A new multistage construction chronology for the Great Serpent
Edward W. Herrmann
, G. William Monaghan
, William F. Romain
Timothy M. Schilling
, Jarrod Burks
, Karen L. Leone
, Matthew P. Purtill
Alan C. Tonetti
Department of Geological Sciences, Indiana University, 1001 E. Tenth St., Bloomington, IN 47405, USA
Indiana Geological Survey, Indiana University, Bloomington, IN, USA
Newark Earthworks Center, Ohio State University, Newark, OH, USA
Midwest Archeological Center, National Park Service, Lincoln, NE, USA
Ohio Valley Archaeology, Inc., Columbus, OH, USA
Gray &Pape, Inc., Cincinnati, OH, USA
Department of Geology and Geography, West Virginia University, Morgantown, WV, USA
ASC Group, Inc., Columbus, OH, USA
Received 10 February 2014
Received in revised form
30 May 2014
Accepted 4 July 2014
Available online 12 July 2014
Efﬁgy mounds occur across the midcontinent of North America but their cultural purposes and con-
struction chronologies are rarely known and often controversial. Determining the age and construction
history of monuments is important to relate religious symbolism, scientiﬁc knowledge, and cultural
continuity to groups within a region. Based mainly on circumstantial evidence, researchers have long
held that Serpent Mound in Ohio, USA, was constructed 2000e3000 years ago during the Early
Woodland (Adena) or Middle Woodland (Hopewell) periods. Excavations in 1991 recovered charcoal
buried at shallow depths (35e45 cm) in ﬁll units of the mound and the
C ages from two of these units
indicated that Serpent Mound was built ~900 years ago, during the Late Prehistoric (Fort Ancient) period,
much later than originally thought. Our recent multidisciplinary work provides a more complex, robust
construction history of Serpent Mound. We used geophysics to map the mound, and solid-earth cores to
provide accurate stratigraphy and organic samples for
C age estimates from the base of the mound.
Bayesian statistical analyses of the seven
C ages from Serpent Mound suggest that it was ﬁrst con-
structed ~2300 years ago during the Early Woodland (Adena) period but was renovated 1400 years later
during the Late Prehistoric (Fort Ancient) period, probably to repair eroded portions of the mound.
Modiﬁcation of the mound is also indicated by a possible abandoned coil that is located near the head of
the Serpent and visible only in the magnetometer survey.
©2014 Elsevier Ltd. All rights reserved.
The Great Serpent Mound is one of the most iconic efﬁgy
mounds in the North American midcontinent (Fig. 1). Squier and
Davis (1848: 96) described it as “probably the most extraordinary
earthwork thus far discovered in the West.”It has been designated
as an Ohio State Memorial and a National Historic Landmark. In
spite of these facts, relatively little modern archaeological work has
been conducted within Serpent Mound or on the surrounding
promontory. Our research seeks to deﬁne the construction chro-
nology of Serpent Mound.
1.1. Previous work at Serpent Mound
Putnam (1890) conducted the earliest excavations at Serpent
Mound. He placed trenches in the efﬁgy and several nearby earthen
mounds during 1887e1889, and later he led efforts to restore and
preserve Serpent Mound. Putnam (1890) recognized that people of
two different time periods occupied the Serpent Mound area and
attributed the efﬁgy to the earlier of these groups. Although the
terms “Adena”and “Fort Ancient”had not yet been deﬁned in the
late 19th century, subsequent analyses of the artifacts recovered
*Corresponding author. Tel.: þ1 812 856 0587.
E-mail address: firstname.lastname@example.org (E.W. Herrmann).
Contents lists available at ScienceDirect
Journal of Archaeological Science
journal homepage: http://www.elsevier.com/locate/jas
0305-4403/©2014 Elsevier Ltd. All rights reserved.
Journal of Archaeological Science 50 (2014) 117e125
elsewhere within the site area are attributed most to either the
Early Woodland Adena culture (ca. 500 BCeAD 200) or Late Pre-
historic Fort Ancient culture (ca. AD 1000e1650) occupations
(Abrams and Le Rouge, 2008; Clay, 2005; Cook, 2008), which sug-
gests a long occupational history spanning >1500 years (Fletcher
et al., 1996; Grifﬁn, 1943).
Greenman (1934), Grifﬁn (1943),
Webb and Snow (1945: 341), and Webb and Baby (1957: 106) all
believed that Serpent Mound was constructed during the Early
Woodland Period, by people of the Adena culture.
The assessment of Serpent Mound as an Adena construction was
circumstantial and based mainly on the Adena cultural afﬁliation of
a conical mound about 200 m southeast of the Serpent (Fig. 1). First
excavated by Putman (1890), the conical mound contained multiple
burials and associated grave goods, including pottery and projectile
points. These grave goods were later analyzed and assessed as
Adena by Grifﬁn (1943:56e64), who also found both Adena and
Fort Ancient materials in nearby cultural features. Even though
Adena and Fort Ancient occupations both occurred near Serpent
Mound, because Adena was the earliest well-documented occu-
pation, its burials and artifacts were given precedence as temporal
Fig. 1. Maps of the Great Serpent Mound, Ohio. (A) Base map showing topography nearthe top of the Serpent Mound ridge, the footprint of Serpent Mound, and the locations of cores
and trenches placed into the mound. Contour interval 50 cm; core numbers labeled adjacent to core location; cores with
C ages labeled. See Table 1 for details of the
C samples. Note
closed depressions near the ends of loops within middle part of Serpent. (B) Magnetometer map of Serpent Mound; image displayed at ±5 nT; readings <5 nT displayed as 5nT;
readings >þ5 nT displayed as þ5 nT.Magnetometer data collected on a 12.5 50 cm grid; possible abandoned coil near the head of Serpent Mound labeled. Dark is positive. (C) Map of
Serpent Mound site showing locations of adjacent mounds and prehistoric habitation areas, erosional gullies, and steep valley wall. (D) Map showing regional location of Great Serpent
Mound, Ohio, within eastern North America. (Images provided courtesy of G. William Monaghan. Magnetometer map provided courtesy of Jarrod Burks.)
