3 3 0
0 0 0

31 31 O

15 15 0
0 0 0

27 27 0
0 0 0
1 1 0

2273 2273 0
9 9 0
117 117 0

40 40 0

17 17 0
3 3 0

20 20 0

28 28 0

16 16 0
7 7 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
3 3 0
9 9 0

24 24 0

31 31 0
4 4 0
5 5 0
9 9 0

113

Table 13: K-Means Cluster Analysis Table for 5ft. Units -- 3 Clusters

CLUSTER NUMBER: 3

Members

CASE DISTANCE VARIABLE

5 0 FCR
FLAKES
STF RAG

SCNAILS
WNAILS
SHCAS
METAL
SHMETAL
BWIRE
UBCOAL
BCOAL
BOTGLS
F LTGLS
GWWARE
PWARE
OTHCER
WALSHEL
WOOD
MOLSHEL
CHAR
SLATPEN
SLATE
MORTAR
PECHPT
BRICK
PLASTR
CONCRT
CHALK
OTHER
PLASTIC
BONE

Statistics

MINIMUM MEAN MAXIMUM ST.DEV.

A

d0
00004000400

0.90.;
0N00-I

425

N
009

40‘040000004

—L

—L

‘00
00010040004003

ANA—L
ONOO-fi

425

N
d0¥040000004003

—L

d

an:
0000-3000400

ANA—L
ONOCD-h

425

N
fi0h0-‘000000-‘00

..l

0000000000000000000000000000000

1 14
Table 14: K-Means Cluster Analysis Table for 5Ft. Units - 5 Clusters

Summary Statistics for 5 Cluster Analysis

VARIABLE BETWEEN SS DF WITHIN SS DF F-RATIO PROB
FCR 1 14.756 4 50.8 4 2.259 0.225
FLAKES 6.889 4 12 4 0.574 0.698
STFRAG 8.222 4 0 4 . .
SCNAILS 1546.089 4 10.8 4 143.156 0
WNAI LS 320.8 4 3.2 4 100.25 0
SHCAS 0 4 0 4 . .

METAL 619.2 4 0.8 4 774 0
SHMETAL 926.089 4 0.8 4 1 157.61 1 0
BWIRE 186.356 4 3.2 4 58.236 0.001
UBCOAL 45648028 4 347.2 4 1 3147.474 0
BCOAL 1 095.022 4 195.2 4 5.61 0.062
BOTGLS 1 1326.089 4 84.8 4 133.562 0
FLTGLS 20315.422 4 120.8 4 168.174 0
GWWARE 3066.089 4 4.8 4 638.769 0
PWARE 128 4 0 4 . .

OTHCER 348 4 8 4 43.5 0.001
WALSHEL 157814.756 4 6.8 4 23208.052 0
WOOD 21 0.8 4 7.2 4 29.278 0.003
MOLSHEL . 40.356 4 3.2 4 12.61 1 0.015
CHAR 4299.556 4 34 4 126.458 0
SLATPEN 3.556 4 0 4 . .

SLATE 52.8 4 17.2 4 3.07 0.151
MORTAR 42.089 4 0.8 4 52.61 1 0.001
PECHPT 1 .022 4 3.2 4 0.319 0.853
BRICK 16.089 4 20.8 4 0.774 0.595
PLASTR 72 4 0 4 . .

CONCRT 889.2 4 60.8 4 14.625 0.012
CHALK 854.222 4 0 4 . .

OTHER 195.2 4 6.8 4 28.706 0.003
PLASTIC 16.756 4 6.8 4 2.464 0.202
BONE 65.422 4 0.8 4 81 .778 0

Table 14: K-Means Cluster Analysis Table for 5Ft. Units - 5 Clusters

CLUSTER NUMBER: 1

CASE

(DUI-500d

Members

DISTANCE

3.15
2.25
2.66
2.36
2.24

VARIABLE

FCR
FLAKES
STFRAG

SCNAILS
WNAILS
SHCAS
METAL
SHMETAL
BWIRE
UBCOAL
BCOAL
BOTGLS
FLTGLS
GWWARE
PWARE
OTHCER
WALSHEL
WOOD
MOLSHEL
CHAR
SLATPEN
SLATE
MORTAR
PECHPT
BRICK
PLASTR
CONCRT
CHALK
OTHER
PLASTIC
BONE

115

MINIMUM

0000000000000000000000000000000

Statistics

MEAN MAXIMUM ST.DEV.

3.8

1

0
1.2
0.6

0
0.2
0.2
0.4
8.6
3.6
3.2
4.2
0.8

0

1
0.8
0.6
0.4

2

0
1.6
0.2
0.6
2.8

0
2.2

0
0.8
1.8
0.2

Nd-bONwO'bN

d—LN
004

d
dUUOOOQN-‘UONNwaONQ

3.19
1.55
0
1.47
0.8
0
0.4
0.4
0.8
8.33
6.25
4.12
4.92
0.98
0
1.26
1.17
1.2
0.8
2.61
0
1.85
0.4
0.8
2.04
0
3.49
0
1.17
1.17
0.4

116
Table 14: K-Means Cluster Analysis Table for 5Ft. Units - 5 Clusters

CLUSTER NUMBER: 2

Members Statistics
CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

8 0 FCR 14 14 14
FLAKES 3 3 3
STFRAG 0 0 0
SCNAILS 31 31 31
WNAILS 15 15 15
SHCAS O 0 O
METAL 27 27 27
SHMETAL O O O
BWIRE 1 1 1
UBCOAL 2273 2273 2273
BCOAL 9 9 9
BOTGLS 117 117 117
F LTGLS 4O 40 4O
GWWARE 17 17 1 7
PWARE 3 3 3
OTHCER 20 20 20
WALSHEL 28 28 28
WOOD 16 16 16
MOLSHEL 7 7 7
CHAR O O 0
SLATPEN O 0 O
SLATE 0 O 0
MORTAR O 0 0
PECHPT O 0 0
BRICK 3 3 3
PLASTR 9 9 9
CONCRT 24 24 24
CHALK 31 31 31
OTHER 4 4 4
PLASTIC 5 5 5
BONE 9 9 9

