You are on page 1of 11

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 144:30–40 (2011)

Bioarchaeological Analysis of Diet During the Coles


Creek Period in the Southern Lower Mississippi Valley
Ginesse A. Listi*

Department of Geography and Anthropology, Louisiana State University, Baton Rouge, LA

KEY WORDS dental pathologies; Coles Creek diet; dietary transition

ABSTRACT The timing of the dietary shift from for- rated into Pre-Coles Creek and Coles Creek categories
aging to maize agriculture, and the speed at which such and statistical analyses were used to assess temporal
practices were adopted, are important considerations in variation in pathology load. Results indicate that pathol-
the cultural evolution of the New World. In the southern ogy load in the Coles Creek sample is slightly heavier
Lower Mississippi Valley, maize agriculture traditionally than the Pre-Coles Creek sample; however, the differen-
was believed to have been practiced during the Coles ces are not substantial. Furthermore, data suggest that
Creek period (A.D. 700–1200); however, direct evidence regional differences in resource exploitation existed
for maize is lacking in the archaeological record prior to between the Lower Mississippi Valley and populations
A.D. 1000. The present study examines Coles Creek diet elsewhere in the eastern United States. Specifically, the
from a bioarchaeological perspective. Oral-health indica- presence of starchy native plants other than maize in
tors, including abscesses, antemortem tooth loss, calcu- the diet likely contributed to a high pathology load for
lus, carious lesions, periodontal disease, and tooth wear, early hunter-gatherers. Ultimately, data from this study
were evaluated in a regional, temporal context. Data complement the archaeological, botanical, and zooarch-
were collected from 288 dentitions from eight sites in the aeological records and indicate that Coles Creek subsist-
southern Lower Mississippi Valley that range in date ence was not based on maize agriculture. Am J Phys
from 800 B.C. to A.D. 1200. The sample then was sepa- Anthropol 144:30–40, 2011. V 2010 Wiley-Liss, Inc.
C

Archaeologists have long recognized the relationship Creek period from hunting-gathering to maize agricul-
between diet, population size, and social complexity (e.g., ture would be evident in the pathology load. Specifically,
Schoeninger and Shurr, 1994). In the Lower Mississippi when compared to the Pre-Coles Creek sample, the
Valley, complex societies, as indicated by increasing frequency of all pathological conditions would be higher
population density, regional mound centers, and the in the Coles Creek sample, with the exception of tooth
emergence of hierarchically ranked societies are first wear, which would be lower.
recognized during the Late Woodland Coles Creek period
(A.D. 700–1200; Kidder, 1992a, 2002). Based on the
number and widespread distribution of sites and the COLES CREEK CULTURE AND DIET
assumed correlation between social complexity and agri-
culture, archaeologists originally believed that Coles The Coles Creek period in the Lower Mississippi River
Creek diet was based on maize agriculture (Byrd and alluvial valley is delineated by radiocarbon dates from
Neuman, 1978; Haag, 1978; Williams and Brain, 1983). approximately A.D. 700 to 1200; thus, it straddles the
However, little direct evidence for maize has been found divide between the Late Woodland and Early Mississip-
in the Lower Mississippi Valley prior to A.D. 1000 (Fritz pian periods as defined for the southeastern United
and Kidder, 1993; Fritz, 1995; Kidder, 2002). States as a whole (Jeter and Williams, 1989; Kidder,
The present study investigates Coles Creek diet from 1992a). The Coles Creek period often is described as a
a bioarchaeological perspective within the context of the dynamic time during which significant changes occurred
dietary transition in the southern Lower Mississippi throughout the Lower Mississippi Valley. These changes
Valley by examining the frequency and interaction of included variation in ceramic styles as well as increases
multiple oral health indicators (including dental caries, in population size and site size and complexity (Kidder,
calculus, periodontal disease, antemortem tooth loss, 1992b, 2002). Along with these changes, archaeologists
abscesses, and tooth wear). Two hundred eighty-eight originally supposed that subsistence practices changed
adult dentitions were examined from eight sites that as well, namely, the dietary transition from hunting-
range in date from 800 B.C. to A.D. 1200. The sample gathering to maize agriculture (Byrd and Neuman, 1978;
then was divided into Pre-Coles Creek and Coles Creek
categories and statistical analyses were used to assess
*Correspondence to: Ginesse Listi, LSU Geography and Anthro-
variation in pathology load between the two groups. As pology, 227 Howe-Russell Building, Baton Rouge, LA 70803.
numerous bioarchaeological studies have demonstrated a E-mail: glisti1@lsu.edu
relationship between high carbohydrate diets and a
decline in oral health (Turner, 1979; Larsen, 1984, 1995, Received 13 January 2010; accepted 26 May 2010
1997; Norr, 1984; Perzigian et al., 1984; Smith, 1984;
Berry, 1985; Powell, 1985; Schmucker, 1985; Lukacs, DOI 10.1002/ajpa.21364
1989; Larsen et al., 1991), the hypothesis underlying Published online 18 August 2010 in Wiley Online Library
this research is that a change in diet during the Coles (wileyonlinelibrary.com).