We recognize several taxonomic issues concerning the term “Adena”[Brown,
2005; Clay, 2005; Greber, 2005; Mainfort Jr., 2005]; however, for the purposes of
this discussion and to maintain historic continuity, we will use “Adena”as a heu-
ristic term of convenience.
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125118
indicators for when the mound was built. Grifﬁn (1943:57)
concluded that: “Although artifacts taken by Putnam from the
conical mound south of the serpent and from the lower level of the
near-by village site cannot be positively assigned to the builders of
the efﬁgy, it is considerably less likely that the later Fort Ancient
occupants built the serpent.”This supposition remained the
accepted construction age of Serpent Mound until 1996 when the
C ages from within its ﬁll were reported.
Fletcher et al. (1996) reopened one of Putman's trenches
through the western end of a coil of Serpent Mound (Fig. 1). They
documented the proﬁle and collected charcoal from three contexts:
two from mound ﬁll just above what they considered the mound
base and one from “natural sediment”below the mound base
(Fletcher et al., 1996:119,132e133).
C age estimates from these
contexts provided the ﬁrst direct evidence for when Serpent
Mound was built. The results indicated to Fletcher et al. (1996:133)
that the mound was ﬁrst constructed at 920 ±70 BP (Table 1),
which surprised most because these dates implied that the mound
was built during the Late Prehistoric (Fort Ancient) Period, 1400
years later than originally suspected.
In 2011, we began a multidisciplinary project to reevaluate when
and how Serpent Mound was built. Bayesian analysis of
obtained during our work using an OxCal model suggests that
Serpent Mound was ﬁrst constructed soon after ca. 2300 years ago
during the Early Woodland (Adena), as originally assumed. Despite
the fact that these ages appear to contradict those provided by
Fletcher et al. (1996), we believe that they are compatible and
reveal a more complex construction and use history for Serpent
Mound than previously thought.
2. Material and Methods
The characteristics and lateral continuity of the paleosol buried
during mound construction are essential to resolve the chronology
and construction sequence for Serpent Mound. Consequently, our
methods focused on subsurface data collection at multiple loca-
tions in the mound. We combined magnetometer (gradiometer)
survey, LiDAR data (1 m resolution) (OGRIP, n.d.) and data from a
series of continuous, solid-earth cores, to construct a detailed
stratigraphy and map of the mound and surrounding promontory.
The methods employed were selected to minimize impact to the
site, which was a requirement of our permission to work at the
site, but still provide data capable of addressing our research
The magnetometer survey (Geoscan Research FM 256 ﬂuxgate
gradiometer) provided images of near-surface anomalies, such as
trenches, structural remains, and burned features. Magnetometer
data were collected on a 12.5 50 cm grid, and processed through
Geoplot 3.0; the resulting image is displayed at ±5nT(Fig. 1), dark
is positive (Burks, 2012). Magnetometer imaging revealed an
arcuate-shaped anomaly near the neck of the efﬁgy that is similar
in size, shape and form to the existing coils in Serpent Mound
(Fig. 1A, B). Because the anomaly has no surface expression, we
excavated a 1 5 m trench across the anomaly revealing lithic
debitage, ﬁre-cracked rock and a projectile point fragment that may
date to the Late Archaic period predating mound construction
(Burks, 2012). Although no distinct features were present, the
presence of light-colored silty sediment matched the location of the
When coring, we avoided delicate areas of the mound (e.g., oval
and tail) and previous excavation locales (Fig. 1), focusing instead
on relatively thick portions of the mound. Although trench exca-
vation would likely produce more detailed results, it is a much
more destructive process. Small-diameter (3-cm) solid earth cores
can provide similarly comprehensive and ﬁnely detailed mound
stratigraphy and organic samples, but are far less invasive than
trenching or other excavation techniques (Monaghan and Peebles,
2010; Stein, 1986).
Eighteen continuous solid-earth cores (Fig. 1) were collected
using a Geoprobe (Model 54 TR) and dual tube sampling system,
which drives continuous casing along with the sample tube to
avoid core-hole collapse or sample contamination. These cores not
only allowed us to directly determine the mound base and the
thickness and character of the ﬁll deposits at several places, but
they also provided a spatially broader set of samples from which to
determine the initial age of mound construction. Core depths var-
ied from 120 to 280 cm and provided stratigraphically intact
sediment samples from which the mound building sequence was
interpreted. Core holes were grouted with bentonite upon
completion. The cores were described for stratigraphic, pedogenic,
weathering, and physical characteristics (color, texture, bedding,
soil horizons, contacts, cultural or other inclusions, etc.). We
particularly focused on the presence, depth and character of pale-
osols buried within or under the mound. The depositional prop-
erties and contacts between various mound ﬁll units, and the
pedogenic characteristics of the submound paleosol also provide
important details concerning when and how the mound was con-
structed and provided important contexts for organic matter
C ages from serpent mound.
