4. Geophysical Investigations
of the Hopewell Earthworks (33RO27), Ross County,
Ohio; By Arlo McKee
Abstract
The Hopewell site in southern Ohio is one
of the most important and famous prehistoric
sites in North America. Built about AD
100, it consists of more than 3 miles of
earthworks and over 40 mounds. The earthworks
of the Ohio Hopewell suggest a substantial
investment of human labor and cultural
organization that is still poorly understood
after more than 150 years of study. Geophysical
survey techniques can provide the means
to study these earthworks in a rapid and
nondestructive manor. For the past 200
years, these earthworks have been subject
to the degradation of agriculture. A study
comparing multiple geophysical survey techniques
on the Mound 23 area of the Hopewell site
was conducted in 2004. The results of magnetic,
resistance and conductivity surveys were
compared.
Introduction
Throughout the Ohio River valley, during
the period between 200 BC and AD 400,
inhabitants of this region built hundreds
of earthen mounds and dozens of earthen
enclosures.
|
| |
|
click on image to enlarge |
| Most of the mounds appear
to be associated with mortuary activities,
and often contain elaborate and artistic
objects made from a wide range of exotic
raw materials. The earthen enclosures
are typically geometric in form, and
range in size
from a few acres to more than one hundred
acres. Archeologists have named the archeological
record associated with these mounds and
earthworks the Hopewell Culture, after
a site located about 6 miles northwest
of Chillicothe, Ohio. |
|
Figure
1. 1846 map of the Hopewell Earthworks
(Squier and Davis, 1848).
|
Site Background
The Hopewell mound group was first documented
in 1820 by Caleb Atwater. At that time,
the land was owned by W. C. Clark,
so the site was named Clark’s
Fort. Since that time there have been
only three major investigations into
the archeological remains of the site.
These major investigations focused
primarily on the burial mounds associated
with the earthworks. The site was first
excavated by Ephraim G. Squire and
Edwin H. Davis in 1845. During their
investigation, they estimated that
the embankments cover over three miles
in length and the earthworks contain
three million cubic feet of placed
earth (Squire and Davis, 1847). A structure
of this size suggests a substantial
investment of human labor and social
organization. This is truly remarkable
considering the land surrounding the
Scioto river valley is littered with
dozens of these earthworks. |
click
on image to enlarge |
|
|
 Figure
2. 1891 plan-view map of mound 23
(Moorehead, 1922).
|
|
The next excavation was led by
Warren K. Moorehead in 1891. At that
time, Moorehead renamed the site
after the landowner, M. C. Hopewell.
His work, too, focused almost exclusively
on the burial mounds. Among the mounds
he excavated, one of the largest
included mound 23, which was surveyed
during this project. |
More than 50 burials
were found within the mound and hundreds
of exotic burial goods were recovered.
At the base of the mound, Moorehead
(1922) noted a floor of varied sized
gravels.
It is important to note the method
Moorehead’s team employed
for excavation. Moorhead’s
team used horse drawn scrapers
to reduce the mound to about four
feet above the modern surface.
The team then hand excavated the
remainder. The team excavated a
trench to the base floor of the
cultural remains and placed the
backfill behind them. They then
stripped mostly vertical layers
from the wall of the trench, thus
moving their excavation perpendicular
to the minor axis of the mound.
This excavation strategy that means
that, excepting artifacts, nearly
all of the original materials of
the interior portion of the mound
are still more or less in situ.
The soil may be moved, but it was
moved uniformly and only by a distance
of ten feet or so.
In 1922 the last major excavation
of the Hopewell site was conducted
by Henry C. Shetrone. The excavation
was conducted to reexamine all
of the previously excavated mounds
(Greber and Ruhl 1989, Shetrone
2004). Shetrone found that the
western third of the mound had
not been adequately examined. Again,
Shetrone found the gravel base
floor and noted that it contained
a great degree of burned soils.