C 2010
V WILEY-LISS, INC.
BIOARCHAEOLOGICAL ANALYSIS OF COLES CREEK DIET 31
Haag, 1978; Williams and Brain, 1983; Neuman, 1984). and carious lesions (Rose et al., 1984, 1991). These
The evidence for this supposition was indirect: the researchers conclude that the change in pathology load
locations of Coles Creek sites along the fertile alluvial was due to an increase in carbohydrates and grit in the
bottomlands beside streams and rivers would have been diet. However, since direct evidence of maize is virtually
ideal for plant cultivation; the relative predictability and absent in the archaeological record, they suggest the
high product yield of an agricultural subsistence base high carbohydrate content came instead from native
not only would have supported the increase in popula- starchy plants or seeds (Rose et al., 1984, 1991). Lastly,
tion size that occurred during this time, but also would unpublished stable isotope data, though rare, are avail-
have been necessary to provide sustenance for the large able for several sites in the southern Lower Mississippi
labor force required to build the extensive earthworks Valley and indicate that maize consumption was
that are characteristic of Coles Creek sites. While such minimal or absent at these sites (Bruce Smith, personal
indirect evidence is no longer accepted as a basis for communication to author, 2003).
inferring subsistence, this logic was used as recently as Prehistoric diet is best understood using a multi-
the 1980s (Williams and Brain, 1983). Meanwhile, disciplinary approach that combines botanical, zooarch-
archaeological and botanical data began to accumulate aeological, and bioarchaeological evidence (Sobolik,
which suggested that a dietary transition had not 1994). While archaeological and botanical research in
occurred. the southern Lower Mississippi Valley challenges the
Botanical evidence from the interior Lower Valley indi- traditional view of Coles Creek subsistence as maize
cates that maize was present as early as the eighth or agriculture, bioarchaeological research that incorporates
ninth centuries A.D. (Scarry, 1993); however, its inci- a wide range of osteological data has not been under-
dence in the archaeological record is rare. For example, taken. The current study addresses this deficiency by
at one Coles Creek site in which maize has been found, using multiple oral health indicators to assess Coles
the overall low frequencies recovered and the contexts in Creek diet in a comprehensive, systematic way. The
which it was found (e.g., middens, specialized features distinct pathology profiles exhibited in hunter-gatherers
with tobacco) led researchers to conclude that maize was and agriculturalists provide contrasting expectations by
a relatively minor food source and may have been tied to which data can be interpreted (Lukacs, 1989). The broad
ritual behavior rather than subsistence (Fritz and temporal span, yet relatively narrow geographic focus, of
Kidder, 1993; Kidder and Fritz, 1993). Conversely, sub- the skeletal populations included in this study allow
stantial data exist which demonstrate that Coles Creek pathology load to be analyzed through a regional di-
populations exploited a variety of other botanical resour- achronic framework.
ces, including nuts, fruits, and wild, or possibly,
cultivated plants (Kidder and Fritz, 1993; Fritz, 1995;
Roberts, 2000; Kidder, 2002). MATERIALS
With regard to material culture, not only are Coles A total of 288 adult dentitions were examined from
Creek pottery assemblages indistinct from other Late eight sites in the southern Lower Mississippi Valley.
Woodland cultures, but also artifacts suggestive of plant Skeletal remains from the individual sites were exca-
cultivation (e.g., hoes) are not abundant at Coles Creek vated at different times by different people over a span
sites (Kidder, 1992a). On the other hand, artifacts of nearly 40 years (from the mid 1930s to the early
related to the procurement (including the bow and 1970s). Currently, remains from individual sites are
arrow) or processing of food are ubiquitous throughout catalogued and curated in museum collections in compli-
the Lower Mississippi Valley and provide ample support ance with NAGPRA. The temporal span for the sites
that Coles Creek populations exploited natural resources extends from 800 B.C. to A.D. 1200. Based on these
through hunting, gathering, and fishing (Williams and dates, all remains were classified into two broadly-
Brain, 1983; Neuman, 1984; Kidder, 1992b; Fritz, 1995; defined temporal categories that separated the Coles
Jackson and Scott, 2002). Correspondingly, zooarchaeo- Creek sample from the Pre-Coles Creek sample. Figure 1
logical evidence from Coles Creek sites consistently depicts the location of each site in the Lower Mississippi
attests to the importance of wild game and/or aquatic Valley. Table 1 provides a summary of the skeletal sam-
resources in the diet (Springer, 1980; Mariaca, 1988; ples and MNI used in this study.
Jeter and Williams, 1989; Colburn and Styles, 1990; The Pre-Coles Creek sample is comprised of 144 indi-
Kelley, 1992; Smith, 1996; Jackson and Scott, 2002; viduals from five sites. These mound and/or midden sites
Kidder, 2002). include Little Woods (16OR1-16OR5) and Tchefuncte
Bioarchaeologically, data from the Lower Mississippi (16ST1), located in coastal regions of Louisiana, and
Valley, including the Coles Creek period, generally are Lafayette Mounds (16SM17), Greenhouse (16AV22), and
sparse: human remains typically are fragmented and Mount Nebo (16MA18), which are located in interior
poorly preserved, thus, sample sizes are small, individu- regions (see Fig. 1). The majority of the Pre-Coles Creek
als are often incomplete and commingled, and few patho- sample dates to the Tchula period, Tchefuncte culture
logical conditions have been assessed (Jeter et al., 1989; (800 B.C. to 100 B.C.); however, the burials from
Rose and Harmon, 1999). Most of the previously avail- Greenhouse and Mount Nebo date later to the Baytown
able bioarchaeological data relating to the Coles Creek period, Troyville culture (approximately A.D. 400 to
period, compiled and presented in two noteworthy publi- 700), (Ford and Quimby, 1945; Ford, 1951; Giardino,
cations (Jeter et al., 1989; Rose and Harmon, 1999), 1977, 1982; Lewis, 1991; Kidder, 2002).
consisted of demographic profiles or the analysis of one Pre-Coles Creek burial patterns include both primary
or two pathological conditions appended to site reports, and secondary, single and multiple interments. Many of
monographs, or unpublished manuscripts. The few osteo- the remains were flexed and/or bundled, but some indi-
logical studies that have been undertaken report that viduals were extended. Grave goods generally were not
Coles Creek skeletons exhibit a pathology load that is found; therefore, burials were dated based on site
different from earlier groups, particularly for tooth wear stratigraphy (Ford and Quimby, 1945; Neuman, 1984).

American Journal of Physical Anthropology


32 G.A. LISTI

Fig. 1. Approximate locations of sites used in current study.

TABLE 1. Sites and minimum number of individuals (MNI) for (Neuman, 1984; p 118). Thus, subsistence for the Pre-
Pre-Coles Creek and Coles Creek samples Coles Creek populations is believed to have been a mix
Pre-Coles Creek Coles Creek of hunting, gathering, and fishing (Ford and Quimby,
(800 B.C. to A.D. 700) (A.D. 700 to 1200) 1945; Neuman, 1984; Jeter and Williams, 1989).
Site N Site N
The Coles Creek sample is comprised of 144 individu-
als from five sites, four of which are located in interior
Greenhouse 16 Greenhouse 30 regions (Greenhouse (16AV22), Lake George (22YZ557),
Lafayette 31 Lake George 25 Lake St. Agnes (16AV26), and Mount Nebo (16MA18)),
Little Woods 45 Lake St. Agnes 3 one of which is coastal [Morton Shell Mound (16IB3); see
Mount Nebo 11 Morton Shell Mound 75
Fig. 1]. All sites are multicomponent, but the burials
Tchefuncte 41 Mount Nebo 11
Total 144 144 included in this sample date to the Coles Creek period
(A.D. 700–1200). All are mound sites and may have
served ceremonial or ritual purposes; none is strictly
habitational or residential. Therefore, individuals buried
The exception is Mount Nebo: artifacts, which might at these sites could represent ‘‘elite’’ or ‘‘special’’ mem-
have been grave offerings, were found with several of bers of the population and may not be representative of
the individuals and were used to obtain a relative date the entire population that inhabited the area or utilized
for the burials at that site (Giardino, 1982). the sites (Ford, 1951; Belmont, 1967; Robbins, 1976;
Detailed subsistence data are not available for any of Giardino, 1977, 1982; Toth, 1979; Williams and Brain,
the Pre-Coles Creek populations used in this study. 1983; Neuman, 1984; McGimsey, 2003); (Saunders and
However, during this period in the Lower Mississippi Byers, unpublished).
Valley, deer generally is the most common mammal A common characteristic of Coles Creek burials is the
species present in site middens (Neuman, 1984). Various apparent lack of order in which human remains were
smaller mammals, reptiles, fish, and birds are present as deposited (Ford, 1951; Williams and Brain, 1983). Buri-
well, and Rangia remains are prevalent at coastal sites als include single and multiple interments, bundled,