0000000000000000000000000000000

117
Table 14: K-Means Cluster Analysis Table for 5Ft. Units — 5 Clusters

CLUSTER NUMBER: 3
Members Statistics
CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

5 0 FCR
FLAKES
STFRAG

SCNAILS
WNAILS
SHCAS

METAL

SHMETAL
BWIRE

UBCOAL

BCOAL

BOTGLS
FLTGLS
GWWARE
PWARE
OTHCER

WALSHEL 425 425 425

WOOD
MOLSHEL
CHAR
SLATPEN
SLATE

MORTAR

PECHPT
BRICK

PLASTR
CONCRT

CHALK
OTHER
PLASTIC

BONE

u—S
—L
.8

AU
0000-5000-5003

‘0’
OOUIU-‘OOO-‘OO,

‘00
000100-‘000-‘003

...sN—L—L
MOO-I-
dN—L—h
NCDCD-h
ANA—s
N004

‘1
000

N
40#0-*000000-5

d0#0-‘000000-‘00’
N

d0#0-‘000000-¥00J

A
d

sub
0
0
0
0000000000000000000000000000000

1 18
Table 14: K-Means Cluster Analysis Table for 5Ft. Units - 5 Clusters

CLUSTER NUMBER: 4
Members Statistics
CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.

7 0 FCR
FLAKES
STFRAG
SCNAILS
WNAILS
SHCAS
METAL
SHMETAL
BWIRE
UBCOAL
BCOAL
BOTGLS
FLTGLS
GWWARE
PWARE
OTHCER
WALSHEL
WOOD
MOLSHEL
CHAR
SLATPEN
SLATE
MORTAR
PECHPT
BRICK
PLASTR
CONCRT
CHALK
OTHER
PLASTIC
BONE

d0}

NON«F\J0-*UIUO%
‘0

U
UNONhVO-‘U’UO-fi
‘0

U
“N0N45N0-‘UIO’0uh

158

..s
010'“
N0“

d
010'
~10

0|
0‘]

d
dN000000VONNNU|000

—l
dNOOOOOONONNNUOOO

d

dN000000NONNN000
0000000000000000000000000000000

 

119
Table 14: K-Means Cluster Analysis Table for 5Ft. Units - 5 Clusters

CLUSTER NUMBER: 5

Members Statistics

CASE DISTANCE VARIABLE MINIMUM MEAN MAXIMUM ST.DEV.
2 0 FCR O 0 0 0
FLAKES O O 0 O
STFRAG 0 O 0 0
SCNAILS O 0 0 0
WNAILS 1 3 1 3 1 3 O
SHCAS O O 0 0
METAL 3 3 3 O
SHMETAL 6 6 6 0
BWIRE O O 0 O
UBCOAL 4 4 4 0
BCOAL 38 38 38 O
BOTGLS 4 4 4 0
FLTGLS 9 9 9 O
GWWARE 2 2 2 0
PWARE 0 O O 0
OTHCER 0 0 O 0
WALSHEL 4 4 4 0
WOOD O O 0 0
MOLSHEL O 0 0 0
CHAR O O O 0
SLATPEN 0 0 O O
SLATE 5 5 5 0
MORTAR O 0 O 0
PECHPT 1 1 1 O
BRICK 0 O O 0
PLASTR O 0 0 0
CONCRT 27 27 27 0
CHALK O O 0 0
OTHER 7 7 7 O
PLASTIC O 0 0 0
BONE 2 2 2 O

APPENDIX B
MAPS AND GRAPHS

120

Map 1:
Map of Michigan, Showing Lansing and
the Location of the Merle Beach School Site

  

Merie Beach School Site

" Lansing

 

121

 

C Map 2: Merle Beach Schoolhouse Site Map 3

 

 

V N1165

a;
O
O
o

I

 

 

 

 

 

Merle Beach
Schoolhouse

 

o
o
5
<5
I

 

 

>
J
>

 

 

V moss

 

(P
O
I
>
J

> <3 0 <5

N1015
151000 A E1025 E1050 A E1075 A E1100 A 51125

V V v V W

 

 

 

 

 

 

 

 

0 Post holes I 5x5 Test Units _ O Pre-existing obstacles
I I 1 I

 

 

 

 

122

Map 3:
Surface Collection Contour Map for Bottle Glass

150 L I 1 I I I 1 1 l

 

135 3 —

1204 000

60 Q -

90+ e
9

>- 75~ 0%

50.. -

00

454 ~
30“ I-

15 J _

 

 

 

O I I I I I I I I I

X

Map 4:
Surface Collection Contour Map for Metal

 

150 I J l l I l I I i
135 ~ —
120 s
105 T o

O
90 s
> 75 — “‘5

60*

 

45—

BO~ —

 

 

 

 

 

O
O
)
“N
(j
p
r 5‘
K, I
()3
«X
0‘
O“:
o
O
G)\
xx-
L
.4
0‘
O
‘8.