Geophysical Survey Methods in
Archeology
Geophysical surveys in archeology
can be a very useful tool to
guide excavations and to examine
variation in subsurface features
and soils over large areas. The
use of geophysical survey techniques
can supply an excavation with
a “road map” to subsurface
features without disturbing the
site. There are many different
types of surveys, each with its
own sensitivities and drawbacks.
As Weymouth (1986) notes, there
are two general types of geophysical
methods: active
and passive.
The vast majority of methods
employed for archeological purposes
fall under the active category.
The active survey methods used
for this research include electromagnetic
conductivity and electric resistance.
In each of these methods, a signal
is sent from the measuring instrument
into the ground. Depending upon
the electrical properties of
the soils, the signal will then
be altered from its original
state and sent back to the instrument
to be measured.
Resistivity. Of all the geophysical
methods currently utilized for
archeological purposes, resistivity
surveys were the first to gain
popularity. This method sends
an electric current into the
soil via two conducting probes
placed a few centimeters in the
ground. Two other probes measure
the voltage that is transferred
through the ground. The ratio
of the voltage to the current
that is applied yields the resistance
according to Ohm’s Law
(R=V/I where R=resistance, V=volts,
and I=current). The principal
factor that determines soil resistance
is its water content and distribution
(Weymouth, 1986; Clark, 1990).
When the ground is completely
saturated, soil resistance will
be at a minimum because water
is a good conductor of electricity.
Conversely, when the ground is
completely devoid of moisture,
electrical resistance will be
at a maximum. Optimally, a survey
would be conducted when the ground
is moderately saturated. In this
case, the water distribution
between soil horizons, and buried
cultural remains, would be uneven.
For example, the soil in a buried
ditch or pit tends to be less
dense than the surrounding earth.
This would allow for a greater
accumulation of moisture within
the fill than in the surrounding
area. Hence, the ditch would
appear as a low resistance area
compared to the surrounding soils.
The depth the current will travel
through the soil depends on the
spacing of the probes and the
physical characteristics of the
soil itself. In a uniform matrix,
the voltage will travel in regular
hemispheres between the current
probes. Generally, the depth
of penetration will equal the
horizontal spacing of the current
probes (Clark, 1990; Kvamme,
2001). This survey used the Geoscan
Research RM-15 parallel twin
configuration and a 0.5-meter
probe separation. This allowed
for roughly a 0.5-meter maximum
penetration.
Conductivity. Conductivity is
the theoretical inverse of resistivity.
However, because of the way in
which a conductivity meter senses
information, it can yield differing
results. Conductivity meters
employ non-contact transmitting
and receiving coils. An electromagnetic
signal is sent out by the transmitter
and induces a current in the
soil. This current creates a
secondary magnetic field which
is then sensed and measured by
the receiving coil. Introducing
a new magnetic field in the soil
makes this method sensitive to
metals which will appear as extreme
values. This survey, however,
still retains sensitivity to
the same types of features as
a resistance method. The Geonics
EM-38 was used for the conductivity
portion of this survey. This
instrument operates on a frequency
of 14.6 kHz and houses a 1 m
coil separation, it is capable
of delivering four measurements
per second. The EM-38 also has
two settings for depth. The vertical
dipole mode, which measures a
depth up to 1.5 m, was used in
this survey.
Magnetic. Magnetometry was the only passive geophysical method employed in this
study. This method measures the relative strength of the earth’s magnetic
field. The magnetic properties of a soil depend on the concentration of iron
compounds like hematite, magnetite, and maghaemite (Weymouth 1986). Undisturbed
earth will yield a uniform magnetic field, whereas buried ditches, etc., will
tend to be more or less magnetic than the surrounding matrix. Magnetometers are
generally most sensitive to metals and fired materials. Metals will appear in
magnetic data as paired strong positive and negative values which is referred
to as a dipole. When a material is heated to a certain point the heat will “reset” the
material’s magnetic clock. This process is called thermoremanent magnetism
(Weymouth 1986). The atoms of the substance will align themselves to magnetic
north at the time of cooling. For this reason burned features, such as hearths,
kilns, fired bricks, and burned house floors, are readily visible in magnetic
data. Typically, most anomalies will range between ± 5 nanotesla (nT).