American Journal of Physical Anthropology


BIOARCHAEOLOGICAL ANALYSIS OF COLES CREEK DIET 33
flexed, or extended individuals, and isolated elements. Data collection procedures
Grave goods generally were not present; therefore, either
stratigraphy or radiocarbon dates obtained from samples Every tooth or tooth position was assessed as either
taken from the same stratigraphic layers were used to present, lost antemortem, lost postmortem, or not able to
date the burials (Ford, 1951; Belmont, 1967; Giardino, be analyzed. Each tooth or alveoli subsequently was
1977, 1982; Toth, 1979; Williams and Brain, 1983; examined macroscopically for pathological conditions. A
Neuman, 1984; McGimsey, 2003). low magnification binocular microscope was used if
the determination of the pathological condition was
questionable.
METHODS For the assessment of carious lesions, the presence
and location of each carious lesion were recorded follow-
Skeletal remains from all sites were in various states ing guidelines in Buikstra and Ubelaker (1994). To dif-
of preservation and fragmentation, ranging from friable ferentiate between true carious lesions and natural non-
fragments held together by dirt matrix and preservative carious pits and fissures (i.e., buccal pits), an attempt
to well-preserved specimens. Some elements were heav- was made to determine whether or not and the extent to
ily fragmented, while others were whole. Some individu- which the dentin underlying the enamel was affected
als were only partially represented; others were nearly (Aufderheide and Rodriguez-Martin, 1998; Ortner, 2003).
complete. MNIs reported in Table 1 were determined by If a buccal pit was present but the extent of dentin dam-
the number of dentitions present. For single interments, age could not be discerned, the lesion was noted, but not
a whole or partial dentition represented one individual. included in the analyses of carious lesions prevalence.
In the case of multiple interments, if a maxilla and man- Tooth wear was evaluated for the first molars only.
dible could be reliably associated together based on simi- This tooth was selected because it was preserved more
lar size and tooth morphology, they were counted as a often per individual than second or third molars and,
single individual; otherwise, each element was consid- therefore, was the most numerous. Wear was analyzed
ered separately. based on Scott’s methodology (1979; see also Buikstra
All individuals were ‘‘Adults’’ (greater than 18 years and Ubelaker, 1994): the occlusal surface was divided
of age), based on eruption of the third molars. However, into quadrants, a score of 1–10 was assigned to each
specific age ranges and sex were not able to be assigned based on the amount of enamel present, and the quad-
for the majority of the individuals due to fragmentation rant scores were then totaled. If the crown was broken
and commingling. In single burials, while some individu- or portions were missing, the tooth was not scored.
als had an os coxae complete enough to suggest an age For recording dental calculus, a score was assigned to
range or sex, the ossa coxarum of others were too frag- each tooth based on the amount of calculus present
mented to assess, or were not present at all. Age ranges (based on Brothwell, 1981; see also Buikstra and Ube-
and sex were not assessed in commingled remains due to laker, 1994). The ordinal categories (examples of which
the impossibility of reconciling postcranial elements with are pictured in Fig. 2) ranged from no calculus present,
dentitions. through small, moderate, and heavy amounts present.

Fig. 2. Dentitions showing different levels of calculus: none (A), slight (B), moderate (C), and heavy (D).

American Journal of Physical Anthropology


34 G.A. LISTI

Fig. 3. Dentitions showing different levels of periodontal disease: none (A), less than one-half root absorption (B), greater than
one-half root absorption (C), alveolar remnants (D, arrow), and complete alveolar resorption (D, posterior corpus).

The assessment of periodontal disease was based on a sample (Lukacs, 1989). This method is used to report
scale that standardizes the progressive deterioration of summary statistics and the average number of carious
the alveolar region for each tooth (Lukacs, 1989). The lesions per individual for each individual site as well as
ordinal categories (examples of which are pictured in for the collective Coles Creek and Pre-Coles Creek
Fig. 3) ranged from no resorption, through minimal (in samples. However, for samples in which individual pres-
which less than one-half the tooth root is exposed), ervation is varied and many specimens are incomplete,
moderate (in which more than one-half the tooth root is the individual count method can result in skewed data.
exposed), heavy (where only remnants of the alveolus Thus, an alternative technique called the mean number
remain), and complete resorption of the alveolus. of pathologies per specimen (MNPS) also was used to
Antemortem tooth loss (AMTL), which also corre- calculate pathology load (Moore and Corbett, 1971;
sponds to the most advanced stage of periodontal Lukacs, 1989). In this method, the number of teeth or
disease, was assessed using the initial recording of the alveoli affected with the pathological condition is divided
presence or absence of each tooth. Teeth lost prior to by the number of teeth or alveoli available for assess-
death were indicated by the complete resorption of the ment per individual (Moore and Corbett, 1971). The
alveolus (depicted in Fig. 3D). MNPS values are reported for each individual site as
For abscesses, Buikstra and Ubelaker (1994) recom- well as for the collective Coles Creek and Pre-Coles
mend recording the presence or absence of periapical Creek samples.
abscesses and their location (buccal or lingual). In this Temporal variation in pathology load was tested statis-
study, the same information was recorded for periodontal tically using an independent samples t test for the aver-
abscesses as well. Data on both types of abscesses were age tooth wear and MNPS for each pathological condi-
combined for analyses of temporal variation. tion. A chi-square was used to assess variation in the se-
verity of calculus and periodontal disease between Coles
Creek and Pre-Coles Creek samples. For all analyses,
Data analysis the null hypothesis was that subsistence was the same
for both Pre-Coles Creek and Coles Creek populations.
Several methods were used to analyze the pathology That is, no significant differences would exist between
load in the two samples. Population frequencies were the two samples in the MNPS or average tooth wear.
calculated using the tooth count and individual count Significant differences (if P \ 0.05), therefore, would be
methods for all pathological conditions except tooth wear indicative of a change in diet.
(for which only the average per molar was calculated).
With the tooth count method, the number of teeth
affected with pathology is counted relative to the total RESULTS
number of teeth available for study (Lukacs, 1989).
These frequencies are reported for each individual site Tables 2–5 list the individual and tooth counts for
as well as for the collective Coles Creek and Pre-Coles each site and for the collective Coles Creek and Pre-
Creek samples. Coles Creek samples. Based on individual counts, the
In the individual count method, the number of individ- frequency increased through time for carious lesions,
uals with the particular pathology is counted and di- abscesses, AMTL, and calculus, but decreased through
vided by the total number of individuals present in the time for periodontal disease. Based on tooth counts, the

