ELEMENTAL ANALYSIS OF MARKSVILLE-STYLE PREHISTORIC
CERAMICS FROM MISSISSIPPI AND ALABAMA
By
Keith Allen Baca
A Thesis Submitted to the Faculty of Mississippi State University In Partial Fulfillment of the Requirements for the Degree of Master of Arts in Applied Anthropology in the Department of Sociology, Anthropology, and Social Work
Mississippi State, Mississippi
May 3, 2008
ELEMENTAL ANALYSIS OF MARKSVILLE-STYLE PREHISTORIC
CERAMICS FROM MISSISSIPPI AND ALABAMA
By
Keith Allen Baca
Approved:
Janet Rafferty Evan Peacock Graduate Coordinator of Anthropology Associate Professor of Anthropology Professor of Anthropology (Director of Thesis)
S. Homes Hogue Gary Myers Professor of Anthropology Interim Dean of College of Arts and Sciences
Name: Keith Allen Baca
Date of Degree: May 3, 2008
Institution: Mississippi State University
Major Field: Applied Anthropology
Major Professor: Dr. Janet Rafferty
Title of Study: ELEMENTAL ANALYSIS OF MARKSVILLE-STYLE PREHISTORIC CERAMICS FROM MISSISSIPPI AND ALABAMA
Pages in Study: 129
Candidate for Degree of Master of Arts
Distinctive Marksville-style pottery is characteristic of the Middle Woodland period (200 B.C. – A.D. 500) in the Lower Mississippi River Valley and adjacent regions. Marksville material is common in the Lower Mississippi Valley, and the scarcity of similar pottery in northeastern Mississippi and western Alabama has caused claims that Marksville pots were imported into those areas; however, they may have been locally made. To test these alternative possibilities, the elemental composition of some Marksville-style potsherds, other pottery, and clays from various archaeological sites spanning the above regions was characterized using laser ablation-inductively coupled plasma-mass spectrometry. The results show that the analyzed Marksville-style pottery shares similar elemental profiles with locally common wares and local clays in the sample, allowing the conclusion that all of these
Marksville specimens were made in the regions where they were found.
ACKNOWLEDGMENTS
Numerous persons generously assisted me in many ways during the process of
producing this thesis. Dr. Janet Rafferty, who served as my major professor and
director of this thesis, and my other thesis committee members Dr. Evan Peacock and
Dr. S. Homes Hogue, patiently guided me through every stage of the project.
I am indebted to Dr. Yunju Xia, Mr. Dean Patterson, and Dr. Ron Palmer of
the Institute for Clean Energy Technology at Mississippi State University for
contributing so much of their valuable time and effort in the processing of my
samples with the Institute’s Laser Ablation-Inductively Coupled Plasma-Mass
Spectrometer. In addition, my fellow graduate student Bradley Carlock assisted in
the laboratory work. Dylan Karges of the Cobb Institute of Archaeology, Mississippi
State University, prepared the clay samples for analysis.
The following people were instrumental in providing permissions to borrow
the specimens for analysis: Dr. Evan Peacock of the Cobb Institute of Archaeology;
Joe Giliberti and Ernie Seckinger of the U.S. Army Corps of Engineers, Mobile
District; John O’Hear of the Cobb Institute; Eugene Futato of the Alabama Museum
of Natural History; Pam Edwards-Lieb of the Mississippi Department of Archives
and History; Jim Turner of the Mississippi Department of Transportation; and Dr. Ed
Jackson of the University of Southern Mississippi.
ii Joe Greenleaf, another fellow graduate student, kindly prepared the map. I single out for special thanks Jeffrey Alvey of the Cobb Institute for his efficient and timely work in formatting my thesis manuscript files into a finished document.
Above all, I wish to express my eternal gratitude to my wife, Julie, who steadfastly supported me throughout the long course of my graduate work. Her constant encouragement, perseverance, and, most importantly, her sheer belief in me, saw me through to the finish. I could not have done it without her.
iii
TABLE OF CONTENTS
ACKNOWLEDGMENTS ...... ii
LIST OF TABLES...... vi
LIST OF FIGURES ...... vii
CHAPTER
I. INTRODUCTION ...... 1
II. PROBLEM STATEMENT...... 5
III. IMPLICATIONS OF SOME PREVIOUS REGIONAL SOURCING STUDIES ...... 7
IV. THEORETICAL AND METHODOLOGICAL CONSIDERATIONS...... 12 Pottery Sourcing and the Potential of Evolutionary Archaeology..... 12 Pottery Classification...... 18
V. MIDDLE WOODLAND PERIOD CERAMICS IN MISSISSIPPI AND WESTERN ALABAMA BY REGION ...... 21 The Yazoo Basin, Western Mississippi ...... 22 North-Central Mississippi...... 25 Northeastern Mississippi and West-Central Alabama ...... 28 Southeastern Mississippi and Southwestern Alabama...... 30 Coastal Mississippi...... 30
VI. LITERATURE REVIEW OF PURPORTED NONLOCAL MARKSVILLE POTTERY IN EASTERN MISSISSIPPI AND WESTERN ALABAMA ...... 32
VII. THE STUDY SAMPLE AND SITE INFORMATION...... 40 1CK21 Porter Village, Clarke County, Alabama...... 43 1GR2 Craig’s Landing, Greene County, Alabama ...... 45 22CL527 Kellogg Village, Clay County, Mississippi ...... 47
iv 22CO502 Dickerson, Coahoma County, Mississippi ...... 48 22CO666 Dry Bayou 1, Coahoma County, Mississippi...... 50 22CO682 County Line Church, Coahoma County, Misssissippi ...... 51 22GN685, Greene County, Mississippi ...... 51 22GN687 Mossy Ridge, Greene County, Mississippi...... 52 22HR591 Godsey, Harrison County, Mississippi...... 53 22LO538 Vaughn Mound, Lowndes County, Mississippi ...... 54 “M MD 13” Holifield, Madison County, Mississippi...... 55 22NO503/534 Monk Place, Noxubee County, Mississippi...... 56 22OK888, Oktibbeha County, Mississippi ...... 58 22SH500 Spanish Fort, Sharkey County, Mississippi...... 59 22SH520 Leist, Sharkey County, Mississippi ...... 60 22UN500 Ingomar Mounds, Union County, Mississippi ...... 61 22WI516 Stinking Water, Winston County, Mississippi...... 64 22YZ515 Milner, Yazoo County, Mississippi...... 65
VIII. ANALYTICAL METHODS...... 67 Determination of Elemental Composition of Samples by LA-ICP-MS...... 67 Elemental Data Analysis...... 70
IX. RESULTS ...... 73 Correspondence Analysis...... 73 Principal Components Analysis...... 83
X. SUMMARY AND CONCLUSIONS ...... 88
REFERENCES CITED...... 93
APPENDIX
A. SAMPLE ELEMENTAL CONCENTRATION DATA...... 109
v LIST OF TABLES
TABLE Page
A1. Sample elemental concentration data…………………………………… 109
vi LIST OF FIGURES
FIGURE Page
1. Map of Mississippi and Western Alabama showing physiographic zones and locations of archaeological sites from which the analysis specimens originated...... 44
2. Correspondence analysis (CA) ordination graph of all samples, axes 1 and 2...... 74
3. Correspondence analysis (CA) ordination graph of all samples, axes 1 and 3...... 75
4. Correspondence analysis (CA) ordination graph of all samples, axes 2 and 3...... 76
5. Correspondence analysis (CA) ordination graph of sample elements, axes 1 and 2 ...... 80
6. Correspondence analysis (CA) ordination graph of sample elements, axes 1 and 3...... 81
7. Correspondence analysis (CA) ordination graph of sample elements, axes 2 and 3...... 81
8. Principal components analysis (PCA) ordination graph of all samples by sodium (Na) and potassium (K) concentration ...... 85
vii
CHAPTER I
INTRODUCTION
An archaeological research theme of longstanding interest is the identification of
regional trade or exchange of raw materials and finished artifacts between prehistoric social groups (Bishop and Canouts 1993:162; Earle 1982:2; Seeman 1979). A key goal of exchange studies is to identify the areas from which regionally rare or anomalous materials originate, a research emphasis commonly referred to as “sourcing” (Bishop and
Canouts 1993:162; Bishop et al. 1982; Earle 1982:4; Ericson and Baugh 1994:3; Ericson and Earle 1982:xv; Harbottle 1982:15-17).