123

Map 5:
Surface Collection Contour Map for Other Artifacts

I I l I L

 

150

135

1 [fit J
____2./ /

120
105
90
> 75
60
45

80-

151

 

   

A

 

 

O #0 r r

I I I I I

00 (3.5 25.0 37 E3 500 62.5 75.0 87 EJQQOMQEDAflgO

X

Map 6:
Surface Collection Contour Map for Plastic

 

 

 

 

150 l \\ l i ' Ix I I J l l 1 ‘1 \
\\ (If ’1 K
135 \ °/ ° ~
\v IW/
12o ' r
90 11 "v x ~
> 75 i ‘ fixJ ”
60 —J ‘1‘ r
5 /
40 '4. 0.x \ ‘/ '-
30 ] hf. ”
15 ‘. C a / IO ‘—
‘11‘ r ,1
3 J I
O '1 1 f I I I I r I
n r' ’1 ' i 6‘
QO UE 9V 0 Q. E. 50C 625 75.0 87 ”10001 @2250

Y

124

Map 7:
Surface Collection Contour Map for Wood
1 50 I I 4 1 l l I J J

 

135

120

105

90

75

60

45

SO

15

 

 

 

 

O I I I I I I I I I

00 «25 25.0 '37 5 500 60.5 75.0 81 5100-01 «2.5.25.0

 

 

 

><
Map 8:
Surface Collection Contour Map for Bone
150 l I i i I l I l 1
135 4 _
00
120 ~ ~
.05 l
90 —
75 1’ Q
60 —
0
.5 - 6
08
15 -1
>%
O I I I I.[ I

 

 

 

X

125
Map 9:

Surface Collection Contour Map for Plaster

150

135

120

105

90

75

60

45

30

15

O

 

1 1 1 1 1 1 1 1
_ \ _
Q 9
fl 0 \I-
’0
- O _
_. I—
0
Q % O
1 -
} O
0 r
WAT 1 1 1 1 i 1 T

 

 

 

 

X

Map 10:

Surface Collection Contour Map for Shingles

150

50

45

3O

15

O

l

I I I

 

<1) 0

j

I I

I

 

 

 

I I I I I I I I

00 «”225 250 315 500 6’25 160 81 5100011151250

126

Map 11:
Surface Collection Contour Map for Slate

150 I I I I L I I I L

 

135 ~ —

120 '-

105 0

45‘

530-1

15-1

 

 

 

 

o
00 «215 250 31 .5 500 625 15.0 31 5.000.125.2610

X

Map 12:
Surface Collection Contour Map for Concrete

150 I I I I I 1 I l L

 

1351 E
120 J
105 —
901

75-

I

60

454

30-

15"

 

 

 

 

O 1 f 7 T I
00 (A5 25C) 37 E1 500 625 150 $1 5100011251250

_1
.0
d
.4

X

127

Map 13:
Surface Collection Map for Flat Glass

 

150 1 I 1 1 1 1 I l ’

135 —/ —
120 ~ 1-
105 - o F

90- —
> 75-+ 0 e
60- ' _
454 ~
30— —

15« ~

 

 

 

A
O I I I I I I I I I

0.0 (25 250 37 .5 500 596 75.0 $1 5,009 (22.250

X

Map 14:
Surface Collection Contour Map for Square Nails

150 I I L l I I L I I

 

135 ~ —

120 -

105 0

903

> 75‘

50*

 

45~

30~ ~

15— —

 

 

O A A #

I I I I I I I I I

 

()9 17,5 259 31 5 500 6’26 150 81 200% \2225-0

><

128

Map 15:
Surface Collection Contour Map for Unbumed Coal

 

I50 I I I I I I I I I
135 ‘I b

120 *

°°Q
1051 (30 /q *
90~ L —

05

> 75— Q _

O
O

60‘ ~

00

45~ _
304 _

15~ ~

 

 

 

 

 

O I I I I I l I T I

00 \7’5 250 37 5 500 525 150 E57 6N0001\26‘\260

X

Map 16:
Surface Collection Contour Map for \Mre Nails

150 I I I I I l I I I

 

1:35 - ~
120 a 0
105 — _

(01
90 —
> 75 - (3}

50—

 

453 e
30~ —

15— 1-

 

 

 

Q .n A ‘

I I I I I I j . I

09 95 25.0 31 5 59.0 635 750 87 5.909. 125.0259

X

129

Map 17:
Posthole Contour Map for Paint Chips

 

150 L\‘/1 1 I 'fwf‘o
/ WMN K

C)

135 —~

120 4

105 -‘

90

75

50

45

30

155

 

 

00
/I\

 

 

O

I I I I I

I

X

Map 18:

Posthole Contour Map for Other Artifacts

L

# J 1 IV
I.
"I
l
1

150
120
..I

10:: '

90 E2.

7.. {3; O
,' ‘Q

504

.1: _I‘

30

7 I I I I I I I

I

 

(7 \o o O
. \Q I C‘ r p, \? 0
A \ u i ‘\ A I\ / \I

X

130
Map 19:

Posthole Contour Map for Mollusk Shell

150

135

120

105

90

75

60

45

3O

15

O

p.

I ; i

 

O

 

144

V\
O

L

    
 

 

 

00 125 25.0 31 .5 500 625 150 9.1 51003122250

X

Map 20:

Posthole Contour Map for Burned Slate

150

135

120

105

90

75

5O

45

30

15

O

 

 

\ AWJ

 

 

 

00 125 250 37 5 50.0 625 150 g1 2000025960

X

131
Map 21:

Posthole Contour Map for Fire Cracked Rock

150

 

135 -

120 '-

105 -

90—

15-

 

O

914
p
Q:
7 g o _
.2. oi <1
45* 00% _

.. 5
x.