American Journal of Physical Anthropology


BIOARCHAEOLOGICAL ANALYSIS OF COLES CREEK DIET 35
TABLE 2. Frequencies of dental caries based on individual and tooth counts
Individual count Tooth count
Naff/Ntota % Avg. Naff/Ntota %
Pre-Coles Creek sites
Greenhouse 2/15 13.3 0.13 2/143 1.4
Lafayette 3/15 20.0 0.33 5/87 5.7
Little Woods 0/7 0 0 0/17 0
Mount Nebo 3/11 27.3 0.72 8/111 7.2
Tchefuncte 6/30 20.0 0.33 10/240 4.2
Total 14/78 17.9 0.32 25/598 4.2
Coles Creek sites
Greenhouse 3/29 10.3 0.17 5/234 2.1
Lake George 15/25 60.0 1.60 40/477 8.4
Lake St. Agnes 1/3 33.3 0.33 1/14 7.1
Morton Shell Mound 17/68 25.0 0.29 20/247 8.1
Mount Nebo 3/9 33.3 0.44 4/68 5.9
Total 39/134 29.1 0.52 70/1,040 6.7
a
Naff 5 number of teeth or individuals affected with pathology; Ntot 5 total number of teeth or individuals available for
assessment.

TABLE 3. Frequencies of abscesses and antemortem tooth loss (AMTL) based on individual and tooth counts
Abscesses AMTL
Individual count Tooth count Individual count Tooth count
Naff/Ntota % Naff/Ntota % Naff/Ntota % Naff/Ntota %
Pre-Coles Creek sites
Greenhouse 5/14 35.7 7/164 4.3 4/16 25.0 7/225 3.1
Lafayette 4/24 16.7 7/178 3.9 9/31 29.0 33/343 9.6
Little Woods 0/18 0 0/71 0 11/45 24.4 44/287 15.3
Mount Nebo 4/11 36.4 12/95 12.6 3/11 27.3 4/133 3.0
Tchefuncte 12/29 41.4 25/193 13.0 9/41 22.0 30/611 4.9
Total 25/96 26.0 51/701 7.3 36/144 25.0 118/1,599 7.4
Coles Creek sites
Greenhouse 11/29 37.9 31/374 8.3 15/30 50.0 55/553 9.9
Lake George 11/25 44.0 19/422 4.5 11/25 44.0 49/608 8.1
Lake St. Agnes 0/3 0 0/20 0 0/3 0 0/51 0
Morton Shell Mound 14/65 21.5 29/382 7.6 13/75 17.3 40/566 7.1
Mount Nebo 2/10 20.0 4/67 6.0 4/11 36.4 6/110 5.5
Total 38/132 28.8 83/1,265 6.6 43/144 29.9 150/1,888 7.9
a
Naff 5 number of teeth or individuals affected with pathology; Ntot 5 total number of teeth or individuals available for
assessment.

TABLE 4. Frequencies of dental calculus and periodontal disease based on individual and tooth counts
Calculus Periodontal disease
Individual count Tooth count Individual count Tooth count
Naff/Ntota % Naff/Ntota % Naff/Ntota % Naff/Ntota %
Pre-Coles Creek sites
Greenhouse 9/15 60.0 49/125 39.2 14/14 100.0 106/130 81.5
Lafayette 12/13 92.3 53/84 63.1 17/24 70.8 56/132 42.4
Little Woods 4/7 57.1 12/17 70.6 19/21 90.5 65/75 86.7
Mount Nebo 7/11 63.6 31/110 28.2 7/11 63.6 63/90 70.0
Tchefuncte 28/31 90.3 190/240 79.2 28/34 82.4 148/218 67.9
Total 60/77 77.9 335/576 58.2 85/104 81.7 438/645 67.9
Coles Creek sites
Greenhouse 21/27 77.8 89/223 39.9 25/28 89.3 277/308 89.9
Lake George 21/25 84.0 294/465 63.2 24/25 96.0 306/420 72.9
Lake St. Agnes 3/3 100.0 11/12 91.7 3/3 100.0 10/16 62.5
Morton Shell Mound 53/68 77.9 178/244 73.0 34/69 49.3 125/263 47.5
Mount Nebo 6/8 75.0 37/61 60.7 6/10 60.0 31/50 62.0
Total 104/131 79.4 609/1,005 60.6 92/135 68.1 749/1,057 70.9
a
Naff 5 number of teeth or individuals affected with pathology; Ntot 5 total number of teeth or individuals available for
assessment.