Lab number Core# Depth (cm) Material
C age (BP)
Median cal age
calibrated modeled age
Context for OxCal model: ZOD/A
Beta 337163 core 6 87e90 organic sediment 2170 ±30 22.2 334 BC 382e182 BC(360e116 BC)
Beta 337168 core 7 105e110 organic sediment 2320 ±30 23.9 390 BC 414e257 BC(429e235 BC)
Beta 337169 core 11 95e98 organic sediment 2310 ±40 23.2 382 BC 421e231 BC(482e209 BC)
Beta 337170 core 13 100e105 organic sediment 2300 ±50 22.6 375 BC 419e228 BC(488e204 BC)
Context for Oxcal model: mound ﬁll (postconstruction)
Beta 337162 core 3 75e80 organic sediment 2510 ±30 24.9 639 BC 790e540 BC(791e540 BC)
Beta 337166 core 7 75e80 organic sediment 2530 ±80 23.8 639 BC 808e416 BC(808e416 BC)
Beta 337167 core 7 85e90 organic sediment 2180 ±30 21.7 303 BC 362e173 BC(361e168 BC)
Beta 55277 n.a. 35 oak/ash charcoal 920 ±70 eAD 1106 AD 998e1225(AD 997e1253)
Beta 55278 n.a. 45 oak/ash charcoal 920 ±70 eAD 1106 AD 998e1225(AD 997e1253)
Context for Oxcal model: Submound (relationship uncertain)
Beta 47212 n.a. 132 oak charcoal 2920 ±65 e1068 BC 1261e907 BC(1368e925 BC)
As reported by Fletcher et al., 1996.
Median calibrated age (calendar years) based on 2
Calibrated using OxCal 4.2.3; ages in parentheses are OxCal modeled ages (Bronk Ramsey, 2013) using IntCal13 atmospheric curve (Reimer et al., 2013); calibrated ages
show entire 2
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125 119
collected for AMS dating from mound ﬁll and the submound A
horizon of the paleosol that was buried during initial mound con-
struction. We sampled all charcoal from the top of the buried
paleosol because organic matter from the surface of buried soil
horizons tend to be younger than those from deeper paleosol ho-
rizons (Matthews,1985; Martin and Johnson, 1995:236; Wang et al.,
1996:287). Collecting charcoal from such contexts is important
because the youngest dates from the uppermost portion of a
paleosol (i.e., A
horizon) marks the age of the most recent charcoal
deposited within the paleosol, which also represents the maximum
age for when the surface was buried and cut off from additional
organic matter inputs (Haas et al., 1986:473; Holliday, 1995:120;
Holliday, 2001:90; Holliday, 2004:179).
The selected samples were submitted to BETA Analytic who
treated them with an acid bath and dated the residue through AMS
methods. Thus, even though our samples included much charcoal
in the matrix, they were run as bulk sediment-type dates. The
timing of construction of Serpent Mound was estimated using
Bayesian statistics through OxCal 4.1 (Bronk Ramsey, 1994, 1995,
2001, 2008, 2009, 2013; Bronk Ramsey et al., 2010). The model
was based on an analysis of the ages and contexts of the samples
derived from the submound A
horizon (i.e., Beta 337163, Beta
337168, Beta 337169, and Beta 337170; Fig. 2;Table 1).
3.1. Serpent Mound Stratigraphy
Our interpretation of the Serpent Mound chronology is based
primarily on the stratigraphic relationships between mound ﬁll and
the premound-construction ground surface, particularly by the
presence of an intact soil proﬁle upon which mound ﬁll rests.
Mound ﬁll, which was normally ~1 m thick, was readily distin-
guishable from submound deposits in all cores (Fig. 3). Mound ﬁll
sediments are typically very ﬁne sandy silt loams or silty clay loams
with loose, subangular blocky structure. Small pebbles (<0.5 cm),
charcoal ﬂecking, ﬁre-cracked rock and rootlets are typical in-
clusions, and abrupt, sharp contacts between different mound ﬁll
units are common. The surﬁcial portion of the mound is typically
disturbed, probably by historic plowing. However, the uppermost
mound ﬁll units in cores 5, 10, 11, 12 and 14, were compositionally
different (sandier) with weaker pedogenic development compared
to other cores and may indicate recent, probably historic, repairs to
the mound (Fig. 3, core 11).
The deposits directly underlying Serpent Mound were generally
silty clay loam that had weathered into a well-developed, 100e150-
soil horizon sequence. Typically, the A
horizon of this paleosol is friable, dark brown (10 YR 3/3-3/2) silty
clay loam and is usually underlain by an E
horizon composed of
powdery, very friable silt loam (10 YR 7/2-7/3). The E
underlain by a dense silty clay B
horizon (7.5 YR 4/6) that is in turn
underlain by weathered bedrock (C
horizon) or dense, oxidized
horizon). Overall, the characteristics of this paleosol, such
as well-developed E and B
horizons, suggest that it underwent a
relatively long period of pedogenesis, and represents the ground
surface buried during mound construction. The characteristics of
the buried A
horizon (i.e., color, organic-matter content, structure,
etc.) and its position within the underlying sequence of soil hori-
zons clearly distinguish it from overlying mound ﬁll deposits.
Although it represents the ground surface upon which the Ser-
pent Mound was constructed, the surface of the A
horizon of the
submound paleosol is occasionally disturbed, probably during the
construction process, and sometimes mixed with the underlying E
horizon in cores 3, 5, 7, 10, 11, and 12. These zones of disturbance
(ZOD) occur only in the upper several cm of the paleosol (e.g., A
horizons) and included some charcoal ﬂecking and occa-
sional fragments of ﬁre-cracked quartzite (i.e., cores 7 and 12). No
charcoal or organic matter was observed beneath the buried A
horizon from our cores, but was noted by Fletcher et al. (1996:119).
We believe that the ZOD was created as a result of mound-building
processes during the initial construction of Serpent Mound.