It is not uncommon, however, for a feature to fall within .1 nT of the background
magnetic strength.
The Geometrics G-858 cesium gradiometer was used for the magnetic portion of
the survey. The sensors were mounted on a vertical shaft at a 1-meter probe separation
with the lower sensor measuring the magnetic field of the soils. The upper sensor
measures the background magnetic field. The bottom reading was then subtracted
from the top to cancel out the magnetic variance not related to soil conditions.
Survey area
The survey area was a rectangular block 60 meters north-south and 120 meters
east-west. The western portion of the block was positioned to cover roughly
two-thirds of mound 23. The eastern half of the block covered the east wall
and associated ditch of the large enclosure. Additionally, the eastern portion
of the block was thought to cover the approximate location of the south wall
of the small square enclosure.
|
| |
|
click
on image to enlarge |
| At the time of Moorehead’s
excavation in 1891, mound 23
stood 3-4 meters high and was
roughly 46 meters in its greatest
diameter. Since that time, the
site has been seriously degraded
from agriculture. At the time
of the survey, mound 23 and the
large earth wall stood no more
than one meter above the normal
ground surface. Nothing could
be seen of the earthen ditch
or the wall of the small square. |
|
 Figure
3. 1891 photograph of mound 23
(Moorehead, 1922).
|
| |
|
click
on image to enlarge |
| The survey area was situated
to answer three main questions.
First, what archeological
remains, if any, were left
in mound 23? Second, what
can be detected within the
earthen wall and ditch of
the large enclosure? Finally,
what is the preservation
quality and exact location
of the small square enclosure? |
|
 Figure
4. 2004 photograph of mound
23.
|
| Survey and
interpretation methods
The base lines of eighteen
20 by 20-meter grids were
first laid out with a transit.
Wooden stakes were then placed
in the corners of each survey
grid. One hundred meter tapes
were then stretched between
the east-west running baselines
to place the remaining corner
stakes. During the survey
two 20-meter ropes were placed
along the east-west lines
of each grid to form a boundary.
A third rope was moved along
the ground at one meter intervals
to serve as a guide. The
guide ropes were marked with
colored electrical tape every
half meter and flagging tape
every 5 meters. At the end
of data collection, the instruments
were then downloaded into
a portable laptop computer
for processing. Data manipulation
was carried out using Geoplot,
Surfer 7, and ArcGIS 8. For
final presentation, the data
was transferred into Surfer7.
The grids were set up in
Surfer using the Krigging
method.
The interpretation process
was designed to answer
the survey design goals
stated
previously. Additionally,
the survey was intended
to provide a means to compare
multiple data sets. The
goal
was to produce an easily
interpretable map of the
differing data sets that
would clearly show the
anomalous areas of each
survey. To
this end, the conductivity
and resistance data were
filtered and smoothed using
Geoplot. The datasets were
then exported into Surfer
7 for interpretation. Two
differing contour plots
were created to highlight
the
anomalies of each data
set. The first contour
plot depicted
the typical value range
of the anomalies present.
The
second plot reflected a
range of standard deviations
from
the mean of the data set.
The contour plots were
then exported to ArcGIS
8.3. Each
of the three contours were
assigned a different color
and displayed in partially
transparent fashion. The
resulting image served
as the finished product
of this
attempt at correlation.
Results
Resistance. Resistance
data was collected at one
meter traverses with a
half-meter sampling interval.
A total of 14,400 readings
were collected for the
survey area. The data ranged
in values from 27.2 to
95.0 Ohms. The mean of
the data was 44.4 Ohms
with a standard deviation
of 5.75 Ohms |
| click
on image to enlarge |
|
|
 Figure
5. Resistance Data.
|
|
Mound
23 is displayed clearly
in the resistance data.