M\

W.
6:) Mr

I

I

I

I

 

 

00 125‘ 250 31 5 500 5’25 150 91 5 000 «2“ 250

X

 

Map 22: Posthole Contour Map for Wrre Nails
1 50 o /I I I I 1 lo I I I

135 ~—/ V K \K—
120 ~ 0 0 I \‘QS/e

0 g5 (5
105 L
90 — I, $1 2 ~

‘~. \v' \\_/

75 — g G O o —

 

 

 

45 — —
0
30 — 1—
15 -\ _
0
o \‘W//A\‘QF\‘ Wk fl 6/
I I I I I I I I
00 «1‘3 259 431 5 500 5’25 150 {51 2000.1251950

X

150

135

120

105

90

75

50

45

30

15‘

O

1 32
Map 23:
Posthole Contour Map for Concrete

 

. O '
o <

 

 

 

0.0 «2.5 r250 37 .5 50.0 625 750 BI 5’\OQQ\ (lbw/5.0

150

135

120

105

90

75

50

45

3O

15

O

X

Map 24:
Posthole Contour Map for Slate

 

 

 

 

/
I1 1’ \x I \ 'L_J/
I I I I I I I I I

 

0.0 «25 250 .37 .5 50.0 61.5 75.0 87 5N000flgbflgf)

X

I

133

Map 25:
Posthole Contour Map for Charcoal

 

 

150 I I L I_ I 1 1 V 1 \0
135 0 .1], r-
o

105 — —
90 —
75 — —
so —
45 —

30-

    

15—
o \ LA

I
00 1’25 250 31 5 500 6’25 150 <51 5.90on (Pl-5.950

 

 

 

X

Map 26:
Posthole Contour Map for Chalk

150 I 1 I

 

135 s

120 4

105 -

 

45* —

SO- —

 

 

 

I I I I I r
00 «26 .250 ,3; 5 50,0 6'35 4'59 9] 5109011961250

X

134
Map 27:

Posthole Contour Map for Shot Gun Shell Cases

>_

>.

150

185

120

105

90

75a

60

45

BO

15

O

l

 

l 4 l l L l l
O

 

 

 

 

00 «2.6 25.0 37 .5 50.0 6'25 15.0 6 5,00% «2,225.0

150
135
120
105
90
75
60
45
30

1a

0

 

X
Map 28:
Posthole Contour Map for Plaster
l 1 l l l 1 V1 "\1_./1 \

/d\\r"\

9 m N ' rm
1 O 0 0 0%

O x\__/O
Q

00
\\ ‘fIc, jf\\\ /

\___(

I

 

 

 

f‘\ M; 4
T T l T T r T l

00 \25 369 37 5 500 5’25 750 g? 5‘0ng (12250

X

>_

>.

150

185

120

105

90

75

50

45

30

15

O

1 35
Map 29:

Posthole Contour Map for Metal

L l

 

   
 

K x111 \11/ L

q> 000 C—
I ~
_I 1.
I 11 _

 

 

 

O-O V25 950 37 5 500 6‘26 750 87 200% (12250

150

135

120

105

90

75

60

45

30

15

O

X

Map 30:

Posthole Contour Map for Brick

 

//

L

 

 

' V
K
\fo

 

I I I I 7 I

X

150

185

120

105

90

75

60

45

8O

15

O

136

Map 31:
Posthole Contour Map for Wood

 

—-4

l llllll
;/ V71 VVW\

 

V

 

 

\mm/KA/

 

0.0 «2.5 250 37 .5 50.0 625 750 87 5\OQO\\25\250

150
185
120
105
90
75
50
45
80
15

O

X

Map 32:

Posthole Contour Map for Pennies

 

l

 

 

 

 

()0 125 250 37 ‘5 50.0 62.5 15.0 g1 2000025950

X

137

 

Map 33:
Posthole Contour Map for Glazed White Ware Ceramics
150 1 1 11 {/1 /.2\\ I\J‘ 11
135 4K @ l-
120 -> 0 Q r
105 C3

90

75

60

45

80

15

 

 

 

O

 

\k/CXL/r}\1 A ka/

j

00 «25 250 37 5 50.0 625 150 g’I 2000\«22250

 

><
Map 34:
Posthole Contour Map for Barbed \Mre
150 l l l l J l l l
W 010W koa‘
135 I Q @0 —
. 00 ./
120 ~ ‘ :,( ~
105 " 9 \P
90 ~ (x0 —
75 -/ K o \09 _
\
60 a \1 \_
v 0
4 4 _
5 01.0
80 —1> —
15 J’\ 00 C? /L
o We f‘\1/1\%\%

 

 

 

\/
A

1 38
Map 35:

Posthole Contour Map for Bottle Glass

150

185

120

105

90

75

60

45

80

15

O

l l I

 

1)» \J

l

l

 
 
 

11: \27
01%? CDC
[1

Oo

    

 

A M./
l I I I

 

 

I I

0.0 «2.5 25.0 37 .5 50.0 62.5 15.0 BI .5\00.0,\»\Q,.5,\25.O

150

185

120

105

90

75

60

45

8O

15

O

X

Map 36:

Posthole Contour Map for Flat Glass

1 l

1 l l l

 

l
\

 
  

O

Q

\l_,/
l
P
0

l
V

C)

  

 

n
f

 

I I I I I I I I

 

0.0 (2.5 250 37 5 50.0 62.5 15.0 €17 200% (2,225.0

y.

150

185

120

105

90

75

60

45

8O

15

O

139

Map 37:
Posthole Contour Map for Screws

 

pg 1 1
1? (30 o o o 14:
- <_
_ <:
\A AA/z 1/

 

 

 

00 V25 250 31 5 500 6’25 150 <51 5\QOO\\’Z-51\250

X

Map 38:

Posthole Contour Map for Walnut Shells

150

185

1204

105

90

75

5O

45

8O

15

O

 

 

 

 

 

I I I T I I I I

I

00 «25 350 37 ”'3 500 625 150 37 2900\52250

X

140

Map 39:
Posthole Contour Map for Burned Coal

150 /\If1 I IVIIQI

 

 

135 A _
(‘9 <7
120 -* . —
o
105 —> (L
90 /"
75 / —
60 l—
45 _
30 _ *-
15 — —

 

 

 

o K AAAN

00 125 250 37 6 500 61.5 15.0 6! 2009122250

 

 

 

X
Map 40:
Posthole Contour Map for Unbumed Coal
150 J ‘fi\ %\ 1 : i} I A 4 l 1 1 1
1 351 \Q l : WQ ’” o
135 1 Q; \, (V) _
120 1/ P
105 —1 L
l A
9C) 1M (E \KJ/M {—0 < +—
_/ ) ,’
75 A ‘ : L
i JA 7
50 -: , /"A l _
l '