3.2. Radiocarbon dating and OxCal chronological model
To estimate when Serpent Mound was ﬁrst constructed, char-
coal or charcoal-rich sediments were sampled from near the base of
the mound. These samples were derived from the top of the sub-
mound paleosol (ZOD/A
horizon) and from ﬁll just above the base
of the mound and provided seven
C ages from ﬁve different cores
(e.g., Cores 3, 6, 7,11, 13; Figs. 1 and 3;Table 1). The charcoal pieces
sampled were too small or degraded for taxonomic identiﬁcation
beyond noting them as carbonized plant materials and most were
too small to remove from the matrix or to submit as standalone
dates. Notably, the locations for the ﬁve cores from which
were obtained are from widely separated areas of the mound. Four
of these ages derive from the ZOD (Cores 7 and 11; Figs. 1 and 3;
horizon (Cores 6, and 13; Figs. 1 and 3;Table 1) that
marks the base of the mound and the three others were from
emplaced mound ﬁll directly above the base of the mound (Cores 3
and 7; Figs. 1 and 3;Table 1).
C ages were calibrated used OXCAL 4.1 and their calibrated
probability ages range from 639 to 303 BC (combined 95% proba-
bility range of 808e116 BC; Table 1;Fig. 2). While this is up to a
~400 year spread, the range is small compared to ~1400 years
Fig. 2. Diagram showing construction chronology for Great Serpent Mound. Results and interpretation of OxCal model are shown as well as the distribution and probability density
diagram (PDD) of individual calibrated
C ages of all samples reported from Serpent Mound; ages shown are OxCal modeled ages (Bronk Ramsey, 2013) using IntCal13 atmospheric
curve (Reimer et al., 2013). See Table 1 for details of the
C sample and unmodeled calibration results. (Images provided courtesy of G. William Monaghan.)
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125120
separating our youngest age and that suggested by Fletcher et al.
(1996) for the construction of Serpent Mound (i.e., 303 BC and AD
1106, respectively; Fig. 2;Table 1). The range of our dates is even
tighter if the ages derived from the emplaced mound ﬁlls, espe-
cially Beta 337162 and Beta 337166, which occur 15e25 cm above
the mound base and are 200e300 years older than those from ZOD/
A(Table 1), are discounted. This age discrepancy may have resulted
because charcoal deposited in submound A
horizons or pre-
construction cultural features were incorporated into ﬁll when the
mound was ﬁrst built. The older dates likely represent charcoal
from deeper, older horizons that were borrowed from their original
context in order to be placed on the mound, which has also been
noted at other mound locales (Saunders et al., 2005; Monaghan and
Peebles, 2010; Black, 1967).
The initial construction chronology of Serpent Mound was
modeled through OxCal as a single-phase event that occurred very
rapidly (i.e., within several years). The model was constructed to
yield the maximum possible age for burial of the submound pale-
osol. The results indicate that the paleosol was buried sometime
after 321 BC (95% probability range of 381e44 BC; Fig. 2), which
also represents the earliest possible time of mound construction.
4.1. Initial construction
Although the OxCal model provides a maximum construction
age, we believe that taken as a whole our data strongly support that
Serpent Mound was ﬁrst constructed ~2300 years ago, rather than
~1400 years later. Our results indicate the presence of a pre-
construction paleosol beneath Serpent Mound, and that charcoal
from different locales along its surface dates consistently to ~300
BC. The youngest calibrated age within the 95% probability range is
116 BC and we obtained no dates associated with the later Fort
Ancient occupation of the site. However, the basic problem is that
the age of the charcoal in paleosol A
reﬂects timing of the event
through which it was deposited and is not necessarily associated
with the burial of the paleosol. Consequently the mound could have
been constructed any time after 300 BC. The precision of the model
and exactly when the paleosol was buried is complicated by sample
genesis and contexts. Unlike the construction ages reported from
other mounds using similar techniques (e.g., Monaghan and
Peebles, 2010; Monaghan et al., 2013), the estimate for when Ser-
pent Mound was built, including that of Fletcher et al. (1996),is
based on detrital charcoal which can impart additional errors.
The disturbance of the premound A
horizon that formed the
ZOD likely occurred when Serpent Mound was built; however, the
charcoal fragments within it were not necessarily deposited when
the ZOD was formed. They were likely deposited in the soil during
earlier, preconstruction natural and cultural events and their resi-
dence time in the soil (i.e., the time lag between when charcoal was
ﬁrst deposited in the paleosol and subsequently buried and pre-
served with the paleosol during the initial mound construction) is
unknown. The uncertain soil residence time of the charcoal is
further complicated because it could be derived from heart- or
sapwood. If from heartwood, the tree could be a few hundred years
older than the event during which it was burned and deposited.
Given the ~1400 year difference between an Early Woodland
(Adena) and Late Prehistoric (Fort Ancient) construction, the
discrepancy of a few hundred years related to relic wood is not a
serious impediment to resolving when Serpent Mound was ﬁrst
Our approach to resolving when Serpent Mound was ﬁrst built is
similar to that commonly used to date paleosols, which assumes
that the time of paleosol burial is approximated by the youngest
ages found (Haas et al., 1986:473; Holliday, 1995:120; Holliday,
2001:90; Holliday, 2004:179; Martin and Johnson, 1995:236).
Saunders et al. (2005:636e637) followed a similar process to
Fig. 3. Diagram showing sediment proﬁle logs of dated cores; core numbers labeled above logs; locations of cores shown in Fig. 1.
C sample depths in core marked and calibrated
ages labeled; see Table 1 for details of the
C samples and calibration results. Mound ﬁll sediments are indicated by light gray; bottom of mound indicated by thick black line. The
submound paleosol was clear in all cores and is located beneath the thick black line; soil horizons labeled in the co re logs. Mound ﬁll units are separated by clear, abrupt contacts
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125 121
determine when mounds were ﬁrst constructed at Watson Brake.
They used bulk sediment (humate) and charcoal from submound
paleosols to establish when the soil was buried (i.e., the earliest age
for mound construction). This initial construction age was then
further constrained by the
C ages of overlying mound-use layers.
