Hopewell geometric earthworks: a case study in the referential and experiential meaning of monuments Wesley Bernardini Department of Sociology and Anthropology, University of Redlands, 1200 E. Colton Ave., USA Received 15 October 2003; revised 3 February 2004 Abstract Archaeological landscapes with dispersed settlements often contain widely spaced, morphologically similar, non-res- idential monuments (e.g., Neolithic megaliths and enclosures, Eastern Woodlands conical burial mounds, Southwestern great kivas, and Hopewell geometric earthworks). These monuments are commonly interpreted as ‘‘village surrogates,’’ places at which members of a local, dispersed community gathered to express and reproduce social ties. Some applica- tions of the village surrogate model have privileged referential meaning (what a monument symbolized) at the expense of experiential meaning (how monuments were experienced), obscuring important variability in relationships between monuments and use-groups. Focusing on a cluster of five likely contemporary Hopewell geometric earthworks in south- central Ohio, this paper emphasizes that the construction of monuments in dispersed settings was not always experi- enced as the aggregation of autonomous, isomorphic communities. An analysis of labor involved in earthwork construction demonstrates that in the Hopewell case, a very widely dispersed population, not exclusively affiliated with individual monuments, gathered repeatedly to build a related set of ceremonial centers. Parallels with the Chaco Phe- nomenon of northwest New Mexico are explored, and the importance of distinguishing between referential and expe- riential meaning in the broader study of prehistoric monuments is discussed. Ó 2004 Elsevier Inc. All rights reserved. The construction of large-scale monuments by dis- persed populations was not uncommon in prehistory. Examples include the Neolithic henges and enclosures of Europe, Woodland Period mounds in the eastern United States, and great houses and great kivas in the American Southwest. Hopewell geometric earthworks, built between ca. A.D. 1–500 in and around southern Ohio, are among the most impressive examples of such monuments, consisting of earthen embankments ar- ranged into precise geometric shapes up to 5.2 m (17 ft) tall and more than 300 m (1000 ft) in diameter (Fig. 1). Monuments in dispersed social landscapes have typi- cally been interpreted as focal points for a surrounding community. In the absence of a large, fixed settlement, monuments are thought to have served as ‘‘village surro- gates,’’ venues at which relationships among a local, dispersed population were created, symbolized, and repro- duced (Hodder, 1984; Sherratt, 1984, 1990). Hopewell earthworks have been interpreted using a variant of the ‘‘village surrogate’’ idea, the ‘‘vacant ceremonial center model’’ (Dancey and Pacheco, 1997a), which proposes that each earthwork organized a number of small, dispersed set- tlements into a community through group ceremonies. In the case of Hopewell earthworks, and other simi- lar monuments, the village surrogate interpretation is 0278-4165/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jaa.2004.06.001 E-mail address: [email protected]. Journal of Anthropological Archaeology 23 (2004) 331–356 www.elsevier.com/locate/jaa
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Journal of Anthropological Archaeology 23 (2004) 331–356
www.elsevier.com/locate/jaa
Hopewell geometric earthworks: a case study in thereferential and experiential meaning of monuments
Wesley Bernardini
Department of Sociology and Anthropology, University of Redlands, 1200 E. Colton Ave., USA
Received 15 October 2003; revised 3 February 2004
Abstract
Archaeological landscapes with dispersed settlements often contain widely spaced, morphologically similar, non-res-
around a monument because they use it as a central
place). Such circularity can be especially difficult to
break when, as in the case of Hopewell earthworks,
there are so little available data directly pertaining to
the behavior being explained (e.g., material residue of
gatherings at monuments by local communities; evi-
dence of the dispersed settlements themselves) to provide
‘‘resistance’’ to interpretation (Wylie, 1994).