Sourcing studies increasingly have involved chemical and/or petrographic materials analysis in conjunction with stylistic analysis. With respect to chemical analysis, its quantitative precision, on a scale unachievable with stylistic or petrographic techniques, is seen to have revolutionized the study of exchange by bringing objective data to bear on sourcing studies (Earle 1982:4; Bishop and Canouts 1993:160), although the high cost of the various chemical techniques used historically has been an impediment to their widespread adoption by archaeologists (Earle 1982:4).
In the archaeology of Mississippi, as in southeastern archaeology generally, studies of prehistoric pottery traditionally have been dominated by stylistic analysis to further the longtime preoccupation with the delineation of local cultural sequences.
While this culture-historical orientation is still quite prominent, a newer focus on non-
1 historical southeastern ceramic studies such as sourcing is now evident (Gibson 1993:33-
34; Lafferty 1994:7). Steponaitis et al. (1996), Lynott et al. (2000), and O’Brien et al.
(1995) note that pottery sourcing studies based on chemical characterization of ceramic fabrics, along with clays from potential source zones, have begun to appear. As reviewed by Neff (n.d.), most of these southeastern pottery sourcing studies have employed neutron activation analysis (NAA), which has for several decades constituted a predominant chemical analytical method for pottery (Rice 1987:312, 390, 396), as well as for a host of other archaeological materials such as rocks and minerals, marine and freshwater shell, amber, metals and glass (Harbottle 1982).
Overall, however, the use of chemical characterization to source suspected nonlocal pottery in the Southeast has remained sporadic (Bishop and Canouts 1993:176).
As noted by Johnson (1994:102), “‘[t]rade sherds’ have been a common topic of discussion . . . but . . . the designation has been almost exclusively in terms of style and macroscopic temper characteristics.” Gibson (1994:133) makes a similar observation regarding instances of pottery believed to have been imported into the Lower Mississippi
Valley: “[the] trade ware identifications are based strictly on typology (i.e., decoration and temper similarities), not on objective analyses, and consequently remain . . . tenuous .
. .”
A key characteristic of the Middle Woodland period (ca. 200 B.C. – A.D. 500) of the prehistoric Southeast is the widespread, if geographically uneven and usually sparse, occurrence (usually in burial mound contexts) of exotic artifacts. Such occurrences are conventionally labeled as a manifestation of the so-called “Hopewellian Interaction
Sphere,” a widely used but vaguely defined rubric (named for the Hopewell type site in
2 Ohio) referring, in part, to the interregional spread of certain distinctive kinds of artifacts
(Brose and Greber 1979; Seeman 1979). At the Bynum/22CS503 (Cotter and Corbett
1951) and Pharr/22PS500 (Bohannon 1972) sites in northeastern Mississippi, burial
mound items made of minerals not naturally occurring within the state, such as copper
earspools, copper/silver panpipe fragments, mica, galena, and greenstone platform pipes,
are easily identified as exotics; indeed, southeastern sourcing studies tended to focus on
such uncommon materials (e.g., Goad 1978; 1979; Walthall 1981).
In contrast, considerably less attention has been devoted to the source
identification of common archaeological materials like pottery, which in sheer volume
and ubiquity ranks with lithics in abundance. In northeast Mississippi, a distinctively
decorated ceramic usually named Marksville is often deemed to be imported. The
characteristic set of Marksville decorative motifs, consisting of curved, broad incised
lines (often with some of the spaces or “zones” between them filled with dentate-stamped
or rocker-stamped patterns), is easily recognized in Middle Woodland artifact
assemblages throughout the state of Mississippi and adjacent areas. Pottery with such
surface treatment, when composed of ceramic paste tempered with particles of grog
(crushed potsherds or fired clay), is often classified, following Phillips’s (1970) trinomial typology, under several varieties of the types Marksville Incised and Marksville Stamped
(e.g., Atkinson et al. 1980; Blitz and Mann 2000; Brookes 1976; Brown 1998; Jenkins
1975, 1979, 1981, 1982; Jenkins and Krause 1986; Rucker 1974; Toth 1979, 1988;
Williams and Brain 1983).
The recognition of nonlocal ceramics in archaeology traditionally has been
through stylistic analysis, in which the appearance in a given region of anomalous styles
3 more common to other regions is inferred through the “criteria of abundance” principal; i.e., the geographic zone where a particular style of artifact occurs most frequently is assumed to be its center of production (Bishop and Canouts 1993:162; Rice 1987:413;
Seeman 1979). However, in addition to merely stylistic studies, recent decades have seen the increasing use by archaeologists worldwide of chemical and petrographic materials analysis to “locate the origins of particular kinds of pottery by characterizing the composition of the ceramic and comparing it with the composition of pottery of known origin or with raw resources. These . . . sourcing studies attempt to find the geographical source of the pottery through its chemical or mineral characterization” (Rice 1987:413; see also Rapp 1985). When the characterization of ceramic assemblages of a given region is established, the recognition of intrusive imported pottery, differentiated by its own characterization “fingerprints,” should then be possible (Maher 1983:6).
Accordingly, the research reported in this thesis was undertaken in an attempt to determine whether certain so-called “Marksville” ceramics found in eastern Mississippi, claimed to be extraregional imports, can be differentiated into local and nonlocal groups by variant elemental concentrations. The analysis employs Laser Ablation-Inductively
Coupled Plasma-Mass Spectrometry, a technique heretofore not attempted with archaeological materials from the region.