For example, the submound paleosol from Mound A dated to 3515
BC while the ﬁrst overlying use surface dated to 3345 BC. A similar
few hundred year difference generally occurs between submound
paleosol ages and those of mound surfaces, which range
3500e3950 BC and 3263e3710 BC, respectively (Saunders et al.,
2005:640e642), and supports the contention that the youngest
ages of a paleosol closely corresponds with its burial, as suggested
by several other studies (Holliday, 2004:182; Holliday, 1995:120;
Martin and Johnson, 1995:236; Haas et al., 1986:473).
Studies of charcoal from modern soils also support the
contention that the youngest ages from the paleosol are generally
the best approximate of the time of burial (Payette et al., 2012;
Sanborn et al., 2006). Payette et al. (2012) and Sandborn et al.
(2006) collected charcoal from the A-horizon and sub-A-horizon
of forest soils in Quebec and British Columbia (respectively) to
document the timing of ﬁres. They showed that the youngest
charcoal occurred in the A-horizon, the majority of which is <400
years old, and that older charcoal was more common at depth in the
soil proﬁle. Studies focused on black carbon (including charcoal)
turnover rates conﬁrm that surﬁcial soil horizons up to 20 cm deep
are dominated by younger, fast-cycling charcoal with turnover
rates of about 300 years (Hammes et al., 2008;Fontaine et al.,
2007). Such a distribution of charcoal in the soil proﬁle (young in
surface soil horizons and older charcoal preserved at depth) has
been noted and discussed by others (e.g., Matthews, 1985; Martin
and Johnson, 1995:236; Wang et al., 1996:287). The concentration
of younger charcoal in A-horizons results because high rates of
biological activity, turbation and accelerated leaching can promote
rapid oxidation and accelerate organic debris destruction. Although
charcoal that is thousands of years old can be preserved in the
upper portion of an A horizon, it is not common. For example,
Payette et al. (2012:9,10) found that although charcoal in the
modern A-horizon ranged from modern to 1750 BP, the mode of all
ages (145 BP) was skewed very young and more in line with a
contemporary age. If this soil were buried today, most of the
charcoal would date to within a few hundred years of present.
Similarly, if the submound paleosol at Serpent Mound were buried
during a Fort Ancient construction, the ages of its A horizon should
be concentrated during a few-hundred-year interval before AD
1100. The distribution of ages (Fig. 2) strongly suggests that the
submound paleosol was more likely buried during an Adena
Such a conclusion is warranted because all of the (95% proba-
bility) calibrated age ranges reported from any context within
Serpent Mound, except for those recorded by Fletcher et al. (1996),
were older than 116 BC (Fig. 2;Table 1). If Serpent Mound was ﬁrst
constructed after AD 1100, as suggested by Fletcher et al. (1996),
then the presence of 900 year old organic matter would be ex-
pected on the surface of the paleosol. Because of the lack of any
younger charcoal, a more acceptable explanation for the age of
submound charcoal is that it approximates when the paleosol was
buried and the initial mound construction ~2300 years ago (Fig. 2).
4.2. Multi-stage construction
An initial Fort Ancient construction of Serpent Mound proposed
by Fletcher et al. (1996) is contradictory with our OxCal construc-
tion model and
C age ranges (Fig. 2;Table 1). Our data suggest
that the mound was ﬁrst built ~1400 years earlier and contempo-
rary with an Adena occupation as presumed throughout most of
the 20th century. Such an ostensibly substantial temporal in-
compatibility could be resolved by suggesting that Fort Ancient-age
charcoal was introduced into the mound historically during Put-
man's excavations or his subsequent mound restorationdan
assumption that is not necessarily warranted. Although historic
repair may be indicated by our core data and the Fletcher et al.
(2006:122) excavation, the historic repair is stratigraphically higher
in the mound ﬁll units than the samples Fletcher et al. obtained for
The evidence compiled by Fletcher et al. (1996) concerning the
reliability of their
C ages is generally convincing and supports the
charcoal as authentically related to a Fort Ancient (re)construction
episode 900 years ago, which leaves the contradiction between the
two initial construction chronologies unresolved. To settle this
contradiction, we propose that Serpent Mound was constructed
and then later modiﬁed during two distinct episodes: an Adena
construction ~2300 years ago during which the mound was ﬁrst
built, followed ~1400 years later by an episode of Fort Ancient
renovation or repair. Lynott (2007:558) found evidence of a similar
repair episode at the Hopeton Earthworks where prehistoric
groups likely renovated the Hopewell age monument some 56 km
northwest of Serpent Mound 800 years after its construction. The
reason for renovating Serpent Mound 900 years ago is unknown,
but might relate to repairs necessary after parts of the mound were
damaged as a result of erosion, sinkhole subsidence, or both.
Damage to some of the mound extensive enough to require repairs
is not surprising given its low pre-restoration height (generally
1 m high), antiquity of initial construction, ridge-top landscape
position, and associated karst topography (Figs. 1, 2 and 4).
Several lines of evidence support multiple episodes of con-
struction for Serpent Mound. The set of dates obtained by Fletcher
et al. (1996) and our set of dates are each internally consistent,
albeit separated by 1400 years, but are also from different contexts.
These factors imply that they may mark two distinct construction
events within different parts of the mound (Fig. 2;Table 1). Except
for the two AD 1106 dates reported by Fletcher et al. (1996)
(Table 1), all
C ages reported from Serpent Mound predate ~300
BC, which further demonstrates that an initial mound construction
during the Fort Ancient period is implausible. We recovered no Fort
Ancient age organic matter from the mound base anywhere along
the extent of the mound (Fig. 1). In fact, all the ages from the base of
the mound, which were spread along most of the length of the
mound, were at least 1400 years older and are remarkably similar
to each other (Figs. 1 and 2). Even the exceptions to these ages (Beta
337162 and Beta 337166; Fig. 2;Table 1) are older, not younger,
than the model median age of 321 BC. The absence of charcoal
younger than ~300 BC is strong evidence that Serpent Mound was
constructed during the Early Woodland (Adena) Period (Fig. 2).