One benefit of the proposed heuristic division is that
it encourages more critical examination of physical
experience separate from inferred meaning. It also per-
mits the experiential aspect of meaning to be explored
as an independent variable, rather than simply as an
implication of referential meaning. For ‘‘vacant’’ monu-
ments like Hopewell earthworks, where little but earthen
architecture remains for analysis, experience may be
productively measured through energetic analysis
(Abrams, 1994)—the quantification of a manual con-
struction event in terms of the number of people in-
volved, the duration of the project, the area from
which participants were drawn, etc. Establishing the
energetic parameters of experience helps us to under-
stand the nature and social scale of the relationships ex-
pressed and created among participants in a common
venture.
This study presents an experiential analysis of Hope-
well earthwork construction, an energetic assessment of
the experiences of the people who built them. The results
of this analysis show that some earthworks in the core
Scioto Valley and Paint Creek areas of the Hopewell sys-
tem were not constructed by local populations affiliated
with each earthwork, as assumed by the village surrogate
(and vacant ceremonial center) model. That is, earthwork
construction was not �experienced� by participants as the
convergence of a local community. Instead, earthwork
construction was the product of labor pooled at a regio-
nal level, and thus was experienced as a much larger so-
cial phenomenon than previously recognized. This
conclusion reorients Hopewell research from questions
about the intra- and inter-community dynamics of earth-
work polities (e.g., papers in Dancey and Pacheco,
1997b) to questions about pan-regional ceremonial sys-
tems. It also demonstrates the importance of distinguish-
ing between referential and experiential meaning in the
broader study of prehistoric monuments.
Referential and experiential meaning
The concepts of referential and experiential meaning
are well illustrated by a consideration of Neolithic tombs
(Hodder, 1994). It has long been suggested that linear
Neolithic tombs �mean� houses for the dead (Childe,
1949). This referential meaning is inferred from a num-
ber of formal similarities between long barrow tombs
and earlier long houses, including an elongated trapezoi-
dal shape of similar length and width, an entrance at the
broader end, a northwest–southeast alignment, and a
linear internal division (Hodder, 1984, 1990). Tombs
are also often built on top of earlier long houses (Hod-
der and Shand, 1988), apparently referencing the older
structures in space as well as form. Thus, it seems likely
that a Neolithic citizen of north-west Europe would
have understood linear tombs to symbolize �houses for
ancestors.�To accept a broad referential meaning for all tombs,
however, is to ignore the many different ways in which
people can experience a monument at any one point in
time, or over time (Bradley, 1993; Holtorf, 1998). As
Hodder (1994, p. 85, italics added) notes,
Even if the tombs were called �houses� of the ancestorsand were built on house or settlement sites, they presum-ably came to have meaning in their own right as associ-
ated with a specific set of activities. Those who dug thesoil together, who carried the stone or timbers to makethe chambers, who carried in the dead, moving asideearlier remains of their ancestors, who gave gifts, who
burned, mounded over and closed the tomb, in theirjoint activities developed a common tradition. For themthe tomb acted less through reference and more through
direct experience.
For those involved in the communal labor project,
the experiential significance of asserting common
ancestry and continuity of rights likely superseded the
334 W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356
broader referential meaning of tombs in society. Once
tombs were closed off and moved into their ‘‘afterlife’’
(Bradley, 1993), it was no longer possible to experience
meaning through them directly, and they ‘‘came to act
as reference points on the landscape (e.g., Barrett
et al., 1991), now being meaningful less through direct
experience and more through reference to the past.
Thus, ‘‘the meaning of tomb material culture may have
shifted through time from referential to experiential to
referential again’’ (Hodder, 1994, p. 85).
Like Neolithic tombs, Hopewell geometric earth-
works have lately come to be interpreted through the
lens of a homogeneous referential meaning—in this
case, as ‘‘village surrogates.’’ This blanket interpretation
obscures potentially important variability in the ways
that sub-groups on the landscape experienced monu-
ments. Decoupling referential from experiential mean-
ing can provide productive new insights into this
variability.
Fig. 2. The broadest territorial definition of the Hopewell
phenomenon.
Fig. 3. Primary distribution of Hopewell geometric earthworks.
1 As usual for Hopewell there is at least one exception, at the
Turner Site.