4
CHAPTER II
PROBLEM STATEMENT
A number of archaeologists have asserted that certain Middle Woodland period
(200 B.C. – A.D. 500) ceramic specimens found in northeastern and east-central
Mississippi, which they have classified using stylistic types of the Marksville series, were imported from the Lower Mississippi Valley (LMV), the region where Marksville pottery is most abundant. This widespread belief is the result of the extension of the Marksville typological concept, designed for the LMV, to subsume all northeast Mississippi ceramics exhibiting grog/clay tempered paste with broad-line incised surface decoration.
As a result, a perception has arisen that all such material found in northeast Mississippi was imported from the LMV (see Galaty n.d.:388). The fact that grog-tempered broad- line incised sherds are always a rarity in northeast Mississippi assemblages has reinforced the view that the material is of extraregional origin. However, this presumption obscures the possible, even likely, occurrence in northeast Mississippi of locally produced
“Marksville” pottery.
This thesis tests the often assumed, but so far unproven identification of
Marksville pottery in northeast Mississippi as extraregional imports through the quantitative analysis of elemental composition of samples of curated Marksville series pottery from various archaeological sites across Mississippi and adjacent regions.
Control analysis was done on samples of assumed local pottery, and on fired clay (also
5 safely assumed to be of nearby origin), from the same sites. (If fired clay is unavailable in a site collection, clay samples suitable for ceramics were collected from at or near each site.) The chemical analysis sought to detect possible elemental concentration similarities between common Middle Woodland ceramics and potential ceramic clays from a cross- section of several physiographic provinces. If this could be demonstrated, the various
Marksville pottery specimens from sites in the northeast and east-central parts of
Mississippi can then be identified as locally produced or extraregional based on, respectively, their chemical similarity or dissimilarity to common local pottery and local clay deposits.
6
CHAPTER III
IMPLICATIONS OF SOME PREVIOUS REGIONAL SOURCING STUDIES
Steponaitis et al. (1996:555) note that while pottery is sometimes identified as
nonlocal on the basis of stylistic comparisons, this method is unreliable because it is often
impossible to rule out the chance that a particular specimen thought to be an import is
actually a locally made item incorporating a foreign style. Hence, these authors affirm
that stylistic comparisons should also be complemented by chemical or mineralogical
compositional analysis in an effort to tie the raw material from which the artifact is made
to its geological source area (Steponaitis et al. 1996:555). The fundamental concept
supporting this raw materials characterization approach to artifact sourcing is termed the
“provenience postulate,” which holds that “differences between distinct sources of raw
materials can be recognized analytically and that compositional variations will be greater
between sources than within sources” (Rice 1987:413-414; see also Bishop, Rands, and
Holley 1982:301); i.e., it is assumed that quantitatively homogeneous compositional
groups represent geographically limited source zones (Neff 2000; Neff n.d.:333; Weigand
et al. 1977).
It is therefore apparent that any sourcing study of prehistoric pottery employing
quantitative analysis of the pottery’s physiochemical composition must be done in
conjunction with parallel analysis of clays, pottery’s chief raw material. Archaeological ceramic studies traditionally have focused on aplastic tempering materials added by
7 potters during manufacture. Clays, which form the bulk of any ceramic vessel, have been comparatively neglected due to the cryptocrystalline structure of the aluminum oxide and silica molecules composing clays, a factor which precludes the usual macroscopic, and to a lesser degree, microscopic qualitative analysis most familiar to archaeologists.
However, the application of quantitative chemical characterization techniques to clay analysis is increasing. The comparative analysis of clays occurring in the same localities from whence the analyzed pottery samples were taken is a key aspect of “ceramic ecology” (Matson 1965), characterized by a contextual approach which seeks, in part, to determine how the properties of the “ceramic environment,” i.e., the natural resources available for pottery production, are reflected in the finished ceramic product. Although clays are relatively ubiquitous minerals, as a ceramic resource they are not of equal quality: regarding Mississippi clays, Logan (1907, 1908, 1909) documents wide variability in the distribution and suitability of local clay deposits for modern ceramic products such as brick and crockery. The clay mineral smectite, found in abundance in
Mississippi’s Yazoo Basin, has a high shrink/swell ratio, and therefore is subject to cracking when dry. As a ceramic clay, it is inferior to kaolinite, which is more common in Mississippi’s uplands to the east of the Yazoo Basin (Galaty n.d.:382).
Employing the ceramic ecology approach on a geographically broad scale,
Steponaitis et al. (1996) provide an informative panregional baseline study of gross
chemical composition variability trends of pottery and clays by geographic zone in the
lower Southeast. These investigators assumed from the outset, based on previous studies
of other regions of the world (e.g., Maggetti 1982; Tobia and Sayre 1974), that alluvial
clays in particular drainage basins would tend to demonstrate certain broadly shared
8 compositional similarities, and that compositional differences from clays in different
drainages should be evident, especially if the geological formations crossed by the
respective drainages are distinctive. For the Steponaitis et al. (1996) study, a sample of
Mississippian period potsherds from twenty-one different regions across the Southeast was subjected to elemental composition analysis (sherds stylistically typical of each area were selected, and were thus assumed to have been made locally). Four major ceramic compositional groups were revealed, and were seen to correspond geographically with different clay-mineral provinces. Two of these compositional groups, the Western
(Steponaitis et al. 1996:562) and Southern (Steponatis et al. 1996:565-566), represent two of the major geographic zones considered in this thesis: the Western group includes the
Mississippi Valley, one of the subdivisions of which is the Yazoo Basin, from which part of my study sample is drawn, while the Southern group includes the Tombigbee Basin in western Alabama and eastern Mississippi, the latter being where most of the other specimens in my sample originate.
Based on the generalized compositional data presented by Steponaitis et al.
(1996:565-566), Lynott et al. (2000), O’Brien et al. (1995), and Neff (n.d.:337, 346, 348,
353) note that the quaternary alluvial sediments blanketing the Mississippi Valley (which includes the Yazoo Basin) show relatively high levels of sodium. In the lower
Mississippi Valley, the clay minerals of the alluvial sediments consist predominantly of smectite and illite (Potter et al. 1975:371); therefore, the high-sodium tendency of the region’s sediments applies to these two types of clays. Outside the Mississippi Valley, both to the east and west, sediments are correspondingly low in both sodium and potassium; this trait is taken as a correlate of the high incidence in these areas of
9 kaolinite, which is rare in the Mississippi Valley. In addition to the variability in the
concentrations of these major elemental constituents, variability in trace element
concentrations can serve as an even more specific indicator of geographic origin: in
Neff’s (n.d.:348, 354) reference group of Moundville ceramics, a profile distinguished by low antimony/high thorium is seen, while material from some areas west of the
Mississippi Valley exhibit a high chromium profile (Neff n.d.:348).