Fort Ancient-age charcoal occurs only in the Fletcher trench
(Figs. 1 and 4), and their contexts are atypical compared to those
from other locales. For example, cores show that a well-developed,
generally continuous paleosol underlies most of Serpent Mound
but is absent in the Fletcher trench (Fletcher et al., 1996:121e122)
(Fig. 1). Coring and LiDAR data also show the mound is typically
~100 cm thick and directly underlain by a ZOD, which incorporated
the premound A
horizon during mound construction (Figs. 1 and
4). Where the Fletcher trench was placed, at the western end of
coil 4, however, is different. Not only is the submound paleosol
absent here, the mound is only ~50 cm thick, and blocks downslope
drainage from the edge of a surface depression (Fig. 4). This
depression and that within coil 5 (Figs. 1 and 4) form the head of an
erosional gully that drains the Serpent Mound promontory.
Considering that karst and sinkholes are common across the
landform, the depressions may be active sinkholes (Figs. 1A and
3B). Erosion was likely where coils of Serpent Mound block the
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125122
natural drainage through gullies (e.g., west end of coil 4 where the
Fletcher trench was placed; Figs. 1A and 3C) and would have been
particularly extensive where the drainage was headed by a sinkhole
(Fig. 4B). When these sinkholes were active, they would have
promoted mound instability and erosion through submound
drainage or collapse.
Because of its landscape position within an upland drainage
gully and adjacent to a possible sinkhole, the area of the Fletcher
trench has probably long been prone to intermittent erosion that
was likely extensive enough to require periodic repairs. Erosion and
redeposition within this part of the mound are supported by the
age of charcoal buried 70 cm below the Fort Ancient mound ﬁll in
the Fletcher trench. The depth and age of this material (2920 BP,
95% probability range 1368e925 BC, BETA 47212; Table 1) led
Fletcher et al. (1996:132e133) to suggest that it was related to an
earlier Late Archaic occupation and was “carried to the lower depth
by bioturbation or some other mode of transport.”We agree but
suggest that the “other mode of transport”was likely long-term
processes associated with gulley formation (Fig. 4). Given the age
of the buried charcoal, gully formation was ongoing during the
Early Woodland. Whether the erosion and redeposition occurred
before or after the initial mound construction is uncertain, but the
lack of submound A
horizons in the Fletcher trench is consis-
tent with erosion extensive enough to completely remove the
original (Adena) mound ﬁll and the upper part of the paleosol.
The remaining coil arms (e.g., coil 4, Fig. 4) that parallel the
erosional gully were apparently not signiﬁcantly altered by erosion.
Other Serpent Mound alterations, however, may have occurred
at the site prehistorically. A possible reconﬁguration of Serpent
Mound, which may mark an erased coil, was identiﬁed as a
sinusoidal-shaped magnetic anomaly at the neck of the Serpent
(Fig. 1A, B) (Burks, 2012). Because of its similar size and geometry to
the other Serpent Mound coils and lack of surface expression, the
anomaly is believed to be an abandoned and erased coil. Lithic
debitage, ﬁre-cracked rock, and projectile point fragments, one of
which may date to the Late Archaic Period, were recovered from a
trench excavated through the eastern arm of the anomaly (Fig. 1)
(Burks, 2012). Even though none of the artifacts were associated
with apparent features, the presence of Late Archaic cultural ma-
terials within the footprint of the erased coil suggests that it may
have been constructed upon a Late Archaic occupation horizon or
that these artifacts were incorporated into the base of the mound
during construction. However, no direct evidence of mound con-
struction was noted within the trench,which is not surprising given
that evidence of mound construction was also lacking below the
horizon in cores or trenches within the mound. Moreover,
the extensive historic disturbances around Serpent Mound, espe-
-century land clearance, plowing, and erosion and post-
Putnam historic renovation, pathways, etc., would have also
signiﬁcantly affected the upper 20e30 cm of the ground surface. If
Fig. 4. Three-dimensional (3D) relief maps and topographic proﬁles for Great Serpent Mound, based on LiDAR data (OGRIP, n.d.). (A) Topographic proﬁle through major body coils of
Serpent Mound; view generally east and northeast. Location of proﬁle shown in B. Proﬁles based on LiDAR contours maps (Fig. 1). Positions of coil arms and Fletcher trench labeled.
(B) 3D relief map of Serpent Mound showing position of Serpent Mound in the uplands and erosional gullies on east and west side of mound; gullies drain into a deeply incised
(~70 m) Brush Creek valley that developed on Silurian carbonates (Reidel, 1975). Small depressions surrounding the mound are probably karst (sinkhole) features. Depressions
inside the ends of coils 4 and 5 may represent active sinkholes (see Fig. 1). Location of trace for proﬁle A shown; coils, coil arms, and trench locations discussed in text are labeled. (C)
Three topographic proﬁles showing Fletcher/Putnam trenches at the western end of coil 4 and the adjacent SE and NW arms of coil 4 of Serpent Mound. Thickness of mound ﬁll
shown; base and thickness of Fletcher trench after Fletcher et al. (1996); base and thickness of adjacent coils based on Cores 8, 9, and 10 and the LiDAR morphology of Serpent
Mound. (D) 3D relief map of Serpent Mound showing details of coils 4 and 5; Location of image shown by shaded area in panel B. Note karst and closed depressions occur at the ends
of the coils. (Images provided courtesy of G. William Monaghan.)