Hopewell geometric earthworks
The Hopewell phenomenon is usually defined by the
presence of one or more non-utilitarian goods, often
(but not only) recovered from mortuary contexts, such
as copper celts, mica cutouts, and bear canines (Struever
and Houart, 1972). In its most extreme definitions,
Hopewell covers most of eastern North America in the
Middle Woodland Period (ca. A.D. 1–500), from Ontar-
io to Louisiana and from New York to Florida (Fig. 2).
There is, however, an increasing recognition that defin-
ing the boundaries of the Hopewell phenomenon
through a composite of diagnostic artifacts encompasses
a tremendous range of local variability in social and rit-
ual organization (Carr et al., 2002).
This study focuses on the core area of the Hopewell
phenomenon, widely agreed to be centered on the Scioto
Valley of south-central Ohio, but defined more specifi-
cally here by the distribution of geometric earthworks.
The distribution of Hopewell geometric earthworks cov-
ers a much smaller territory than Hopewell artifacts,
with the densest concentration centered on the modern
town of Chillicothe (Fig. 3). As their name implies, geo-
metric earthworks are composed of embankments of
earth arranged into various geometric shapes, including
circles, squares, octagons, and ‘‘roads’’ (parallel
embankments of earth). The scale of these constructions
is immense, with individual geometric shapes enclosing
areas of 30 acres or more, equivalent to more than 25
football fields.
Hopewell geometric earthworks were the most visi-
ble products of Hopewell society, yet despite their
prominence, their use (intended or actual) remains un-
clear. Unlike conical burial mounds, the most common
Middle Woodland earthen construction, geometric
earthworks were not used as repositories for the dead
or their associated offerings,1 though earthworks some-
times enclosed spaces in which burial mounds were con-
structed and mortuary rituals were performed.
Although geometric earthworks consisted of high walls
and ditches, they were almost certainly not used for de-
fense since their embankments contain many openings
and often have interior ditches, rather than exterior
W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356 335
ones. Earthworks are generally free of trash or habita-
tion debris, though some exceptions are known (in most
of these cases, occupation seems to predate earthwork
construction; see below). Some scholars have proposed
that earthworks were astronomical observatories (Hive-
ly and Horn, 1982), noting alignments through gate-
ways and mounds. Most researchers attribute a
symbolic meaning to the precise geometric arrange-
ments that comprise the earthworks, for example, as
representations of winter and summer moiety ‘‘big
houses’’ (DeBoer, 1997), or as broader materializations
of ancient cosmologies (Romain, 1996). Beyond these
outlines, little is known about the details of geometric
earthwork construction and use, and until recently little
field research had been conducted on them (but see Gre-
ber, 1999, 2002; Lynott, 2003; Lynott and Weymouth,
2002).
ig. 4. The vacant ceremonial center model of hopewell
ettlement. Redrawn after Dancey and Pacheco (1997, Figure
.2).
Village surrogates
The presence of isolated monuments within dispersed
settlement systems is common, especially in prehistoric
North America and Europe. An insightful explanation
of these monuments, presented most clearly by Euro-
pean archaeologists, views monuments in dispersed set-
tings as ‘‘village surrogates’’ (Hodder, 1984; Renfrew,
1976; Sherratt, 1984, p. 129; Sherratt, 1990, p. 148–
149). That is, in the absence of an actual nucleated
village, members of a dispersed population created,
reproduced, and symbolized a community through the
construction and use of a central monument. Among
European examples, Neolithic megaliths (Chapman,
1981; Renfrew, 1976; Sherratt, 1990) and enclosures
(Evans et al., 1988) have been profitably interpreted in
this manner. Similar interpretations have been made
for monuments in dispersed settings in North America.