Of particular relevance to this study’s hypothesized likely compositional variation between clays and Marksville ceramics found in the Yazoo Basin and eastern Mississippi
is Steponaitis et al.’s (1996:568) analysis of sherds found at Moundville, Alabama
bearing Plaquemine stylistic traits. Previously suspected as imports from the Lower
Mississippi Valley, their compositional profile confirmed that none of the sherds were
made of clays of the Moundville region; rather, they conformed to the Western
compositional group, showing that, as expected based upon the decorative elements, the
Plaquemine style sherds did indeed originate in the Mississippi alluvial valley, and
possibly also from the hills immediately east of the valley. As such, as noted by Neff
(n.d.:338), “This finding confirms the expectation that east-west movement of ceramics
between the Mississippi valley and adjacent regions should be discernible based on bulk
paste analysis.”
It can be anticipated, then, that clay and ceramic sourcing investigations cross-
cutting the Yazoo Basin and adjacent uplands to the east, as I attempt in this study, may
reveal east-to-west compositional variation in clay deposits and local ceramics made
from them due to markedly different hydrological/depositional regimes: “[M]ultiple
upstream sources contribute to the sedimentary clays of the [Mississippi] valley itself,
10 whereas upland primary clays or sedimentary clays within small tributary drainages [e.g.,
in eastern Mississippi] are derived from a much more restricted set of parent rocks” (Neff
n.d.:336). Such a pattern is apparently reflected in the results reported by Steponaitis et
al. (1996).
It obviously follows that such a raw materials or resource-based perspective requires detailed consideration of geology on both regional and local scales, including the recognition of rock and sediment deposits, hydrological conditions, and soils (Neff n.d.:357-358, 371-372; Rice 1987:314, 319). This necessitates the consultation of available published geological and soil survey bulletin maps for the areas where archaeological sites yielding ceramic samples are located, although it is recognized that such maps are at too small a scale to depict most clay deposits. The survey bulletins can, however, serve as a background guide for clay prospecting at or as near as possible to the site of interest, via surface collecting or auger sampling of sediments in road cuts and
stream banks.
11
CHAPTER IV
THEORETICAL AND METHODOLOGICAL CONSIDERATIONS
Pottery Sourcing and the Potential of Evolutionary Archaeology
The identification of nonlocal artifacts in archaeological assemblages through
material compositional analysis may be quite certain, with the specific locale from which
the exotic material originates being determined. However, even though scientific
materials analysis brought to bear on sourcing research may identify nonlocal artifacts,
the cultural mechanisms by which artifacts were moved from their place of origin to the
archaeological context of discovery, are uncertain. Terms such as “exchange,” “trade,”
and “interaction” are frequently used in sourcing studies, often without definition, and even if explanatory scenarios are attempted to account for nonlocal artifacts, they are
typically speculative and, in any case, empirically unverifiable. Gibson and Melancon
(2004:173) comment that employing such “catch-all” non-explanatory terminology to account for suspected nonlocal pottery “only beg[s] the question,” and that “the real accounting lies in the motives, politics, and social relations behind trade.” They note, however, the continuing lack of success in explicitly modeling “the roles of various actors and social networks involved in the transfers” (Gibson and Melancon 2004:189).
Dissatisfaction with the prevailing explanatory inadequacy of sourcing studies has led to advocacy of another approach, that of conducting sourcing research under the relatively new theoretical paradigm of evolutionary archaeology (Neff 1995, 1998).
12 Proponents of this paradigm seek to adapt and extend principles of Darwinian natural
selection (as applied to variable somatic features of organisms) to artifact variation,
which has been argued to be integral to the human phenotype and hence also subject to
selection, no less so than organic human remains such as bones and teeth (Lyman and
O’Brien 2002:82-89; O’Brien and Lyman 2000:v-vi, 6-8). Accordingly, it has been
asserted that:
“a focus on sourcing is warranted because it describes what is perhaps the most salient human phenotypic characteristic observable in the archaeological record: namely, the propensity to move objects through space. . . . [C]eramics and virtually all other constituents of the archaeological record have all been displaced over some distance by humans. The universality of the human propensity to move objects through space implies that it must initially have been fixed by selection and that its various manifestations must be at least partly under the control of selection. The high archaeological visibility of this propensity implies that we can study its evolution in past human populations” (Neff 1998:115).
However, the development of an approach to archaeology based on Darwinian
evolutionary theory has remained largely unrealized, due in large part to difficulty in recognizing and defining classificatory units which may be used to detect selection- driven change in phenotypic traits through time (Dunnell 1980). The key to resolving this impasse, argues Neff (1995:70), is to re-examine and evaluate “existing modes of describing archaeological remains,” in his case the characterization of chemical variability of pottery, as a way to “recognize both the historical process of diversification and the shifting opportunities that created selective pressure in the past.” Chemical compositional diversity of ceramic pastes within assemblages, if it can be linked to chemical diversity among ceramic clays from spatially distinct geological sources, can thus be conceived of as subject to selection, reflecting opportunities for economic
13 interaction in the form of pottery trade between groups in different regions. In this scenario, “compositional data record past selective pressures that arose out of variation in the opportunities for making a living through pottery production” (Neff 1995:73).
Hence, “interaction,” usually undefined in most archaeological literature, is conceived of by Neff as driven by selective pressures related to economics, which in the case of the interregional transport of pottery (or, by extension, any other manufactured good) is motivated by “opportunities for making a living” other than direct subsistence procurement, i.e., hunting/gathering or cultivation. In Darwinian terms, increasing production and export of pots (which are artifacts and therefore part of the human phenotype) are directly tied to increased biological reproductive success of humans, i.e., both of individual potters and the larger social group (Ramenofsky 1995:135-136).
Viewed this way, although the majority of pottery in any area is produced for local consumption, if extralocal demand arises for whatever reason (e.g., diversion of labor from pottery making to more economically advantageous tasks), selective pressure for the increased production of vessels from local clay sources and their subsequent export to outside groups in which a pottery shortage exists will operate (Neff 1995:107); from the perspective of Darwinian evolution, this trend confers reproductive advantage to the group producing pottery for export.
In this scenario, observed ceramic elemental compositional variability through space attributable to geographic differences in clay composition can be represented as culturally transmitted phenotypic expressions (i.e., learned behavior of clay selection and paste preparation formulae passed from potter to potter and manifested in finished pottery) subject to selection pressure (sensu Dunnell 1995:41). In a hypothetical case of
14 an assemblage with significant occurrence of Marksville-style pottery, if much of that
pottery exhibits a nonlocal configuration of chemical composition, this may be a manifestation of selective pressure for the production of Marksville pots for export from
another region. On the other hand, if Marksville-style pottery in an assemblage is
compositionally similar to sherds bearing regionally common decorative treatments, and
to local clay sediments, that pottery can reasonably be identified as locally made.