The highest values
in the data set mark
the limits of the mound.
This ring is probably
the result of large
gravels that were placed
during the construction
of the mound. The resistance
data also revealed
weaker anomalies in
the interior portion
of the mound. The anomalies
probably result from
gravels that were associated
with the floor of the
mound structures.
The large earth wall
to the east of the
mound is displayed
as a linear anomaly
that runs slightly
northwest to southeast. |
| click
on image to enlarge |
|
 Figure
6. Anomalous areas of
the resistance data.
|
|
This
linear anomaly could
also be the result of
gravels that were placed
during the construction
of the earthwork. Near
the data point E4885
N5060, there is a boundary
that follows the same
direction. This boundary
separates a zone of high
resistance in the east
from the lower resistance
of the rest of the map.
This boundary marks the
edge of the earth wall.
Conductivity. Conductivity
data was collected
at one meter traverses
with a quarter-meter
sampling interval.
A total of 28,800 readings
were collected for
the survey area. The
data ranged in values
from -4.70 to 11.39
milisemens per meter
(mS/m). The mean of
the data was 6.26 mS/m
with a standard deviation
of 1.64 mS/m. |
As
expected, the conductivity
data showed the near
inverse of the resistance
data. The exception
was that the anomalies
were not nearly as
strong. The anomalies
associated with mound
23 appeared as a
light halo compared
to the background
area. Curiously,
these values fell
within one standard
deviation from the
mean. This could
have been due to
the general poor
overall conductivity
of the soil. The same boundary
associated with
the eastern edge
of the large enclosure
is also present
in the conductivity
data. Here, the
boundary is marked
by values of high
conductivity to
the west and low
values to the east.
The values associated
with this boundary
fall within the
same range as the
anomalies of the
mound.
|
|
click
on image to enlarge |
| |
 Figure
7. Conductivity Data.
|
| |
click
on image to enlarge |
| |
 Figure
8. Anomalous areas
of the conductivity
data.
|
|
Magnetic.
Magnetic data was
collected at one
meter traverses.
The instrument was
set to take readings
every 0.2 seconds
which allowed for
roughly 5 samples
per meter to be taken.
A total of 36,918
readings were collected
for the survey area.
The background magnetic
field strength averaged
at 53,3341.58 nT.
The vertical gradient
ranged in values
from -25.5 to 69.4
nanotesla per meter
(nT/m). The mean
of the data was 1.91
nT/m with a standard
deviation of 3.21
nT/m.
|
click
on image
to enlarge |
|
The
magnetometer
clearly displayed
the northern
edge of the mound
as a strong negative
linear anomaly.
The interior
of the mound
is displayed
as strong positive
values in the
south and slightly
lesser positive
values in the
north. These
boundaries could
mark the walls
of burned structures
where the walls
burned longer
or with more
intense heat.
The large
enclosure
is marked
by two
parallel
lines running
through the
length
of the survey
area. The
east line
is of much
greater intensity.
Additionally,
there is
a third
line to the
east
which marks
the eastern
boundary
of the ditch.
|
 Figure
9. Magnetic Data.
|
|
click
on image
to enlarge |
|
 Figure
10. Anomalous
areas of the
magnetic data.
|
|
| Overlay. The comparison
of the three
geophysical
survey types
proved to
be quite
informative.
The plot
of the anomalous
areas of
the surveys
showed the
exact relationship
between the
data sets. |
| |
|
click
on image
to enlarge |
The
anomalies
associated
with mound
23 showed
a close approximation
of the physical
boundary
of the mound.
The conductivity
and resistance
data highlighted
the margins
of the mound
while the
magnetometer
sensed the
interior.
The earth
wall was
displayed
in a similar
manner to
the mound.
The outer
extremities
were displayed
by resistance
and conductivity,
while the
magnetometer
displayed
the interior
of the wall.