For example, Woodland period conical burial mounds
in eastern North America have been interpreted as
markers of a group�s investment and ancestry in an area
(Buikstra, 1979; Buikstra and Charles, 1999; Charles,
1985). Late Woodland period animal effigy mounds in
Wisconsin, often occurring as lines of repeated figures,
are thought to have been built during seasonal aggrega-
tions of totemically affiliated groups of mobile foragers
(Benn, 1979; Storck, 1974). In the American Southwest,
isolated great kivas have been interpreted as integrative
facilities at which the residents of surrounding dispersed
farming settlements converged to conduct ceremonies
and exchange goods, mates, and information (Adler,
1990). Isolated Chacoan great houses (outliers) in the
American Southwest, many of which are associated with
a great kiva, are also frequently interpreted as focal
points of dispersed communities (Breternitz et al.,
1982; Marshall et al., 1979; Lekson, 1991; but see Maho-
ney, 2000).
The vacant ceremonial center model
Many Hopewell researchers accept a variant of the
village surrogate model for Ohio Hopewell geometric
earthworks known as the ‘‘vacant ceremonial center
model,’’ variously defined by Prufer (1964a, p. 71;
1964b;1965, p. 137)., Pacheco (1996), Smith (1992),
and most recently by Dancey and Pacheco (1997a).
The model was originally borrowed from Mesoamerica
where it was used to explain Mayan ceremonial com-
plexes, though subsequent discoveries of substantial
activity and occupation at these complexes led to the
dropping of the term in that area (Morley et al., 1983).
The vacant center model proposes that Ohio Hope-
well peoples organized themselves into settlements, or
‘‘hamlets,’’ of one to a few households that were distrib-
uted around a ceremonial earthwork center. The central
assumption of the model is that a single earthwork
served as the ceremonial center for each community of
dispersed hamlets. Schematically, then, the vacant center
model envisions the Hopewell landscape as divided into
autonomous polities, each consisting of an earthwork
surrounded by its associated, dispersed community, as
depicted in Fig. 4. Under the vacant center model, ham-
lets should be clustered around earthworks, with bound-
aries or gaps in settlement between these dispersed
communities.
Unfortunately, the settlement pattern data needed
to evaluate the expectations of the vacant center model
are lacking for the Hopewell period and for Ohio in
general. A recent tabulation of Middle Woodland (ca.
F
s
1
336 W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356
A.D. 1–500) habitation and non-mortuary sites in Ohio
compiled in 1997 contains only 91 sites for the entire
state (Dancey and Pacheco, 1997a, Table 1.1), many
of which are marginal candidates for residential status.
The few areas that have been intensively covered by
systematic block surveys are concentrated almost exclu-
sively in the immediate area of known mounds and
earthworks (e.g., Lynott, 1982; Lynott and Monk,
1985; Seeman, 1981) and thus do not cover potential
community boundary areas. Further, of the suspected
habitation sites identified from surface remains or sho-
vel testing, few have been tested extensively enough to
determine the relationship between surface and subsur-
face expressions. Dancey (1991, p. 68), for example,
acknowledges that ‘‘[The Murphy Site] may be the
only comprehensively documented [Middle Woodland
Period habitation] site in central Ohio (if not all of
Ohio).2’’
Critiques of the village surrogate model
As research on monuments in dispersed settlement
systems has progressed, some of the original formula-
tions of the village surrogate model have been chal-
lenged. As initially conceived, the village surrogate
model hypothesized that increasing reliance on agricul-
ture and a resultant increase in territoriality motivated
the construction of fixed markers to legitimize control
of restricted resources. Recent research has, however, re-
vealed that ‘‘full-scale agriculture was more the excep-
tion than the rule’’ (Chapman, 1995, p. 39) among
monument builders in dispersed settlement systems.
Without fixed investments to defend, monuments may
have been constructed along lines of movement rather
than in the centers of defended territories, and thus
may not correspond in direct ways to local communities
(Bradley, 1993; Chapman, 1995).