The ceramic chemical composition analysis undertaken in this study is informed
by some of the technical and methodological approaches advocated and employed by
Neff and his colleagues, as cited throughout this thesis. However, unlike in the case of
the abundant amounts of imported pottery identified and used by Neff (1995) to advance
a hypothesis of opportunistic selective pressures influencing the interregional transport of
pottery, my analyzed sample of Marksville-style sherds from sites outside the Lower
Mississippi Valley center of Marksville pottery production is small. Therefore, even if
extralocal Marksville specimens can confidently be identified, it is clear that only minor
quantities of Marksville pottery were ever brought into northeastern and east-central
Mississippi.
In view of the fact that there are never more than scant amounts, if any, of imported Marksville material in northeast Mississippi site assemblages suggests that
Neff’s (1995) hypothesized selective pressure for the large scale manufacture of pottery for export to neighboring regions is not applicable to the question of the origin of
Marksville-style ceramics in Mississippi and Alabama. However, an alternative evolutionary approach is suggested by another study by Blomster et al. (2005), which can
15 take into the account the dimension of pottery style as subject to selection, in addition to
elemental composition characteristics.
The possibility that an alternative evolutionary approach that accommodates
pottery surface decoration, i.e., stylistic attributes, into a selection-based model of artifact
production and distribution, is suggested by the concept of the adaptive value underlying
what may otherwise appear to be archaeological manifestations of wasteful behavior, e.g., long-distance export of minor amounts of elaborately decorated “nonutilitarian”
pots, such as whole vessels which are ultimately deposited in burial mounds. The extra
effort in time and energy invested in material display, in the case of this study elaborately
decorated pots, can be conceptualized as a form of “costly signaling,” defined as:
“the ability to provide an honest index (or signal) of each [animal or human] individual’s quality or motivation can frequently be in the best interests of both signaler and recipient. Often the costs of producing the signal offer the best (sometimes the only) indication of its validity: More able, more highly motivated individuals can afford to provide more expensive (more wasteful) signals” (Bird and O’Connell 2006:163).
From an economic viewpoint, the notion that the elaboration of apparently “wasteful”
material display actually confers competitive advantage to humans is not new – e.g.,
Veblen’s (1994 [1899]) “conspicuous consumption” – but display has recently been
invoked in evolutionary terms in archaeology as “costly signaling” to posit an adaptive
benefit derived from the production of Maya calendrical monuments, which are
expensive in terms of the labor investment needed for quarrying, transport, and
engraving: Neiman (1997:269) points out that such investment may seem detrimental in
that “[t]hese expenditures represent the diversion of resources that members of the Maya
elite might otherwise have invested directly in the maintenance of themselves and their
16 kin, the acquisition of mates, and the production of children.” However, there may be
adaptive value in such costs, if “waste” if signals in the form of material elaboration
benefits both sender and receiver, i.e., selective advantage accrues to the sender (the
owner of an elaborate artifact) if the signal causes the receiver (viewer) to “act in a way that benefits the sender. . . . Yet it must be in the fitness interest of the receiver to interpret the signal so as to benefit the sender. In this scenario, the investment in waste is smart advertising” (Neiman 1997:270). Mutual adaptive benefits accrue to both sender and receiver of costly signaling if, as posited in the case of Mayan monuments, competitive ability on the part of the a given monument’s sponsor is conveyed.
Competitive ability, as defined by Neiman (1997:270), means:
“the likelihood that individuals will win political contests with competitors, contests whose outcomes ultimately do determine access to mates and resources. . . . [S]ignaling competitive ability might be favored by selection on both sender and receiver: if the two parties differ significantly in competitive ability, communication allows both of them to avoid the fitness costs of a contest whose outcome is a foregone conclusion.”
Costly signaling may be manifested not only in monumental Maya stelae, but also, in certain circumstances, in less conspicuous but no less elaborately decorated artifacts like Marksville-style pottery. This is suggested by the work of Blomster et al.
(2005), which has revealed through elemental analysis the significant interregional export throughout Mexico of intricately incised Gulf Coast Olmec pottery of the types Calzadas
Carved, Xochiltepec White, and Conejo Orange-on-White, all of which exhibit broadline
incised surface decoration not unlike that of Marksville ceramics. Blomster et al.
(2005:1071) state that imported Olmec vessels were used by “elites” or “leaders” at
various regional centers. From a costly signaling viewpoint, even though it constitutes a
17 minority in most assemblages (typically 5 to 10 percent) it can be seen that the demand for imported (and locally copied) Olmec-style pottery benefited both consumers and producers, justifying the extra expense in time and effort needed to produce the decorated ware, whether imported or locally made. In some contexts, a similar situation may have existed in the case of Marksville-style pottery in northeastern Mississippi, as discussed in
Chapter X.
Pottery Classification
The classification scheme used on the pottery specimens analyzed for this study is
paradigmatic (Dunnell 2002:70-76), i.e., the ceramic types are based on nonredundant,
mutually exclusive descriptive attributes of temper and surface treatment, as advocated
by Rafferty (1986). “Temper” refers to aplastic material added to raw ceramic clay to
reduce shrinkage during the drying and firing process, thereby preventing cracking; the
specimens analyzed for this study are tempered with either grog (crushed pottery or fired
clay particles) or sand. The exterior surface of a given grog- or sand-tempered potsherd
in this study bears one of several surface treatments: broad-line incised, stamped and
broad-line incised, cordmarked, or plain. (See Chapter VII for the various type designations and abbreviation codes.)
For the purposes of this research, the advantage of a paradigmatic pottery
classification over Phillips’s (1970) more widely used type-variety system is that the
former typology is neutral with respect to implications of the geographical origins of
named specimens (Rafferty 1986:47), which is part of the intent of the type-variety
classificatory scheme (Phillips 1970:27). Phillips’s trinomial type-variety designations
18 incorporate place names borrowed from localities where the pottery assigned the name was first recognized; implicitly, these localities are in regions where pottery identified by the particular type-variety is the most common.
However, a deleterious effect of Phillips’s typology has been the frequent reification of its type-variety names (Peacock 2003:24), that is, the tendency of type designation names, once they become entrenched in the literature, to be misconceived of as “real” (emic) discovered units rather than what they actually are: arbitrary (etic) units created by the analyst to track variation through time and space (Dunnell 1986, 2002:46,
52; Lyman et al. 1997:6-8; O’Brien and Dunnell 1997:6, 1998:19; O’Brien et al.
2002:437; Phillips et al. 1951:62-63, 66). Those who make this fundamental error are often misled into adopting untested provenience assumptions influenced by the application of pre-existing nomenclature to specimens in regions other than those for which the type-variety names were originally devised. Thus, type-variety unit names are often counterproductive in that they tend to obscure variation, both spatial and chronological, that archaeologists wish to detect among artifacts (O’Brien et al.