5 T KT : _
4a A!" \' (5L1 F MA: 7‘ f \\\A/
30 1; l V _

i ‘\‘
15 —{ \ {P
l y /
l a 4 / A” ~ 4-NV4/fl\\__/AJ\\\J
' ‘ T I 7 T 1 I T T
L0 .3,/6 ,va J r; '1) 50C 625 7150 81 6\000\ \25\25,C)

X

141

Map 41:
Posthole Contour Map for Chart Flakes

 

150

135

120

105

90

>75

60

45

30

 

15

 

 

O A”: MAJ/

00 (2’5 (250 37 E) 500 625 750 %7 6\OOO\‘\26\250

 

X

Map 42:
Posthole Contour Map for Square Headed Nails

150 Pk 1

135 a

O
1201?
105 ~ \

90-4

 

> 75~
60 a.
45a

30 >

15 “x 0 "
O \D F\ /\Q\ [\E\ /\\\/
I I I T I I I I , I

00 «25 260 31 5 500 6'25 160 a1 510001 «251259

 

 

 

 

x

150

185

120

105

90

75

60

45

30

15

—1

 

142
Map 43:
5ft. Unit Contour Map for Wood

I

I

I

I

d
I

 

 

O

I

I

 

I
12 5x25 0

 

 

 

 

OO «:26 250 31 6 500 62.5 15.0 ‘57 5\OOO'\
x
Map 44:
5ft. Unit Contour Map for Square Nails
150 l I l l l l I L 1
lo
135 — ~
120 — —
105 - —
90 — _
\\
75 a _ I _
/
60 q/j/ ._
45 — ~
30 — —
15 — A 0 .—
O I I I I l I I I I
00 v25 250 37 b 500 625 150 81 2000022150

X

143

Map 45:
5F . Unit Contour Map for Fire Cracked Rock

150

”

  
 

 

135 -l
120 J

105 A

90-
75*
45‘

O I I I I I m I

30—
IsJ

 

 

 

 

00 «(2’5 250 97 E) 500 625 150 91 E)\QOO\'\'ZE’\’ZE>O

X

Map 46:
5Ft. Unit Contour Map for Chert Flakes
150

1 l l l l I l l l
135 " \b—
120 ‘ ’—
2
105 "

 

90

75

60 ~

45— ..

30-

 

 

 

 

I I
00 «25 250 31 6 500 525 150 (57 2000022250

X

144
Map 47:
5Ft. Unit Contour Map for Brick

 

120

105

90

75

60

45

30

15

 

O

 

 

I I

X

Map 48:

5 Ft. Unit Contour map for Bottle Glass

1 l l J I l

I I
00 «7’5 250 31.5 600 626 150 87 5\000\\25\25O

 

150
135 d

120 —

90—
> 751
60~

45—

 

O

I K \

T

 

 

I I I I I I I I I

«25 260 3’! 5 500 625 150 51 209% «25969

X

 

145

Map 49:
5Ft. Unit Contour Map for Bone

150 1‘4
III

R.

J_4I_l—L__I__l_

185 n
120 - '
O
105 -‘

90—

75-

45—

I

sol
30-
I

 

 

l
O
I I | I l

15
O

 

00 «2.6 250 31 6 500 6‘25 15.0 87 EKOOOI @2250

X

Map 50:
5Ft. Unit Contour Map for Concrete

 

150 l I 1 J l I I

I L
135 — ’ v
120 — —
105 — —
90 — —
75 — —
60 — —
45 ~ -
so —

15 — o

o I

I I I I I I I I

I

I

 

 

 

00 «’25 250 g’I 5 500 6’25 150 g’I 5\QOOI\ \Q-5\'25-O

X

146

Map 51:
5Ft. Unit Contour Map for Wire Nails

150

 

I

135

120

105

90

75

60

 

 

 

 

 

 

 

 

O l I I I I I I I I I
00 «1.6 25.0 37 .6 50.0 6'25 15.0 87 500% \gbflgO
X
Map 52:
5Ft. Units Contour Map for Metal
150 l l i l l l I l
\ \ \ w
135 a \ \‘v—
\
120 — \ N_
\_
105 “N \‘. wk
90 3 “ A ,»-\\\ —
75 “ I ‘I‘ ‘\ .—
I. I I ,f , \
60 ‘1 $7 [I f I], (r -
/ r v' I
45 I" 7 I 2 <.\ ~
30 [,1 Ii}, //, > ,’ \\\
-‘ / r / \ >—
: El \4/ \\
15 « y/ \o _
\
O I I I I fit 1 I I IV
00 .16 25 _» $1 5 500 616 750 %7 5.\QOO,\\’25.\250

X

147
Map 53:

 

5Ft. Unit Contour Map for Unbumed Coal
150 I 1 . l I ' L 1 1
135 — V—
120 — I —
105 — ‘ ~

90

75

6O

45

30

I

15

 

 

 

I I

I
«’25 260 37 5 500 6’25 150 31 E’NQO-ON 0,2250

X

Map 54:

5Ft. Unit Contour Map for Barbed Wire

In-

I

l

 

150

135

120

105

90

75

60

45

l

80

15

 

 

 

O
O

O

X

148
Map 55:
5Ft. Unit Contour Map for Sheet Metal

 

150

185

120

105

90

75

60

45

30

15

 

 

o
00 «15 250 e1 5 500 6’15 169 51 9000015559

 

 

X

Map 56:
5Ft. Unit Contour Map for Charcoal

,_.