The village surrogate model can also be critiqued for
focusing on the distribution of monuments rather than
the surrounding dispersed settlements. Habitation sites
are typically difficult to identify archaeologically in dis-
persed settlement systems; in fact, somewhat circularly,
2 The lack of settlement pattern data does not reflect a lack
of effort by current Hopewell researchers, who have focused
increasing energy on the collection of such information (e.g.,
Dancey, 1991, 1992; Pacheco, 1988, 1993). Rather, it is the
cumulative product of several factors, not least of which is the
longstanding bias towards mortuary contexts over the first 100
years of the development of Hopewell archaeology, leaving
considerable ground to be caught up. The study of Hopewell
habitations is also not helped by the rapid and deep burial of
sites in the eastern Woodlands, extensive site disturbance by
timber cutting and farming, and the wooden architecture of
prehistoric Hopewell settlements, all of which conspire to make
detection of small residential sites difficult.
it is often the ephemeral nature of the dispersed commu-
nity that supports the interpretation of monuments as
fixed, central gathering places. When small sites are
archaeologically visible and adequate survey data are
available, as for example in the American Southwest,
it is clear that the settlement patterns of dispersed sites
do not often conform to the expectations of the village
surrogate model. For example, dispersed settlements
near Chacoan great houses often do not follow the
boundaries of Thiessen polygons separating hypotheti-
cal great house territories (Mahoney, 2000); while small
groups of structures generally do cluster around great
houses, contemporary settlements can be found up to
13miles distant in seeming ‘‘no man�s lands’’ (e.g., Ba-
ker, 1991).
Finally, variability in the function of morphologically
similar earthworks is being increasingly recognized. Sub-
stantial differences in labor investment are often masked
under headings assigned to monuments built in the same
form. For example, Late Woodland period thunderbird
effigy mounds in Wisconsin can be found with wing-
spans of as little as 15m (50ft) to as much as 770m
(2500ft) (Rosenbrough and Birmingham, 2003), imply-
ing use-groups of very different scales. In Illinois and
Ohio, Ruby et al. (2004) document the variety of ways
in which earthen enclosures and mounds can organize
surrounding populations of different sizes and
compositions.
Ethnographic cases have helped to broaden the range
of potential uses of morphologically similar monuments.
For example, while some of the smaller ceremonial cen-
ters used by the Chachi of northwest Ecuador organize
local dispersed communities, the largest and oldest cen-
ter, Punta Venado, serves a much broader population
drawn from at least five distinct communities located
along 16km of river-front territory (DeBoer and Blitz,
1991). A similar pattern of convergence on a single cer-
emonial center by a broadly dispersed population was
also recorded for the Mapuche of south-central Chile
by Dillehay (1990). The Mapuche�s biannual nguillatun
ceremony, hosted in rotation by different local groups,
draws up to 8000 attendants. Significantly, an individual
is ‘‘invited to several different ceremonies each year, as a
member of their own lineage and as an outsider’’ (Dille-
hay, 1990, p. 227); attendance is not limited to members
of the local community.
The recognition that not all monuments in dispersed
landscapes are created equal opens up an exciting range
of possible interpretations. For example, a single com-
munity may build and use multiple monuments with dif-
ferent functions, and multiple communities may
converge on a single monument. The following section
approaches a set of Hopewell geometric earthworks with
an eye toward such variability, and finds evidence for a
very different pattern of use than previously attributed
to these sites.
Fig. 5. The Scioto and Paint Creek Valleys of Ohio showing the
location of the five tripartite Hopewell geometric earthworks
included in the analysis.
Fig. 6. The five tripartite earthworks of the Scioto and Paint
Creek Valleys. North varies for each site.
W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356 337
Tripartite Hopewell earthworks: a unique set
Significant differences in scale, morphology, and
placement on the landscape are evident within the
known assemblage of Hopewell geometric earthworks.
In particular, morphology appears to differ in important
ways across time and space, with local clusters of earth-
works exhibiting strong similarities in the kinds and
arrangements of shapes they contain (DeBoer, 1997).