2002:443). In recognition of this conceptual pitfall, Gibson and Melancon (2004:172) avoid the use of long-established type names on certain prehistoric wares “because those names convey the appearance of cultural relationships where none may exist. . . . Just because we call a [Lower Mississippi Valley] potsherd Wheeler Punctated and automatically assume that the pot, the potter, or the decorative idea came from the
Tombigbee Valley doesn’t make it so.”
As discussed in detail in Chapter VI below, however, some archaeologists have adopted such assumptions in their application of several varieties of Phillips’s (1970)
19 Marksville Incised and Marksville Stamped types, which were established to describe
grog tempered broad-line incised and stamped pottery in the lower Yazoo Basin of
western Mississippi, to specimens of the same tempering and surface decoration found in
northeast Mississippi. This in turn reinforced their belief, which this thesis attempts to
test, that northeast Mississippi specimens designated with various “Marksville” varieties
were imported from the Yazoo Basin or other regions where the style is much more
common, ignoring the possibility that these decorative motifs may have been copied onto locally made grog-tempered ware.
Hence, although traditional Marksville series names as applied to various
published specimens discussed below are frequently retained herein as easily
recognizable labels to facilitate communication (cf. Gibson and Melancon 2004:172),
types consisting of temper/surface treatment designations take priority and are
consistently applied to the specimens analyzed for this study. This is done to avoid
encouraging unwarranted preconceptions regarding the geographical origin of pottery specimens, assumptions that might arise if Phillips’s Marksville type and variety designations were exclusively used to identify the specimens (cf. Rafferty 1986:47-48).
For this reason, I have eschewed coining new type and variety names herein.
20
CHAPTER V
MIDDLE WOODLAND PERIOD CERAMICS IN MISSISSIPPI AND WESTERN
ALABAMA BY REGION
Before proceeding with discussion of putative imported Marksville pottery in
northeast Mississippi and western Alabama, a brief general outline of Middle Woodland
ceramic complexes in relation to their physiographic settings in Mississippi and western
Alabama is warranted.
The Woodland period of southeastern prehistory is conventionally set at about
3000 to 1000 B.P. (radiocarbon years before present) or 1200 B.C. to A.D. 1000 in calibrated calendar years, and is distinguished from the preceding Archaic period mainly by the widespread occurrence of pottery (Anderson and Mainfort 2002:1-3); indeed, artifact assemblages of the period are characterized largely on the basis of various pottery types, usually defined on the basis of highly variable combinations of surface decoration or finish and tempering material.
The Middle Woodland period (ca. 200 B.C. – A.D. 500) is customarily differentiated from the preceding Early Woodland and succeeding Late Woodland periods by the widespread construction of burial mounds, along with the occurrence in some burial mound assemblages of nonlocal grave goods, including items made from the minerals copper, mica and galena (Anderson and Mainfort 2002:9-10). Also occasionally found in burial mounds are pottery vessels with stylistic motifs of regionally infrequent
21 occurrence, and therefore often thought to be imported (including, in northeastern
Mississippi, pots with Marksville motifs, discussed below).
The Yazoo Basin, Western Mississippi
Extending about 200 miles from Memphis in the north to Vicksburg in the south, and with a maximum width of about 60 miles, the Yazoo Basin subdivision of the Lower
Mississippi Valley (LMV) covers much of the western portion of Mississippi (Kelly
1974:7; Saucier 1994:8, 26; Figure 1). The 7,600-sq.-mi. Yazoo Basin (colloquially called “the Delta”) and much of the greater LMV consist of a vast floodplain dominated by natural levee - backswamp meander belt topography created by alluvial processes characteristic of the shifting courses of the Mississippi River and the lower reaches of its tributaries during geologically recent Holocene times (Fisk 1944; Kidder 2002:66-90;
McNutt 1996:155-157; Phillips et al. 1951:5-36; Saucier 1994). Bordered on the west by the present course of the Mississippi River, the Yazoo Basin, with its nearly flat topography, contrasts dramatically with the adjacent Loess Bluffs to the east, which consist of Pleistocene eolian (wind-transported) silt deposits (Saucier 1994:22; Peacock and Fant 2002:94; Snowden and Priddy 1968); the transition is so abrupt that the steep
Loess Bluffs immediately facing the Yazoo Basin are referred to by Phillips et al.
(1951:23) as “the valley wall.” The Loess Bluff belt gradually transitions eastward to the
North Central Hills (see below), moderately dissected uplands of much older Tertiary age.
During flooding episodes in the LMV, the slightly higher land along streams, called natural levees, were built up by coarser sandy sediments, which are heaviest and
22 therefore promptly drop out of suspension as the velocity of floodwater slows as it
overflows the stream banks. Progressively finer and hence lighter particles of silts and
then clays are in turn deposited farther away from the streams at elevations a few feet
lower (Gladfelter 2001:102-105; McNutt 1996:158; Phillips et al. 1951:8-9; Saucier
1994:98-102, Toth 1988:10). The sediments occurring throughout most of the Yazoo
Basin originally were transported down the Mississippi Valley by meandering previous courses of the Mississippi and Ohio rivers from the plains and prairies of the Midwest
(Phillips et al. 1951:23); fine sandy loams and silty clay soil associations characterize natural levees, while backswamp soils of the Sharkey-Alligator association have a high clay content (McNutt 1996:158). The origin of the sediments of the Yazoo Basin, including the clays, in distant geological provinces has implications for differentiating its ceramic resources from those of the upland sediments of north central and east central
Mississippi, which were alluvially transported during much earlier Eocene time from another physiographic and geological province, the southern Appalachians (Grim
1936:45-48).
In the archaeological literature of the LMV/Yazoo Basin, the Middle Woodland period is usually referred to as the Marksville period (e.g., Kidder 2002:72-79; McNutt
1996:171; Toth 1988:9), named for the type site located at Marksville, Louisiana, from which the distinctively decorated Marksville ceramic series was initially discovered by archaeologists (Setzler 1933a, 1933b; Toth 1974, 1988:14, 184-187). The elaborate, archetypal Marksville series decorative elements, applied to grog-tempered ware, are very distinctive. Marksville Incised is characterized by unmistakable broad, curvilinear or rectilinear grooves, which are U-shaped in cross-section. Marksville Stamped and
23 Mabin Stamped combine these broad-line incisions with zones containing dentate
stamping and rocker stamping (McNutt 1996:172; Phillips 1970:110-127, 212-222, 886;
Toth 1988). Marksville Incised and Stamped sherds are reliable Marksville period
diagnostics throughout the Yazoo Basin, although both types occur less frequently in site assemblages in the northern part of the region (Phillips 1970:886). Another uniquely
Marksville-period diagnostic is the cross-hatched rim, which occurs on vessels with both incised and stamped body decoration, and sometimes on otherwise undecorated pots
(McNutt 1996:172). Other common decorated types ascribed to the Marksville/Middle
Woodland period in the Yazoo Basin are Indian Bay Stamped (which differs from
Marksville Stamped by the lack incised lines; Phillips 1970:91-93) and Twin Lakes
Punctated (Phillips 1970:166).