American Journal of Physical Anthropology


36 G.A. LISTI

frequency increased for carious lesions, AMTL, calculus, of these data as indicative of dietary change is con-
and periodontal disease, but decreased for abscesses and founded by several factors which include sample biases
tooth wear. The decrease in tooth wear through time is and potential intersite and interregional differences in
significant (Table 5). resource exploitation.
Table 6 displays the MNPS scores and the results of The condition of the skeletal remains examined in this
the independent samples t tests. Only periodontal dis- study is less-than-ideal; therefore, sample biases associ-
ease (which decreases) shows significant variation ated with preservation and demography (or lack thereof)
through time. could be impacting the results. The former biases are
Table 7 depicts the frequencies for each category of alleviated somewhat by using multiple methods to calcu-
calculus and periodontal based on tooth and/or alveolar late population frequencies. Of the three methods used
counts. For calculus, the distribution of data is similar for in this study, the individual and tooth counts, which are
each sample. Also, though 58–61% of each sample has commonly reported in the literature, assume dentitions
calculus, the majority of teeth have a ‘‘small’’ amount as are complete. Only the MNPS statistic takes into
opposed to ‘‘moderate’’ or ‘‘heavy’’ amounts. While the account individual preservation. Nevertheless, obtaining
frequency of calculus increases slightly through time, the consistent results from these different methodologies
severity decreases. This temporal change is not signifi- might indicate that the impact of differential preserva-
cant. For periodontal disease, as with calculus, the tion is minimal. As demonstrated in Table 8, temporal
frequency increases through time; however, the severity patterns are consistent only for two of the five pathologi-
decreases. This variation is significant (P \ 0.01). cal conditions assessed with multiple methodologies (e.g.,
carious lesions and calculus). Therefore, differential
preservation likely is affecting the results for abscesses,
DISCUSSION AMTL, and periodontal disease.
The present study explores Coles Creek subsistence The lack of demographic data also is inhibiting data
from a bioarchaeological perspective within the context interpretation. Ideally, skeletal samples should be repre-
of the dietary transition in the southern Lower Missis- sentative of the entire population, as opposed to a small,
sippi Valley. As the transition occurred (and with it, an specialized subset (e.g., only adults, children, or elite
increase in the consumption of carbohydrates), the ex- members) though Wood et al. (1992, p 344) argue that
pectation was that all pathological conditions except no skeletal sample can completely represent the original
tooth wear would increase from the Pre-Coles Creek to living population, ‘‘even (one) that is a perfect random
the Coles Creek period. Table 8 summarizes the tempo- sample of all those who died.’’ Additionally, comparabil-
ral trends in data compared to expectations. Though ity in age and sex distribution among samples is impor-
results vary somewhat depending on methodology, data tant because of the influence of biological parameters
initially appear to support expectations. That is, 12 of 18 (particularly age) on pathology load. Most pathological
categories of data are consistent with expectations for an conditions are progressive; therefore, a population con-
increase in carbohydrate consumption during the Coles sisting primarily of older adults could exhibit higher fre-
Creek period. However, a straightforward interpretation quencies of pathology than a population composed of
younger adults, simply because they are older. In this
study, because specific adult age categories could not be
TABLE 5. Mean wear scores for first molars assigned for the majority of the individuals, the extent
Pre-Coles Creek Sites (N 5 89, n 5 99) M1
to which age may be affecting the results cannot be
Greenhouse 28.4 determined. Thus, the possibility that the temporal vari-
Lafayette 26.4 ation in pathology load seen in this study is related to
Little Woods 20.6 differences in age rather than diet cannot be excluded.
Mount Nebo 33.3 Despite the difficulties presented above, the collections
Tchefuncte 30.0 included in this study, while not ideal, represent the
Average 29.1 larger and more well-preserved skeletal samples avail-
Coles Creek sites (N 5 132, n 5 156) M1 able for analysis from the Lower Mississippi Valley. As
Greenhouse 28.4 such, they currently are the best (and only) alternative
Lake George 25.5
Lake St. Agnes 20.0
for researching prehistoric diet in this region from a bio-
Morton Shell Mound 23.4 archaeological perspective. However, in recognition of
Mount Nebo 21.9 the potential biases affecting the samples, results pre-
Average 25.0 sented here are conditional with the provision that,
T-test: t 5 3.993 Sig. 5 0.000* should additional materials become available in the
future, hypotheses for temporal variation in diet can be
‘‘*’’ indicates that relationship is statistically significant. reconsidered.

TABLE 6. Results of the independent samples t test of the MNPS statistic for all dental pathologies
Abscesses AMTLa Calculus Caries PODa
N Mean N Mean N Mean N Mean N Mean
Pre-Coles Creek 96 0.08 144 0.08 77 0.58 78 0.05 104 0.70
Coles Creek 132 0.06 144 0.07 131 0.60 134 0.09 135 0.55
t-value 1.232 0.636 20.410 21.596 2.722
Significance 0.219 0.525 0.682 0.112 0.007*
a
AMTL 5 antemortem tooth loss; POD 5 periodontal disease.
‘‘*’’ indicates that relationship is statistically significant.

American Journal of Physical Anthropology


BIOARCHAEOLOGICAL ANALYSIS OF COLES CREEK DIET 37
TABLE 7. Summary statistics and results of chi-square for severity of calculus and periodontal disease
Severity of calculus
None Small Moderate Heavy Total
N % N % N % N % Naffa Ntota %
Pre-Coles Creek 241 41.8 270 46.9 61 10.6 4 0.7 335 576 58.2
Coles Creek 396 39.4 520 51.7 84 8.4 5 0.5 609 1005 60.6
Chi-square X2 5 4.514; df 5 3; Sig. 5 0.211

Severity of periodontal disease


None \½ Root [½ Root Remnants Resorbed Total
N % N % N % N % N % Naffa Ntota %
Pre-Coles Creek 207 32.1 220 34.1 39 6.0 78 12.1 101 15.7 438 645 67.9
Coles Creek 308 29.1 535 50.6 75 7.1 39 3.7 100 9.5 749 1057 70.9
Chi-square X 5 80.596; df 5 4; Sig. 5 0.000*
2

a
Naff 5 number of teeth or individuals affected with pathology; Ntot 5 total number of teeth or individuals available for
assessment.
‘‘*’’ indicates that relationship is statistically significant.