This study focuses on a set of five uniquely similar
earthworks located in the Scioto and Paint Creek Val-
leys of south-central Ohio (Fig. 5). The five sites are:
Seip and Baum in the main Paint Creek valley; Frank-
fort in the North Fork of Paint Creek; and Liberty
and Works East in the main Scioto valley. These earth-
works represent some of the largest monuments ever
constructed by Hopewell groups, and are located in
what most researchers agree was the ‘‘core’’ of Hopewell
activities. They receive special attention here because
their morphological similarities suggest that they were
built and used within a relatively short time of each
other. The ability to identify earthworks that were likely
built and used within a single human generation is extre-
mely unusual for Hopewell archaeology, since absolute
chronological control is relatively poor.3
Each of the tripartite earthworks consists primarily
of three conjoined shapes (Fig. 6): a square with sides
of approximately 305m (1100ft) in length; a large circle
or partial circle with a diameter of approximately 460m
(1500ft); and a smaller circle with a diameter of about
200m (650ft). While many Hopewell earthworks con-
tain combinations of circles and squares, these five sites
stand out for their tripartite configuration and the stan-
dardized dimensions of their component shapes. Outside
the Scioto Valley area, only the Marietta earthworks
(160km [100miles] east) contains a square of this size,
and this complex lacks circles altogether. With two
exceptions (Newark and Seal), the 460m diameter circle
and 200m circles are exclusive to the five tripartite earth-
works, and neither exception includes a comparable
305m square.
Squier and Davis (1848, p. 56) were the first to point
out the similarities of this set of earthworks: ‘‘This work
[Liberty] is a very fair type of a singular series occurring
in the Scioto valley, all of which have the same figures in
combination, although occupying different positions
with respect to one another, viz. a square and two cir-
cles.’’ The authors went on to comment (Squier and Da-
vis, 1848, p. 57): ‘‘That there is some hidden significance,
in the first place in the regularity, and secondly in the
3 Despite a modest number of radiocarbon dates (Greber,
2002) and an important artifact seriation (Ruhl, 1992), the
construction and use of most sites cannot be dated to intervals
smaller than a century or more.
arrangement of various parts, can hardly be doubted.
Nor can the coincidences observable between this and
the other succeeding works of the same series be wholly
accidental.’’ Succeeding scholars have also pointed out
the striking similarity of this set of sites. For example,
Greber (1979, p. 36) noted that ‘‘within the central Sci-
oto–Paint Creek area, the series of five sites identified
by the tripartite earthwork design form a regional
subunit.’’
The strong morphological similarities among this set
of five earthworks is very likely the result of close inter-
action between the people who planned and built them.
Greber (1997a, p. 219) notes that ‘‘the five square enclo-
sures are almost identical in size, construction material,
338 W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356
iconographic detail, and their nonrandom orientations
on the landscape. This strongly suggests that each was
part of one overall architectural design likely to have
been built over a relatively short time span.’’ The simi-
larities among the earthworks are so strong that it is
not unreasonable to consider that they were all laid
out by the same person or group of people with the req-
uisite ritual and architectural knowledge (James Mar-
shall, personal communication, 2002)—a ‘‘unique
relatively short time within the Ohio Hopewell period
when political and social ties were strong enough within
the central Scioto to provide a basis for the close eco-
nomic and symbolic sharing’’ (Greber, 1997b, p. 246).
The case for contemporaneity of these five earth-
works rests primarily on their pronounced morphologi-
cal similarities. No absolute dates are available from
embankment contexts of the tripartite earthworks; C14
dates from mound contexts at Seip and Liberty overlap
generally for the two centuries between A.D. 200–400
Thompson, 1966). Hopewell people farmed eastern agri-
cultural complex crops in small garden plots, but they
relied on gathered nuts, fruits, and tubers for an impor-
tant component of their diet (Wymer, 1997); they were
most likely only part-time farmers. Thus, it may be use-
ful to consider cross-cultural hunter gatherer population
density as another source of comparative information
about Hopewell density. In a cross-cultural sample of
14 hunter-gatherer groups from temperate forests, the
closest approximation of the Ohio Middle Woodland
environment, Kelly (1995, pp. 224–225, Table 6-4) re-
corded an average population density of .13 people/
km2 (.06 people/miles2). The maximum temperate forest
hunter-gatherer density in Kelly�s sample was .42 people/
km2.