 

”0552/ /‘ ’ ' ‘ ‘

135~\.L/ ~
120~ _
105.: ~,
90—
754

60*

30* ~

 

 

0 7
(30

I

I

X

«25 25C" 31 5 500 @253 150 :51 2900\112950

 

149

Map 57:
5Ft. Unit Contour Map for Misc. Ceramics

 

 

 

 

‘16 25.0 31 f) 50.0 6'15 150 87 @009 @2250

X

Map 58:
5Ft. Unit Contour Map for Mollusk Shell

150

 

135

120

105

90

>- 75

60

45

80

 

OfiI

l I I I

X

150
Map 59:
5Ft. Unit Contour Map for Other Artifacts

 

 

 

 

O I

 

I I I I | I

I I

X

Map 60:
5Ft. Unit Contour Map for Slate

150 '

  
      

 

185 —I

120 -‘

/<

105 r

904

>- 75a

60-

45‘

30-

15*

 

 

O l
00 «(2.5 250

 

I T I I I I I

c1 5 500 625 159 e1 9009\2542‘50

X

150

135

120

105

90

75

6O

45

30

15

O

151

Map 61:
5Ft. Unit Contour Map for Mortar

I I I I I I I I I

 

 

O

 

 

 

 

I I I I I I I

150

135

120

105

90

75

60

45

80

15

X

Map 62:
5Ft. Unit Contour Map for Plastic

 

 

 

 

 

QO

 

I T I

125 250 (51 5 500 6’25 150 $1 610005 @2250

X

152

Map 63:
5Ft. Unit Contour Map for Flat Glass

1* I

150 1 1 1 i 1 1 1

1354 0 \ \-
120-

105—‘5

90-

45~

154 \ 0
O I I I I I I

 

m\

 

 

 

 

X

Map 64:
5Ft. Unit Contour Map for Pearl and Transfer Print Ceramics

 

150

 

Y
\I 0
(Ti 3
I _ “IL“
//I
r
/
/
/
[II
’F/
I I

Q)
C)
I __
/
I

 

X

153

Map 65:
5Ft. Unit Contour Map for Chalk

 

150 1 I ' J—L ' l '

185 ~ M
1204 20 —
105 - ‘° —
90 1 -
75 - —
50 — ~
45 - e r
so — 9 ~
15 - 0 I-
o r I 1

 

 

 

()5 \25I gb-OI 31 5I 505I 67,5 165I g1 5\005\‘\’Z5 1250

X

Map 66:
5Ft. Unit Contour Map for Plaster

b
2%

I I

00 a5 250 131 5 500 525 150 g1 5 QOO,\‘\’Z54255

150

 

1354

I

120 —
105 ~

90-

60 ~°
45 —

30‘

 

 

 

O

 

X

154

Map 67:
5Ft. Unit Contour Map for Walnut Shells

150 W1 l J 1 1 IQV/
135 —* 2 "‘

120 —

 

L—

105

l

 
 
 

90—
> 754
60—

45~

30-
15— /
I‘ I I /I I

O
00 (L5 (250 37 E) 50.0 62.5 150 87 5\OQO\\2,5\250

 

 

 

I I I I

X

Map 68:
5Ft. Unit Contour Map for Glazed White Ware Ceramic

150 l l 1 l l l
135 5/

 

r

120 d —
105 “R —
90 J p
>— 75 — e
50 - ,/ —

 

 

 

,/

45 ‘/ F

%
30 1 /\\\ \ >—
15 _/ ‘0 I» \ \ _

\
/’\\ X \\\
O I )I I I I I I I I ’

 

00 «’15 250 37 5 500 5’25 I50 87 5x000\‘\16«250

x

 

155

Map 69:
5Ft. Unit Contour Map for Peach Pits

 

150

135

120

105

90

75

60

45

30

 

I

 

 

I

V25 9,50 37 5

I

I
500

I l I

:31 5\0001\ «25950

I
5‘25 I50

X

Map 70:
5Ft. Contour Map for Burned Coal

150

 

135 %
12o %
105
90
75
60
45
30

15

V—

 

 

O

I I l
0.0 «25 250 31 .6 50,0

 

I I I

625 760 efI 5’\OOO’\ (2226.0

I I l

x

156

Map 71:
5Ft. Unit Contour Map for Stone Tool Fragments

150 l l 1 l l l l l l
135
120
o
105 -
90 I.
> 75

 

l
I

I

i

I

l

l
I

 

 

 

50 — —
45 — _
:30 - e
15 — —
o
00 «25 25.0 37 .5 50.0 5’15 15.0 9315251250
x
Map 72:

5Ft. Unit Contour Map for Slate Pencils

150 l 1 l l 1 l l l l

 

135 -*

120 d

105 —

90—

 
 

>- 75-

605

454
30— \
15— x
O

I I5)

 

 

 

05 v25 255 575 505 525 I55 57 5x050\ «7 6x250

X

 

<ODCDCQQ‘W

‘<OD(DC.Q¢D"|-«

157

Graph 1:

Frequency Distribution for
Surface Collection

 

 

0 .
0-6 13 20 27 34 41 48 55 62 69 76 83 90 97 104111118125132139
numbers ofartilacts

+ frequency

Graph 2:

Frequency Distribution 2
Surface Collection

 

 

o-
0-51117 23 29 35 4147 53 59 65 7177 83 89 95101107113119
numbersofartifacts

+ frequency

ODQCqu-b

OSGCDGW-t.

  

158

Graph 3:

Frequency Distribution for
Post Hole Tests

d
- '9
1

d
9 <.=
Y Y

 

 

5:1
4::
2‘.
or
1-7 15 23 31 39 47 55 63 71 79 87 95 103111119127135143151159
numberofartifacts
+frequency
Graph 4:

Frequency Distribution for
Post Hole Tests

-715 23 31 39 47 55 63 71 79 87 95103111119127135143151159
number of
+ frequen

‘<O=@C.D0"-h

<O=¢=anfi

159

Graph 5:

Frequency Distribution for
5'x5' Test Units

 

 

943 287 431 575 719 8631007129514391583172718712015215923032447259127352879
number of artifacts
+ frequency
Graph 6:
Frequency Distribution 2
5'x5' Test Units
30.:
251
201
15
104»
5:.