With regard to intersite variation in pathology load, TABLE 8. Expectations and data trends between Pre-Coles
summary data are presented for individual sites. How- Creek and Coles Creek samples
ever, data were not assessed statistically not only Trend
because the focus of the research is regional in scope Expectationa in data
(rather than local), but also because the resulting sam-
Dental pathologies (individual count)
ples sizes would have been too small to yield meaningful
Abscess frequency 1 1
results. Nevertheless, some general observations can be AMTL frequency 1 1
made. For the Pre-Coles Creek populations, while none Calculus frequency 1 1
of the populations has a consistently light or heavy pa- Caries frequency 1 1
thology load, Little Woods has no observed caries or Caries average/individual 1 1
abscesses and is in the lower range of data for AMTL Periodontal disease frequency 1 2
and calculus (based on individual counts). Generally, Dental pathologies (tooth count)
intersite data for the Pre-Coles Creek sample are compa- Abscess frequency 1 2
rable, though the range of variation exhibited in the fre- AMTL frequency 1 1
Calculus frequency 1 1
quencies of some pathological conditions is greater than
Calculus severity 1 2
in others (i.e., calculus and periodontal disease are more Caries frequency 1 1
variable than caries, AMTL, and tooth wear). Periodontal disease frequency 1 1
For the Coles Creek sample, Lake George is conspicu- Periodontal disease severity 1 2
ously different from the other populations in its caries Wear (M1) 2 2
data: the individual count is considerably higher than Dental pathologies (MNPS)
the others and it is the only population that averages Abscess average 1 2
more than one carious lesion per individual (Table 2). AMTL average 1 2
For all other pathological conditions (as well as for caries Calculus average 1 1
Caries average 1 1
tooth counts), none of the Coles Creek populations
Periodontal disease average 1 2
stands out as having a consistently light or heavy pa-
thology load (including Lake George). On the other a
A ‘‘1’’ indicates an increase through time; a ‘‘2’’ indicates a
hand, the range of variation among Coles Creek popula- decrease through time.
tions exhibited in the frequencies of some pathological
conditions is greater than that found in the Pre-Coles
Creek sample. This latter fact, together with the caries periodontal disease (the frequencies of these two patho-
data for Lake George, may suggest that intraregional logical conditions are 58 and 68% (based on tooth counts)
variation in diet existed during the Coles Creek period and 78 and 82% (based on individual counts, respec-
with some populations consuming greater quantities of tively). These results suggest that resource exploitation
carbohydrates than others. for hunter-gatherers in the Lower Mississippi Valley dif-
In addition to the potential intraregional variation in fered from that of other regions in eastern North Amer-
diet during the Coles Creek period, apparent interre- ica. Archaeologists have long reported that populations
gional variation in resource exploitation is hindering the in the Lower Mississippi Valley relied on locally-avail-
assessment of temporal trends in pathology load based able resources that, in addition to the abundant animal
on ‘‘normal’’ expectations. All of the Pre-Coles Creek pop- and aquatic species, included a host of starchy plants
ulations are believed to have been hunter-fisher-gather- other than maize (such as amaranths, chenopods, goose-
ers (Ford and Quimby, 1945; Kidder, 2002). Elsewhere in foot, little barley, maygrass, and knotweed), (Smith,
North America, such populations typically exhibit low 1986; Fritz and Kidder, 1993; Kidder and Fritz, 1993;
rates of dental pathological conditions (with the excep- Fritz, 1995; Roberts, 2000; Gremillion, 2002; Kidder,
tion of tooth wear; Cohen and Armelagos, 1984). The pa- 2002). The presence of these plants in the diet must be
thology load in the Pre-Coles Creek sample, therefore, is contributing to the higher-than-expected pathology load
higher than anticipated, particularly for calculus and for the Pre-Coles Creek populations.

American Journal of Physical Anthropology


38 G.A. LISTI

Despite the unusually high Pre-Coles Creek pathology ences are neither significant, nor substantial. Thus, an
load, the Coles Creek sample nevertheless has higher increase in carbohydrate consumption, whether from
frequencies of most pathological conditions. However, starchy plants already present in the diet or, possibly,
with the exception of periodontal disease and tooth wear, the introduction of maize in some areas (i.e., Lake
the differences are neither significant (when tested stat- George), was minimal and occurred on a local level, if at
istically), nor substantial (when noted through general all. Ultimately, this study complements the botanical,
observation), regardless of which method is used to ana- zooarchaeological, and artifactual records and provides
lyze data. For example, based on tooth counts, increases bioarchaeological support that Coles Creek diet was not
in pathology load between samples vary only by 0.5 to based on maize agriculture.
3%. Increases in pathology load based on individual
counts are slightly higher, ranging from 1.5 to 11.2%.
These slight increases in pathology load could signify a ACKNOWLEDGMENTS
minor increase in carbohydrate consumption during the