Together, archaeological estimates and cross-cultural
figures suggest that Ohio Hopewell population density
was probably not greater than about .5 people/km2, with
a maximum of no more than 1 person/km2. In calculat-
ing labor catchment areas, it is assumed that roughly
half of the population of any given area was of adequate
age and health to participate in the ‘‘heavy lifting’’ that
comprised the bulk of earthwork construction (probably
a liberal assumption).6 Thus, to determine labor catch-
ment areas, the labor crew figures listed in Table 3 must
first be doubled to reflect the total population from
which the laborers would have been drawn.
Several scenarios were evaluated in modeling labor
catchment areas. What is considered to be the most rea-
sonable scenario involves laborers working for 25 days a
year for 10 years to complete each earthwork complex,
and assumes a population density of .5 people/km2. This
scenario would have involved about 350 people/year at
each tripartite earthwork. Keeping in mind that a total
population of 700 people would have been necessary
to field 350 workers, a crew of this size would have been
drawn from a catchment area with a diameter of about
42km. Increasing the number of work days per year to
50 decreases the average diameter to 30km.
Distances separating earthworks from their nearest
neighbors range from 6km for Seip and Baum to
10km for Works East and Liberty, to 22km for Frank-
fort and Works East. Fig. 10A plots labor catchment cir-
cles around each earthwork complex calculated at 25
days/year for 10 years of construction at .5 people/
km2. Portions of catchment circles are shaded when they
overlap with a neighboring earthwork�s labor catchment
area. Under this labor scenario, Works East, Liberty,
and Frankfort overlap more than 50% of each other�scatchments, while Seip overlaps 100% of Baum�s catch-ment. Fig. 10B shows overlapping catchments for the
50 work-days/year scenario, which decreases the average
overlap only slightly to about 45% (not including the
complete overlap of Baum by Seip).
Less conservative labor scenarios produce larger la-
bor crews, larger labor catchment areas, and greater
overlap. If each earthwork complex was built over 5
years, labor catchments range in size from 44 to 60km
in diameter, depending on the number of days worked
(50 or 25). If each geometric shape was built over the
W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356 347
course of 5 years, construction of some of the larger
embankments, such as the square at the Liberty earth-
works, would have required 280–550 laborers; at a den-
sity of .5 people/km2, such a crew would have been
drawn from a catchment 36–53km in diameter. The
Fig. 10. Labor catchment areas for the five tripartite earthworks calcu
days per year; (B) 50 work-days per year.
870–1730 people required to build the average individual
geometric shapes (squares and large circles) in a single
year at 50/25 days/year and .5 people/km2 would have
been drawn from an area 66–94km across. Small
embankment circles, in contrast, required much smaller
lated for 350 laborers at a density of .5 people/km2: (A) 25 work-
348 W. Bernardini / Journal of Anthropological Archaeology 23 (2004) 331–356
labor crews, and could have been built in a single year
under the same parameters by 50–90 laborers drawn
from a 16 to 22km diameter catchment.
Significantly, even the most conservative labor sce-
nario modeled here would have required overlapping
labor pools or laborers from outside the local area sur-
rounding each earthwork. Thus, construction for two
and a half months (50 days) a year for 10 years to com-
plete each earthwork complex, with a high population
density estimate of 1 person/km2, would still have re-
quired that laborers be drawn from a large average
catchment area of 22km in diameter. Even for this con-
servative scenario there is considerable overlap in the
labor catchments of the tripartite earthworks (45%
overlap for Liberty, Frankfort, and Works East; 80%
overlap for Seip and Baum). That is, it would be diffi-
cult for construction to occur on neighboring earth-
works in the same generation without drawing on the
same pool of laborers, or drawing on very distant
populations.
Fig. 11 illustrates labor catchments designed to avoid