   

 

0‘
123 247 371 495 619 743 867 991 111 123 136 148 161 173 185 198 210 223 245
number of artifacts

+ frequency

<03QEQGH-h

ODQCDCD-I-n

160

Graph 7:

Frequency Distribution 3
5'x5' Test Units

 

 

24 49 74 99 124 149 174 199 224 249 274 299 324 349 374 399 424 449 474
number of artifacts

+ frequency

Graph 8:

Frequency Distribution 4
5'x5' Test Units

 

 

0.0‘
35 71 107 143 174 215 287 323 359 395 431 503 539 575 611 647 683
number of artifacts

+ iequency

161

Graph 9: Surface Frequency
Distribution 1 Artifacts Box Plot

 

 

 

 

 

,—

1 l J

O 50 100 150
ARWFACT

Graph 10: Surface Frequency
Distribution 1 Frequency Box Plot

[l o
l l 4 J

~10 O 10 20
FREQUENC

Graph 11: Surface Frequency

 

 

 

 

 

 

Distribution 2 Artifacts Box Plot
l__1 TL___A
O 50 100 150
ARTIFACT

Graph 12: Surface Frequency
Distribution 2 Box Plot

FREQUENC

162

Graph 13: Posthole Frequency
Distribution 1 Artifacts Box Plot

 

 

 

 

 

 

t—————— —————a

L l L J A

O 50 100 150 200
ARWFACT

Graph 14: Posthole Frequency
Distribution 1 Frequency Box plot

D * O O

l J l J

-10 O 10 20
FREQUENC

Graph 15: Posthole Frequency
Distribution 2 Artifacts Box Plot

 

NL——l

 

 

 

 

 

L l J L l 1 l l I

C) 10 2O 30 4O 50 60 7O 80
ARWFACT

Graph 16: Posthole Frequency
Distribution 2 Frequency Box Plot

—10 O 10 20 30 40
FREQUENC

163

Graph 17: 5Ft.Units Frequency
Distribution 1 Artifacts Box Plot

 

 

 

 

 

 

l—-—— ————l
l l 4 1 1 J
0 500 1000 1500 2000 2500
ARWFCTS

Graph 18: 5Ft. Units Frequency
Distribution 1 Frequency Box Plot

l 1 l 1 J

-10 0 10 20 30
FREQUENC

Graph 19: 5Ft. Units Frequency
Distribution 2 Artifacts Box Plot

 

 

 

 

 

l l l i l l

0 500 1000 1500 2000 2500
ARWFACT

Graph 20: 5Ft. Units Frequency
Distribution 2 Frequency Box Plot

1 l l l l

-10 0 10 20 30
FREQUENC

164

Graph 21: 5Ft. Units Frequency
Distribution 3 Artifacts Box Plot

 

 

 

 

 

 

}_______r

l 1 l L 1 J

0 100 200 300 400 500
ARNFACT

Graph 22: 5Ft. Units Frequency
Distribution 3 Frequency Box Plot

-5 0 5 10 15 20
FREQUENC

Graph 23: 5Ft. Units Frequency
Distribution 4 Artifacts Box Plot

 

 

 

>~

 

l J

1' 1
0 100 200 300 400 500

Graph 24: 5 ft. Units Frequency
Distribution 4 Frequency Box Plot

165

Graph 25: Transformed Surface
Collection Frequency Distribution

 

Box Plot
[F o
l J J A
-0.001 0.000 0001 0.002
FREQU1

Graph 26: Transformed Posthole
Frequency Distribution Box Plot

1 J J l J J

-01 00 01 02 (X3 04
FROU2

Graph 27: Transformed 5Ft. Units
Frequency Distribution Box Plot

[1. o
J l l

-1 0 1 2

h—

FREQUS

BIBLIOGRAPHY

BIBLIOGRAPHY

Anonymous A. HiStQDI of Shiawasee and Clinton Counties, Michigan with
1880 Illustrations and Biographical Sketches of mar Prmeent Men and
Their Pioneers. Philadelphia: D. W. Ensign and Co.

Anonymoous B. History of Clinton County. Michigan. St. Johns: County
1980 Clinton County Historical Society.

 

Anonymous C Annual Report of the School Supervisors. State Archives of
1861-1930 Michigan.

Anonymous D (attributed to \Mlliam Locke Smith) Historiwl Sketches of

1880 Education in Michigan. Lansingzw. S. George and Co.

Beaudry, Mary C. “The Lowell Boott Mills Complexand its Housing: Material

1989 Material Expressions of Corporate Ideology” Historical Archaeology
23(1) pp. 1 8-32.

Binford, Lewis R. “A Consideration of Archaeological Research Design.”
1964 American Antiguity 29(4) pp.425-441.

Card, C. G. School Directog : Mecosta County Michigan 1918-1919. Big Rapids:
1919 Press of the Big Rapids Pioneer.

Carillo, Richard F. “Archaeological Variability-Sociocultural Variability,” in
1977 Research Strategies in Historical Archaeology. Ed.Stanley South.
New York :Academic Press. pp.73-89.

Chenhall, Robert G. “ A Rationale for Archaeological Sampling” Sampling in
1979 in Archaeology. Ed. James W. Mueller. Tucson: University of
Arizona Press. pp. 3-25.

Culter, Horace M. and Julia M. Stone The Rural School: lts Methods and

1913 Management. New York: Silver, Burdett and Co.
Daboll, Sherman B. and D. W. Kelley Past and Present of Clinton County,
1906 Michigan. Chicago: T.S. Clarke.

166

167

Foght, Harold, Weinstein The American Rural School: Its Characteristics,
1910 Its future and Its Problems. New York: Macmillan Co.