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125 123
this anomaly is an abandoned coil, other chronological and
morphological data may shed light on the timing of its erasure and
possible mound reconﬁguration. For example, the age of the pre-
mound paleosol in Core 13 (419e228 BC, BETA 337170; Table 1),
which lies adjacent to the abandoned coil (C-13; Fig. 1A), is consis-
tent with other dates reported for the base of the mound. The con-
sistency of this age with those from across the mound suggests that
the abandoned coil may have been built earlier and erased during
the initial construction of Serpent Mound. If so, the erasure of the
coil may have marked an early Adena design change that reconﬁg-
ured Serpent Mound to create new or expanded cultural symbolism.
By integrating previously reported absolute ages (e.g., Fletcher
et al., 1996) and archaeological interpretations (e.g., Grifﬁn, 1943)
C ages from the base of the mound (Table 1), we have
reassessed the construction age of Serpent Mound. This integration
provides a more complete and realistic resolution to the mound's
construction chronology and suggests a more complex history for
the mound. The similarity and consistency of ages within the ZOD/
(Fig.3; Table 1)
their broad distribution along the extent of the
mound (Fig. 1), and a complete lack of post-Adena charcoal in the
provide the strongest evidence for when the submound
paleosol was buried and Serpent Mound construction began. Ser-
pent Mound was initially constructed 2300 years ago during the
Early Woodland (Adena) Period, but was then modiﬁed, repaired,
or renewed ~1400 years later during the Late Prehistoric (Fort
Ancient) Period. Whether the mound was in continuous use during
the 1400 years between its initial construction and subsequent
repair is not clear from these data. However, a possible erased coil
near the head of the serpent indicates that other alterations,
potentially several hundred years earlier than the Fort Ancient re-
pairs, may also have occurred. This suggests a deeper, richer, and far
more complex history for Serpent Mound than previously known.
Renovating or reuse of cultural monuments is not unusual
worldwide and often occurs when a new culture enters a region.
However, another interpretation of the integrated data from Ser-
pent Mound suggests that it was regularly used, repaired, and
possibly reconﬁgured by local groups for more than 2000 years,
which may imply at least some level of long-term cultural conti-
nuity in the use of this iconic monument. Whether cultural conti-
nuity predominated at Great Serpent Mound is unknown, but
should be a focus for new research in the region.
The Serpent Mound Project was initiated in 2010 by William F.
Romain. Its continuing objective has been to conduct multidisci-
plinary scientiﬁc research using noninvasive and minimally invasive
methods. Over the years, many people have contributed time and
funding to the project. Among those we wish to thank are: Jeff
Wilson, Delsey Wilson, Jim McKenzie, Beverly McKenzie, Horton
Hobbs, Mike Zaleha, and Al Pecora. Funding for radiocarbon dating
was generously provided by Friends of Serpent Mound, Gray &Pape,
Inc., ASC Group Inc., and Ohio Valley Archaeology, Inc. We owe a
huge debt of thanks to Kris Phipps, Serpent Mound site manager.
Thanks also to Nancy Stranahan, Megan McCane, Glen Horton, Jeff
Dilyard, Harry Campbell, and Jamie Davis. Permission to conduct
research at Serpent Mound was provided by the Ohio Historical
Society. For various assistance and support we thank George Kane,
Brad Lepper, Bill Pickard, and Karen Hassel. Judson Finley and Mike
Kolb provided editorial advice. We acknowledge the use of equip-
ment from the Glenn A. Black Laboratory of Archaeology.
The authors are solely responsible for the content of this paper.
Abrams, Elliot M., Le Rouge, Mary F., 2008. Political complexity and mound con-
struction among the Early and Late Adena of the Hocking Valley, Ohio. In:
Otto, M.P., Redmond, B.G. (Eds.), Transitions: Archaic and Early Woodland
Research in the Ohio Country. Ohio University Press, Athens, pp. 214e231.
Black, Glenn A., 1967. Angel Site: an Archaeological, Historical, and Ethnological
Study. Indiana Historical Society, Indianapolis.
Bronk Ramsey, C., 1994. Analysis of chronological information and radiocarbon
calibration:program OxCal. Archaeol. Comput. Newsl. 41, 11e16.
Bronk Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy: the
OxCalProgram. Radiocarbon 37 (2), 425e430.
Bronk Ramsey, C., 2001. Development of the radiocarbon calibration program
OxCal. Radiocarbon 43, 355e363. Proceedings of 317th International 314C
Bronk Ramsey, C., 2008. Radiocarbon dating: revolutions in understanding.
Archaeometry 50 (2), 249e275.
Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51 (1),
Bronk Ramsey, C., 2013. OxCal 4.1 Manual. http://c14.arch.ox.ac.uk/oxcalhelp/hlp_
contents. html (accessed November 2013).
Bronk Ramsey, C., Dee, M., Lee, S., Nakagawa, T., Staff, R., 2010. Developments in the
calibration and modeling of radiocarbon dates. Radiocarbon 52 (3), 953e961.
Brown, James A., 2005. Reﬂections on taxonomic practice. In: Applegate, D.,
Mainfort Jr., R.C. (Eds.), Woodland Period Sytematics in the Middle Ohio Valley.
University of Alabama Press, Tuscaloosa, pp. 111e119.
Burks, Jarrod, 2012. Ohio's great serpent Mound surveyed. ISAP News 32, 6e7.
Clay, Berle R., 2005. Adena: rest in peace? In: Applegate, D., Mainfort Jr., R.C. (Eds.),
Woodland Period Sytematics in the Middle Ohio Valley. University of Alabama
Press, Tuscaloosa, pp. 94e110.
Cook, Robert A., 2008. SunWatch: Fort Ancient Development in the Mississippian
World. University of Alabama Press, Tuscaloosa.
Fletcher, Robert V., Cameron, Terry L., Lepper, Bradley T., Wymer, Dee Anne,
Pickard, William, 1996. Serpent Mound: a fort ancient icon? Midcont. J.