The Marksville Incised and Stamped surface treatments were sometime executed on Early Marksville vessels in the form of bird motifs (Kidder 2002:73; Phillips
1970:535, 886; Toth 1979, 1988), specifically, stylized hook-billed head profiles, which
are unique to the period. (See Phillips 1970:110-111, 119, 121 for photographs and
drawings of some Yazoo Basin examples; cf. a similar bird-head motif on a Marksville
Stamped vessel from the Bynum Mounds [Cotter and Corbett 1951:73] in the North
Central Hills of Mississippi, a suspected import from the LMV.)
Also occurring in Marksville period assemblages in the Yazoo Basin is grog-
tempered Mulberry Creek Cord Marked, which gradually superseded grog-tempered
Withers Fabric Marked (a type also occurring during the preceding Tchula/Early
Woodland period) as a surface finish over the course of the Middle Woodland, as well as
grog-tempered Baytown Plain (McNutt 1996:172-174). Cord marked pottery, however,
24 is also common in the subsequent Late Woodland (Baytown) period, and therefore is not as useful as a chronological marker.
Of the total of 99 ceramic and clay samples analyzed for this study, 32 samples represent six Yazoo Basin sites with Middle Woodland/Marksville components:
22CO502, 22CO666 and 22CO682 (Coahoma County); 22SH500 and 22SH520 (Sharkey
County; and 22YZ515 (Yazoo County); see site descriptions below and Figure 1.
North-Central Mississippi
Located east of the Yazoo Basin and Loess Hills, west of the Flatwoods, and north of the Jackson Prairie are the North Central Hills, a wide physiographic belt which extends through much of the northern, central, and east-central parts of the state (Keady
1962:46; Kelly 1973:7-8, 1974:5-6; Peacock 1997:238-240; Peacock and Fant 2002:92-
93). Thomas (1942:12) describes the North Central Hills as “a well-dissected sand hills upland broken by several pronounced cuestas and escarpments . . . The local relief varies from 50 to 150 feet.”
The sediments of the eastern half of the North Central Hills are composed largely of Eocene-age, non-marine, alluvially deposited sands, sandstones, silts, clays and lignites of the Wilcox Group (Bicker 1969; Booth and Schmitz 1983; Grim 1936:19, 50-
52, 116-120; Lowe 1933:32-119; Paulson 1974:9-10; Peacock 2003:5; Peacock and Fant
2002:94). The western half of the region is composed of the Claiborne Group of Eocene sands and clays (both frequently glauconitic) of marine and non-marine origin. Deposits of quartzites are also present in some areas. The Claiborne Group includes (from east to west) the Tallahatta, Kosciusko, Cook Mountain and Cockfield formations (Bicker 1969;
25 Booth and Schmitz 1983; Grim 1936:120-122, 134-136, 207-215; Paulson 1974:8-10;
Peacock and Fant 2002:94; Thomas 1942).
Microscopic analysis by Grim (1936) of detrital heavy minerals in sediment
samples from the North Central Hills led him to posit a southern Appalachian origin for
the Wilcox Group sediments; this contrasts with the largely Midwestern origin of the
recent alluvium covering much of the Yazoo Basin (Phillips et al. 1951:23). The
disparate geographic and geologic origins of the sediments blanketing the two
physiographic zones may be expected to cause chemical compositional differences in
clays and hence in ceramics from the two different regions.
The North Central Hills is the most areally extensive physiographic zone lying
between the Yazoo Basin/greater LMV region, with its abundant, locally made
Marksville-series ceramics, and northeastern Mississippi, where the rare presence of
grog-tempered, Marksville-like broad-line incised/stamped ceramics is often attributed to
importation from the LMV. As such, Middle Woodland period assemblages from the
North Central Hills would be useful for investigating the distribution trends of these ceramic modes in this large intervening area. However, unlike in the Yazoo Basin and northeastern and east-central Mississippi, until recently there has been comparatively
little archaeological investigation in the North Central Hills. This has caused problems in
the interpretation of the prehistoric ceramic traditions of the region, as discussed by Ford
(1980, 1981, 1989), Johnson (1984) and Peacock (1997:237-244).
The comparative dearth of ceramic data for the North Central Hills is reflected in the scant material from the region available for this study: I was able to secure curated
Marksville-like pottery from only one site in the North Central Hills: Stinking Water
26 (22WI515) in Noxubee County (see site and sample descriptions below). The Stinking
Water site is located in the east half of the North Central Hills, on the banks of the
Noxubee River, a major tributary of the Tombigbee River. In recent years, systematic
intensive archaeological surveys of the Ackerman Unit of Tombigbee National Forest
have documented that Woodland assemblages, including that of Stinking Water, are
affiliated with the Miller tradition of the upper and middle Tombigbee drainage (Blitz
1984; Peacock 1995a, 2003). The pottery of the Middle Woodland period Miller I and II
phases is dominated by sand-tempered pottery with plain, fabric-marked or cord-marked
surface finish (named Baldwin Plain, Saltillo Fabric Impressed, and Furrs Cordmarked,
respectively). Fabric-marking was the predominant surface finish during Miller I,
superseded by cord-marking in Miller II (Jenkins 1981, 1982; Jennings 1941:196-199;
Rafferty 2002:207; Formerly, the western boundary of the Miller tradition was thought to be coterminous with the western margin of the Flatwoods (Johnson 1988:52), the physiographic zone immediately east of the North Central Hills. The more recent survey data from the Ackerman Unit of the Tombigbee National Forest reveal Miller ceramic styles are found well within the North Central Hills (Peacock 1997:253).
Marksville-style pottery in North Central Hills site assemblages is uncommon, although some has been recovered from burial mound contexts in the northwest quarter of the region (e.g., Koehler 1966:40), which might be expected given the proximity of those sites to the Yazoo Basin. In contrast, regarding the non-mound Stinking Water site in the eastern sector of the Hills, Peacock (2003:49) notes that “the presence of
Marksville-style ceramics at the site came as a surprise,” occurring as they do in a ceramic assemblage dominated by sherds readily assignable to types defined for the
27 Miller tradition in the Tombigbee River Valley to the east. The grog-tempered broad-line incised and zone-stamped pottery from Stinking Water therefore offers a rare opportunity for elemental analysis to investigate the compositional profiles of some unusually abundant pottery bearing stylistic treatment much more typical of the Yazoo Basin/Lower
Mississippi Valley Marksville pottery than of the North Central Hills.
Northeastern Mississippi and West-Central Alabama
This portion of the study area, most of which lies within the Tombigbee River watershed (Futato 1989), is composed of several physiographic/geological belts. From west to east, the physiographic regions from which ceramic samples in this study originate are the Flatwoods and the Black Prairie.
The Flatwoods, located immediately east of the North Central Hills and west of the Black Prairie, is a narrow belt averaging about ten miles wide. The zone is aptly named, as it has relatively slight topographic relief. It is underlain mostly by clay of the
Porters Creek formation, and in some areas by clays and sands of the Clayton formation.