Coles Creek period, but they would not indicate a com- The author thanks the following individuals for their
plete change in diet. Thus, at a time when populations assistance with various aspects of this research: Dr.
elsewhere in eastern North America were increasing Rebecca Saunders and Mr. Steve Fullen, Louisiana State
their reliance on domesticated plants, those in the Lower University Museum of Natural Science; Dr. Michele Mor-
Mississippi Valley continued to rely on abundant natural gan, Peabody Museum of Archaeology and Ethnology,
resources, some of which included starchy plants. Harvard University; Dr. John Verano, Tulane University;
The supposition that Coles Creek populations were not Ms. Mary Lee Eggart, Cartographer, Louisiana State
maize agriculturalists is further supported when com- University Department of Geography and Anthropology;
paring carious lesions data between the Coles Creek Ms. Mary Manhein, Louisiana State University Forensic
sample and agricultural populations from the southeast- Anthropology and Computer Enhancement Services Lab-
ern United States. Larsen et al. (1991) report carious oratory; and the Editor, Associate Editor, and two anony-
lesion frequencies of 58.9% (based on individual counts) mous reviewers for their constructive comments and sug-
and 19.4% (based on tooth counts) for Precontact Agri- gestions.
cultural populations from the Atlantic Coast of Georgia
and Florida (dating to approximately A.D. 1150–1550).
Similarly, Powell (1991) reports carious lesions frequen- LITERATURE CITED
cies of 54% (based on individual counts) and 18.7% Aufderheide AC, Rodriguez-Martin C. 1998. The Cambridge
(based on tooth counts) for the agricultural population at encyclopedia of paleopathology. Cambridge: Cambridge Uni-
Moundville, located in western, central Alabama (dating versity Press.
to approximately A.D. 1000–1450). Conversely, the fre- Belmont JS. 1967. The culture sequence at the Greenhouse
quencies of carious lesions for the Coles Creek sample Site, Louisiana. Proceedings of the Southeastern Archaeologi-
are well below those values: 29% (based on individual cal Conference. p 27–35.
counts) and 6.7% (based on the tooth counts). Though Berry DR. 1985. Dental paleopathology of Grasshopper Pueblo,
Arizona. In: Merbs CF, Miller RJ, editors. Health and disease
such comparisons to the literature can be problematic in the prehistoric southwest. Tempe: Arizona State University.
due to methodological differences among researchers, the p 253–273.
similarity between the agricultural populations and dis- Brothwell DR. 1981. Digging up bones, 3rd ed. Ithaca: Cornell
similarity between the agricultural and Coles Creek data University Press.
are notable. Buikstra JE, Ubelaker DH. 1994. Standards for data collection
from human skeletal remains. Fayetteville: Arkansas Archaeo-
logical Survey.
CONCLUSION Byrd KM, Neuman RW. 1978. Archaeological data relative to
the prehistoric subsistence in the Lower Mississippi River
Coles Creek diet traditionally was believed to have alluvial valley. Geosci Man 9:9–21.
been based on maize agriculture due to the number, size, Cohen MN, Armelagos GJ. 1984. Paleopathology at the origins
and complexity of Coles Creek sites. However, direct evi- of agriculture. Orlando: Academic Press.
dence in the form of maize kernels and pollen generally Colburn ML, Styles BW. 1990. Faunal remains from the Bangs
has been lacking. Instead, zooarchaeological and botani- Slough Site. In: Schambach FF, editor. Coles Creek and
cal evidence suggests that Coles Creek populations were Mississippi period foragers in the Felsenthal region of the
supported by native plant, animal, and aquatic natural Lower Mississippi Valley. Arkansas Archaeological Research
Series No. 39. Fayetteville: Arkansas Archaeological Survey.
resources rather than agriculture, though some culti-
p 95–108.
vated plants may have supplemented the diet. The pres- Ford JA. 1951. Greenhouse. A Troyville-Coles Creek period site
ent study considered Coles Creek subsistence from a bio- in Avoyelles Parish, Louisisana. Anthropological papers of
archaeological perspective: the dentitions of 288 adults the American Museum of Natural History, Vol. 44. Part I.
from eight sites in the Lower Mississippi Valley ranging New York.
in date from 800 B.C. to A.D. 1200 were examined for Ford JA, Quimby GI Jr. 1945. The Tchefuncte culture. An early
multiple oral health indicators and statistical analyses occupation of the Lower Mississippi Valley. Memoirs of the so-
were used to assess temporal changes in pathology load. ciety for American archaeology. Number 2. Menasha: Society
Results from this study, though provisional because of for American Archaeology.
sample biases, demonstrate that hunter-gatherers in the Fritz GJ. 1995. New dates and data on early agriculture. The
legacy of complex hunter-gatherers. Ann Miss Botan Gar
southern Lower Mississippi Valley have a higher pathol-
82:2–15.
ogy load than typically found in other areas of Eastern Fritz GJ, Kidder TR. 1993. Recent investigations into prehis-
North America due to the presence in the diet of starchy toric agriculture in the Lower Mississippi Valley. Southeast-
plants other than maize. Additionally, the pathology load ern Arch 12:1–14.
during the Coles Creek period is slightly heavier than Giardino MJ. 1977. An osteological analysis of the human popu-
that of the Pre-Coles Creek period; however, the differ- lation from the Mount Nebo Site, Madison Parish, Louisiana.