Archaeol. 21 (1), 105e143.
Fontaine, S., Barot, S., Barre, P., Bdioui, N., Mary, B., Rumpel, C., 2007. Stability of
organic carbon in deep soil layers controlled by fresh carbon supply. Nature
Greber, N'omi B., 2005. Adena and Hopewell in the middle Ohio Valley: to be or not
to be? In: Applegate, D., Mainfort Jr., R.C. (Eds.), Woodland Period Sytematics in
the Middle Ohio Valley. University of Alabama Press, Tuscaloosa, pp. 19e39.
Greenman, E.F., 1934. Guide to Serpent Mound. Ohio Historical Society, Columbus.
Grifﬁn, James B., 1943. Fort Ancient Aspect. Museum of Anthropology, University of
Haas, Herbert H., Holliday, Vance T., Stuckenrath, Robert, 1986. Dating of Holocene
stratigraphy with soluble and insoluble organic fractions at the Lubbock Lake
archaeological site, Texas: an ideal case study. Radiocarbon 28, 473e485.
Hammes, K., Torn, M.S., Lapenas, A.G., Schmidt, M.W.I., 2008. Centennial black carbon
turnover observed in a Russian Steppe soil. Biogeosciences 5, 1339e1350.
Holliday, Vance T., 1995. Stratigraphy and Paleoenvironments of late Quaternary
Valley ﬁlls on the Southern high Plains. Geol. Soc. Am. Memoirs 186, 1e131.
Holliday, Vance T., 2001. Stratigraphy and geochronology of upper Quaternary
Eolian Sand on the Southern high Plains of Texas and New Mexico, United
States. Geol. Soc. Am. Bull. 113, 88e10 8.
Holliday, Vance T., 2004. Soils in Archaeological Research. Oxford University Press,
Inc., New York.
Lynott, Mark, 2007. The hopeton earthworks project: using new technology to
answer old, questions. In: Neusius, Sarah W., Gross, Timothy G. (Eds.), Seeking
Our Past: an Introduction to North American Archaeology. Oxford University
Press, New York, pp. 550e559.
Mainfort Jr., , Robert, C., 2005. Some comments on Woodland taxonomy in the
middle Ohio Valley. In: Applegate, D., Mainfort Jr., R.C. (Eds.), Woodland Period
Sytematics in the Middle Ohio Valley. University of Alabama Press, Tuscaloosa,
Martin, Charles W., Johnson, William C., 1995. Variation in radiocarbon ages of soil
organic matter fractions from late Quaternary buried soils. Quat. Res. 43,
Matthews, John A ., 1985. Radiocarbon Dating of Surface and Buried Soils: Principles,
Problems and Prospects, Geomorphology and Soils. Allen and Unwin, London,
Monaghan, G. William, Peebles, Christopher S., 2010. The construction, use, and
abandonment of Angel Site Mound A: tracing the history of a Middle Missis-
sippian town through its earthworks. Am. Antiq. 75 (4), 935e953.
Monaghan, G. William, Schilling, Timothy, Krus, Anthony M., Peebles, Christopher
S., 2013. Mound construction chronolgy at Angel Mounds. Midcont. J. Archaeol.
38 (2), 155e170 .
OGRIP, n.d. Ohio Geographically Referenced Information Program. (accessed June
Payette, S., Delwaide, A., Schaffhauser, A., Magnan, G., 2012. Calculating long-term
ﬁre frequency at the Stand Scale from charcoal data. Ecosphere 3 (7), 1e16.
Putnam, Frank W., 1890. The serpent Mound of Ohio. Century Mag. 39, 871e888.
Reidel, S.P., 1975. Bedrock geology of the serpent Mound cryptoexplosion structure,
Adams, Highlands, and Pike Counties, Ohio. Ohio Div. Geol. Surv. Report of
Investigations No. 95.
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125124
Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey,
Grootes, P.M., Guilderson, T.P., Haﬂidason, H., Hajdas, I., Hatte, C., Heaton, T.J.,
Hoffmann, D.L., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W.,
Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A.,
Turney, C.S.M., van der Plicht, J., 2013. IntCal13 and Marine13 radiocarbon age
calibration curves 0e50,000 years cal BP. Radiocarbon 55 (4), 1869e1887.
Sanborn, P., Geertsema, M., Timothy, A.J., Hawkes, B., 2006. Soil and sedimentary
charcoal evidence for Holocene Forest ﬁres in an inland temperate rainforest,
East-central British Columbia, Canada. Holocene 16 (3), 415e427.
Saunders, J.W., Mandel, R.D., Sampson, C.G., Allen, C.M., Allen, E.T., Bush, D.A.,
Feathers, J.K., Gremillion, K.J., Hallmark, C.T., Jackson, H.E., Johnson, J.K.,
Jones, R., Saucier, R.T., Stringer, G.L., Vidrine, M.F., 2005. Watson Brake, a
middle archaic Mound complex in Northeast Louisiana. Am. Antiq. 70 (4),
Squier, Ephraim G., Davis, Edward H., 1848. Ancient Monuments of the Mississippi
Stein, Julie K., 1986. Coring archaeological sites. Am. Antiq. 51 (3), 505e527.
Wang, Yang, Amundson, Ronald, Trumbore, Susan, 1996. Radiocarbon dating of soil
organic matter. Quat. Res. 45, 282e288.
Webb, William S., Baby, Raymond A., 1957. The Adena People No. 2. Ohio Historical
Webb, William S., Snow, Charles E., 1945. The Adena People. In: Reports in An-
thropology and Archaeology, vol. 6. University of Kentucky, Lexington.
E.W. Herrmann et al. / Journal of Archaeological Science 50 (2014) 117e125 125