Gregory, John M. Twenty -Seventh Annual Report of t_he Sppprintendent of
1863 Publiclnstruction of the Stfi of Michigan, With Accompanying
Documents for the Year 1863. Lansing: John A. Kerr and Co.

 

Hestor, ThomasR., Robert F. Heizer, and John A. Graham. Field Methods in
1975 in Archaeology 6th. ed. Palo Alto: Mayfield. Publishing.

Hosford, Oramel Thirty- Fourth Annual Report of the Superintendent of
1870 Public Instrmon of the Stgte of Michigan, With Accompanying

Documents for the Year 1870. Lansing: W. S. George and Co.

 

Johnson, Alfred E. and Ann S. Johnson “K-Means and Temporal Variability
1975 Variability in Kansas City Hopewell Ceramics,” American Antiguity
40(3) pp.283-294.

Johnson, Matthew H. “Meanings of Polite Architeture in England,” Historical
1992 Archaeology 26(3) pp.45-56.

Judge, W. James, James I. Ebert, and Robert K Hitchcock “Sampling in

1979 Sixteenth-Century Archaeological Survey,” in Regional Sampling in
Archaeology. Ed. James W. Mueller. Tucson: University of Arizona
Press. pp. 82-123.

Kiddle, Henry, Thomas F. Harrison, and NA. Calkins. How to Teach a Graded
1877 Course of Manual Instruction and Manual of Methods for the Use of
Teachers. New York: Van Antwerp, Bragg and Co.

Kintigh, Keith W. “The Effectiveness of Subsurface Testing: A Simulation

1988 Approach,” American Antiguig 53(4) pp. 686-707.

Kohler, Timothy A. and Sandra C. Parker “Predictive Models for Archaeological

1986 Resource Location” in Advances in Archeaological Method and
Theogy, vol. 9. Ed. Michael B. Schiffer. New York: Academic Press.
pp. 397-452.

Lewis, Kenneth E. “Sampling and the Archaeological Frontier". Regional Models

1977 and Component Analysis,” in Research Strategies in Historical
Archaeology. ed. Stanley South. New York :Academic Press.
pp. 1 51 -201 .

Lightfoot, Kent G. “A Defence of Shovel-Test Sampling: A Reply to Shott,”
1989 American Antiguity 54(2) pp. 413- 416.

168

Mueller, A.D. Prpgressivej'ends in Rural Education: An lntegpretive
1926 Discussion of Some of the Best Tendancies in Rural Education.
New York: Century Co.

Mueller, James W. “The Use of Sampling in Archaeological Survey,” Memoirs

1974 of the Socieg for American Archaeology 28.
Nance, Jack D. and Bruce F. Ball “A Shot in The Dark: Shott’s Comments on
1989 Nance And Ball,” American Antiguig 54(2) pp. 405- 412.

Otto, John Solomon “Artifacts and Status Differences-A Comparison of Ceramic
1977 from Planter, Overseeer, and Slave Sites on an Antebellum

Plantation,” in Research Strategies in Historical Archaeology. Ed.
Stanley South. New York: Academic Press. pp.91-118.

Otwell, George N. “Rural SchoolhouseszBuilding Plans, Requirements,
1923 Suggestions,” Department of Public Instruction Bulletin 54.

Plog, Stephen, Fred Plog, and Walter Wait “Decision Making in Modern

1978 Surveys,” in Advances in Archeaological Method and Theog,
Selections for Students from vols. 1-4. Ed. Michael B. Schiffer. New
York: Academic Press. pp. 607-645.

1976 “Relative Efficiencies of Sampling Techniques for Archaeological

Surveys," in The Early Mesoamerican Village. Ed. Kent V.
Flannery. New York: Academic Press. pp. 1 36-158.

Putnam, Daniel The Development of Primapy and Secondagy Public Education
1904 in Michigan: a Historical Sketch. Ann Arbor: George Wahr.

Redman, Charles L. “Multistage Fieldwork and Analytical Techniques,”
1973 American Antiguity 38(1) pp.61-79.

Schiffer, Michael B. Behavioral Archaeology. New York: Academic Press.
1976

Shott, Michael J. “Shovel-Test Sampling as a Site Discovery Technique: A Case

1985 Study from Michigan," Journal of Field Archaeology 12 pp. 457-467.

1990 “Shovel-Test Sampling in Archaeological Survey: Comments on
Nance And Ball, And Lightfoot," American Antiguity 54(2). pp.
396-404.

Stein, Julie K. “Coring Archaeological Sites” American Antiguity 51(3). Pp.
1986 505-527.

169

Schuldenreien, Joseph “Coring and the Identity of Cultural Resource
1991 Environments:A Comment on Stein,” American Antiguity 56(1) pp.
1 31 -37.

Speth, John D. and Gregory A. Johnson “Problems in the Use of Correlation
1977 for the Investigation of Tool Kits and Activity Areas” in Cultural

Change and Continuity: Essays in Honor of James Bennett Griffin.
New York: Academic Press. pp.35-57.

South, Stanley and Randolph Widmer “A Subsurface Sampling Strategy for

1978 Archaeological Reconnaissance” in Rese_arch Strategies in
Historical Archaeolpgy. Ed. Stanley South. New York: Academic
Press. pp.1 19-150.

Thomas, David Hurst. Archaeology: Down to Earth. Fort Worth: Harcourt Brace
1 992 Jovanovich.

Whallon, Robert. “Unconstrained Clustering for the Analysis of Spatial

1984 Distribution in Archaeology,” in lntrasite Spatial Analysis in
Archaeology. Ed. Harold Hietala. Cambridge: University
Press. pp. 242-277.

 

HICHIGQN STRT

E UNIV. LIBRRRIES
HIJJJIJJJII IIIIJIJJIIJIJIJIJJIJIJJIIJIIJHJI
7102819

 

3129301