American Journal of Physical Anthropology


BIOARCHAEOLOGICAL ANALYSIS OF COLES CREEK DIET 39
Unpublished MA thesis. Department of Anthropology. New Neuman RW. 1984. An introduction to Louisiana archaeology.
Orleans: Tulane University. Baton Rouge: Louisiana State University Press.
Giardino MJ. 1982. Temporal frameworks. Archaeological com- Norr L. 1984. Prehistoric subsistence and health status of
ponents and burial styles. The human osteology of the Mount coastal peoples from the Panamanian isthmus of lower
Nebo site in north Louisiana. Louisiana Arch 9:99–126. Central America. In: Cohen MN, Armelagos GH, editors.
Gremillion KJ. 2002. The development and dispersal of agricul- Paleopathology at the origins of agriculture. Orlando: Aca-
tural systems in the Woodland Period southeast. In: Anderson demic Press. p 463–490.
DG, Mainfort RC Jr, editors. The Woodland southeast. Tusca- Ortner D. 2003. Identification of pathological conditions in
loosa: University of Alabama Press. p 483–501. human skeletal remains, 2nd ed. San Diego: Academic Press.
Haag WG. 1978. A prehistory of the Lower Mississippi River Perzigian AJ, Tench PA, Braun DJ. 1984. Prehistoric health in
Valley. Geosci Man 9:1–8. the Ohio River Valley. In: Cohen MN, Armelagos GH, editors.
Jackson HE, Scott SL. 2002. Woodland faunal exploitation in Paleopathology at the origins of agriculture. Orlando: Aca-
the midsouth. In: Anderson DG, Mainfort RC Jr, editors. The demic Press. p 347–366.
Woodland southeast. Tuscaloosa: University of Alabama Powell ML. 1985. The analysis of dental wear and caries for
Press. p 461–482. dietary reconstruction. In: Gilbert RI Jr, Mielke JH, editors.
Jeter MD, Rose JC, Williams GI Jr, Harmon AM. 1989. Adapta- The analysis of prehistoric diets. Orlando: Academic Press.
tion types. In: Jeter MD, Rose JC, Williams GI Jr, Harmon AM, p 307–338.
editors. Archaeology and bioarchaeology of the Lower Missis- Powell ML. 1991. Ranked status and health in the Mississip-
sippi Valley and trans-Mississippi south in Arkansas and Loui- pian chiefdom at Moundville. In: Powell ML, Bridges PS,
siana. Fayetteville: Arkansas Archaeological Survey. p 355–378. Wagner-Mires AM, editors. What mean these bones? Studies
Jeter MD, Williams GI Jr. 1989. Ceramic-using cultures, 600 B.C. to in southeastern bioarchaeology. Tuscaloosa: University of Ala-
A.D. 700. In: Jeter MD, Rose JC, Williams, GI Jr, Harmon AM, bama Press. p 22–51.
editors. Archaeology and bioarchaeology of the lower Mississippi Robbins LM. 1976. Analysis of human skeletal material from
Valley and trans-Mississippi south in Arkansas and Louisiana. Morton Shell Mound (16IB3), Iberia Parish, Louisiana.
Fayetteville: Arkansas Archaeological Survey. p 111–170. Unpublished manuscript on file, Department of Geography
Kelley DB. 1992. Coles Creek period faunal exploitation in the and Anthropology. Baton Rouge: Louisiana State University.
Ouachita River Valley of southern Arkansas. Midcontinental Roberts KM. 2000. Plant remains. In: Ryan J, editor. Data re-
J Arch 17:227–264. covery excavations at the Hedgeland Site (16CT19), Cata-
Kidder TR. 1992a. Timing and consequence of the introduction houla Parish, Louisiana. Report submitted to the US Army
of maize agriculture in the Lower Mississippi Valley. North Corps of Engineers. Vicksburg District. Baton Rouge: Coastal
Am Arch 13:15–41. Environments. p 10.1–10.33.
Kidder TR. 1992b. Coles Creek period social organization and Rose JC, Burnett BA, Nassaney MS, Blaeuer MW. 1984. Paleo-
evolution in northeast Louisiana. In: Barker AW, Pauketat pathology and the origins of maize agriculture in the Lower
TR, editors. Lords of the southeast. Social inequality and the Mississippi Valley and Caddoan culture area. In: Cohen MN,
native elites of southeastern North America. Archaeological Armelagos GH, editors. Paleopathology at the origins of agri-
Papers of the American Anthropological Association. Number culture. Orlando: Academic Press. p 393–424.
3. Washington, DC. p 145–162. Rose JC, Harmon AM. 1999. Louisiana and south and eastern
Kidder TR. 2002. Woodland period archaeology of the Lower Arkansas. In: Rose JC, editor. Bioarchaeology of the south
Mississippi Valley. In: Anderson DG, Mainfort RC Jr, editors. central United States. Arkansas Archaeological Survey
The Woodland southeast. Tuscaloosa: University of Alabama Research Series No. 55. Fayetteville: Arkansas Archaeological
Press. p 66–90. Survey. p 35–82.
Kidder TR, Fritz GJ. 1993. Subsistence and social change in the Rose JC, Marks MK, Tieszen LL. 1991. Bioarchaeology and sub-
Lower Mississippi Valley. The Reno Brake and Osceola Sites, sistence in the central and lower portions of the Mississippi
Louisiana. J Field Arch 20:281–297. Valley. In: Powell ML, Bridges PS, Wagner-Mires AM, editors.
Larsen CS. 1984. Health and disease in prehistoric Georgia. What mean these bones? Studies in southeastern bioarchaeo-
The transition to agriculture. In: Cohen MN, Armelagos GJ, logy. Tuscaloosa: University of Alabama Press. p 7–21.
editors. Paleopathology at the origins of agriculture. Orlando: Scarry CM. 1993. Variability in Mississippian crop production
Academic Press. p 367–393. strategies. In: Scarry CM, editor. Foraging and farming in
Larsen CS. 1995. Biological changes in human populations with the eastern Woodlands. Gainesville: University Press of
agriculture. Ann Rev Anthropol 24:185–213. Florida. p 78–90.
Larsen CS. 1997. Bioarchaeology. Interpreting behavior from Schmucker BJ. 1985. Dental attrition. A correlative study of
the human skeleton. Cambridge: Cambridge University Press. dietary and subsistence patterns in California and New
Larsen CS, Shavit R, Griffin MC. 1991. Dental caries evidence Mexico Indians. In: Merbs CF, Miller RJ, editors. Health and
for dietary change. An archaeological context. In: Kelley MA, disease in the prehistoric southwest. Tempe: Arizona State
Larsen CS, editors. Advances in dental anthropology. New University. p 253–273.
York: Wiley-Liss. p 179–202. Schoeninger MJ, Schurr MR. 1994. Interpreting carbon stable
Lewis BA. 1991. Analysis of pathologies present in the 16ST1 isotope ratios. In: Sobolik K, editor. Paleonutrition. The diet
Tchefuncte Indian skeletal collection. Unpublished MA thesis, and health of prehistoric Americans. Center for Archaeologi-
Department of Geography and Anthropology. Baton Rouge: cal Investigations. Occasional paper No. 22. Carbondale:
Louisiana State University. Southern Illinois University. p 55–66.
Lukacs JR. 1989. Dental paleopathology. Methods for recon- Scott EC. 1979. Dental wear scoring technique. Am J Phys
_can MY, Kennedy KAR, edi-
structing dietary patterns. In: Is Anthropol 51:213–217.
tors. Reconstruction of life from the skeleton. New York: Smith BD. 1986. The archaeology of the southeastern United
Wiley-Liss. p 261–286. States: From Dalton to de Soto, 10,500-500 B.P. In: Wendorf
Mariaca MT. 1988. Late Marksville/ Early Baytown period sub- F, Close A, editors. Advances in world prehistory, Vol.5.
sistence economy. Analysis of three faunal assemblages from Orlando: Academic Press. p 1–92.
northeastern Louisiana. Unpublished MA thesis, Department Smith BH. 1984. Patterns of molar wear in hunter-gatherers
of Anthropology. Boston: Boston University. and agriculturalists. Am J Phys Anthropol 63:39–56.
McGimsey C. 2003. The Morton Shell Mound project and other Smith RL. 1996.Vertebrate subsistence in southeastern Louisi-
stories of southwest Louisiana history. Regional Archaeology ana between A.D. 700 and 1500. Unpublished MA thesis,
Program. Management Unit III. 2002/2003 Annual Report. Department of Anthropology. Athens: University of Georgia.
Moore WJ, Corbett ME. 1971. The distribution of dental caries Sobolik KD. 1994. Introduction. In: Sobolik KD, editor. Paleonu-
in ancient British populations. I. Anglo-Saxon period. Caries trition. The diet and health of prehistoric Americans.
Res 5:151–168. Carbondale: Southern Illinois University. p 1–20.

American Journal of Physical Anthropology


40 G.A. LISTI
Springer JW. 1980. An analysis of prehistoric food remains from Turner CG. 1979. Dental anthropological indications of agricul-
the Bruly St. Martin Site, Louisiana, with a comparative dis- ture among the Jomon people of central Japan. Am J Phys
cussion of Mississippi Valley faunal studies. Mid-Continental Anthropol 51:619–636.
J Arch 5:193–223. Williams S, Brain JP. 1983. Excavations at the Lake George
Toth EA. 1979. The Lake St. Agnes site. A multi-component Site Yazoo County, Mississippi, 1958–1960. Papers of the Pea-
occupation of Avoyelles Parish, Louisiana. Louisiana State body Museum, No. 74. Cambridge: Harvard University Press.
University Museum of Geoscience. Mèlanges No. 13. Baton Wood JW, Milner GR, Harpending HC, Weiss KM. 1992. The
Rouge. osteological paradox. Cur Anthropol 33:343–370.

American Journal of Physical Anthropology

You might also like