Conclusions
As archeological
investigation
techniques
improve,
it can
often
become
quite
useful
to reexamine
previous
studies.
Mound
23 in
this
survey
had been
previously
excavated
twice.
Thus,
it would
be natural
to assume
that
most,
if not
all,
of the
cultural
remains
were
removed.
However,
this
study
proved
that
there
may still
be features
of archeological
significance
left
in the
mound.
Magnetic
data
pointed
to burned
surfaces
associated
with
the floors
of the
pre-mound
funeral
structures.
|
|
 Figure
11. Image
of the overlapped
data sets.
Resistance
is displayed
from blue
to white,
conductivity
is displayed
from white
to blue,
magnetic
data is displayed
from red
to white.
|
| |
click
on image
to enlarge |
| |
 Figure
12. Image
of the anomalous
areas of
the data
sets overlapped.
Resistance
is displayed
as blue,
conductivity
is displayed
as green,
magnetic
data is displayed
as red.
|
|
|
|
|
|
|
|
|
The
results from the magnetic data also suggested
that the large enclosure is still present below
the surface. Resistance and conductivity showed
an area of higher conductivity/lower resistance
to the east of the large wall. Magnetic data
pointed to internal anomalies possibly resulting
from ritual burning or a stone lining. The
data clearly showed that although the earthworks
may not be visible on the ground surface, much
of the core of the wall still remains below
surface.
Acknowledgments
The author would like to thank Steven De Vore,
Rinita Dalan, Bruce Bevan, and all of the instructors
of the National Parks Service 2004 Prospection
Workshop for their guidance and encouragement.
Also thanks to N’omi Greber, Jennifer Pederson
and the staff of Hopewell Culture National Historic
Park for their cooperation and help with this
project. A special thanks should go to William
Volf, MWAC and the UNL field school for assistance
with data collection and imaging. Finally, thanks
to Mark Lynott, and the UNL McNair program, for
without their guidance and support this work
would not have been possible.
References:
Bevan, B. W.
1998. Geophysical Exploration for Archaeology:
An Introduction to Geophysical Exploration. Midwest
Archeological Center. Copies available from Special
Report No.1.
Clark, A.
1990. Seeing Beneath the Soil: Prospection
Methods in Archaeology. 2nd ed. B.T. Batsford, London.
Conyers, L. B., and Goodman, D.
1997.Ground-Penetrating Radar: An Introduction
for Archaeologists. Alta Mira Press, Walnut
Creek, CA.
Greber, N’omi B., and Katharine C. Ruhl
2000. The Hopewell Site: A Contemporary Analysis
Based on the Work of Charles C.
Willoughby. Eastern National, Ft. Washington,
Pennsylvania. Reprint. Originally
published in 1989 by Westview Press, Boulder,
Colorado.
Kvamme, K. L.
2001. Current Practices in Archaeogeophysics.
In Earth Sciences and Archaeology, edited by
V. T. H. Paul Goldberg, and C. Reid Ferring,
pp. 353-384. Kluwer Academic/Plenum
Publishers, New York.
Moorehead, W. K.
1922. Hopewell Mound Group of Ohio. Field Museum
of Natural History - Anthropological Series
6(5).
Shetrone, H. C.
1927. Explorations of the Hopewell Group
of Prehistoric Earthworks. Ohio Archaeological and Historical
Publications. Volume 35.
2004. The Mound-Builders. The University of Alabama
Press, Tuscaloosa, Alabama. Reprint. Originally
published in 1930 by D. Appleton and Company.
Squier, E. G., and Edwin H. Davis
1998. Ancient Monuments of the Mississippi
Valley. Smithsonian Institute Press, Washington,
D.C. Reprint. Originally published in 1848.
Weymouth, J. W.
1986. Geophysical Methods of Archaeological Site
Surveying. Advances in Archaeological
Method and Theory 9:311-395.
Print this Page
 word
.doc (670K)
 pdf (644)
|