Both formations are subdivisions of the Paleocene-age Midway group (Bicker 1969;
Conant 1942: 12, 14, 38; Futato 1989:165; Kelley 1974:4-5; Lowe 1933:1-32; Paulson
1974:8-9).
While ceramic material from only one site in the Flatwoods is analyzed for this study, that site, the Ingomar Mounds, is very anomalous for northeastern Mississippi in that it has yielded an unusually diverse assemblage of ceramics, noteworthy for conspicuous occurrence of sherds bearing tempering material and surface decorations more characteristic of distant regions (Rafferty 1990:93, 100-101). Most of the pottery
28 found at Ingomar, however, is characterized by regionally common sand- and grog-
tempered plain and cordmarked wares, typical of the Middle and Late Woodland period
Miller II and Miller III phases.
The Black Prairie (or Black Belt) is a crescent-shaped zone that extends through
northeast Mississippi and west-central Alabama. With a maximum width of about 25
miles, the Black Prairie “is a district of subdued topography . . . [t]he surface ranges from
nearly level plains to low broadly rounded hills” (Stephenson and Monroe 1940:29).
The fertile, characteristically black clayey soils of the Black Prairie are formed on chalks of the Mooreville and Demopolis members of the upper Cretaceous-age Selma group
(Bicker 1969; Weaver and Doster 1982:19).
Chalks are composed of calcium carbonate; hence, the calcareous nature of Black
Prairie sediments has pottery sourcing implications, in that pottery made within the Black
Prairie might be expected to exhibit relatively high concentrations of calcium due to the occurrence of calcareous clays. In contrast, ceramics and clays from the other physiographic regions within the study zone, which have non-chalk geology, may show correspondingly lower calcium concentrations.
It is from sites in the Black Prairie that most of the published Marksville-style ceramic specimens said to be extraregional imports were found (see Chapter VI); those subjected to elemental analysis for this thesis are from Kellogg (22CL527) and Vaughn
(22LO538) sites (see Chapter VII). Other analyzed specimens from Black Prairie sites include those from sites 22OK888 and 22LO503 in east-central Mississippi and 1GR2 in western Alabama (see Chapter VII).
29 Southeastern Mississippi and Southwestern Alabama
Covering much of southeastern Mississippi and adjacent areas of western
Alabama, this region is dominated by rolling hills once dominated by longleaf pines,
which have been largely replaced with more commercially desirable pine species. The
geological composition of the region is characterized by Miocene-age sands and clays of
the Hattiesburg and Pascagoula groups, and sandstone of the Catahoula formation. These
deposits are overlain in many areas by materials of the Citronelle formation (Kelley
1974:7).
In contrast to the other regions considered herein, the prehistory of southeastern
Mississippi was largely unknown until quite recently. Grog tempered broadline-incised
and stamped/broadline-incised pottery, often classified under the Marksville rubric,
seems to be rather common in the Middle Woodland ceramic complex of the area, at least
in comparison with northeastern Mississippi (Jackson et al. 2002). For this study,
specimens were chosen from collections representing three sites in the Pine Hills:
22GN685 and 22GN687 in Mississippi and 1CK21 in Alabama (see Chapter VI).
Coastal Mississippi
This area extends from the shoreline of the Mississippi Sound north to the
Southern Pine Hills a few dozen kilometers inland. “The region’s surface is relatively flat with a very slight upslope toward the interior. The surficial materials are largely composed of recently deposited fine sands and silts” (Kelley 1974:7).
30 Marksville-like pottery turns up with some regularity in site assemblages of the region (Blitz and Mann 2000). Ceramic specimens from one site in this region
(22HR591) were analyzed for this study (see Chapter VI).
31
CHAPTER VI
LITERATURE REVIEW OF PURPORTED NONLOCAL MARKSVILLE
POTTERY IN EASTERN MISSISSIPPI AND WESTERN ALABAMA
In the archaeology of northeast Mississippi and adjacent areas of Alabama, the
identification of putative nonlocal pottery has been done in an ad-hoc manner, mainly
through the qualitative description of decorative motifs, and, to a much lesser extent, of
paste texture. Pottery with regionally unusual decorative styles and relatively fine paste texture (as opposed to the generally sandy texture of local ware) is assumed to be imported. The stylistic/macroscopic paste analysis approach to ceramic sourcing has prevailed, as exemplified in the work of Jenkins (1975, 1979, 1981, 1982) and Jenkins and Krause (1986). In these reports, Jenkins, using Phillips’s (1970) type-variety system, classifies some Middle Woodland period (200 B.C. – A.D. 500) pottery specimens found in the region using stylistic types of the Marksville series. Another investigator notes that such “Marksville-type” incised and stamped, grog-tempered material stands out in its scarcity in ceramic assemblages of the Middle Woodland period in northeastern
Mississippi, which are dominated by the sand-tempered types Baldwin Plain, Saltillo
Fabric Impressed and Furrs Cordmarked (Rucker 1974:24). As such, in an evocation of the “criteria of abundance” principal (Bishop and Canouts 1993:162), Jenkins both implies and declares outright that these specimens, of types rarely found in northeast
Mississippi, were imported from the Lower Mississippi Valley, the region where
32 ceramics of the series are most common: in discussing pottery of the Miller I phase of the
Middle Woodland period Miller culture, Jenkins (1979:172) refers to “traded types” (cf.
“foreign ceramics” [Jenkins and Krause 1986:58]), which he exemplifies by discussion of a number of specific specimens.
One of the specimens most prominently cited as being of nonlocal origin is a vessel excavated from Mound E at the Pharr site (22PS500) in Prentiss County,
Mississippi, classified by Bohannon (1972:34, 47, 102) as Marksville Incised, var.
Marksville; this vessel is deemed “an actual import” (Bohannon 1972:47) and
“‘tradeware’ indicat[ing] contacts with the Lower Mississippi Alluvial Valley . . .
Contacts with the Marksville culture included both the importation [to the Pharr site] of vessels and the borrowing of decorative techniques and vessel shape” (Bohannon
1972:85). Following Bohannon, Jenkins (1979:178, 1981:21, 1982:69) also types this pot as Marksville Incised, var. Marksville. He distinguishes this vessel from those of such sand-tempered “indigenous Miller I types” as Baldwin Plain, Saltillo Fabric
Marked, and Alligator Bayou Stamped by stating unequivocally that “[t]he Marksville
Incised, var. Marksville is a type which was produced by the Marksville Culture of the lower Mississippi Valley.”
Similarly, a partial vessel from the Bynum site (22CS503) in Chickasaw County,
Mississippi is classified as Marksville Stamped by Cotter and Corbett (1951:12, 20, 72,
73) following Ford and Willey’s (1940:65) Lower Mississippi Valley type description;
Phillips’s (1970:120-121) updated type-variety classification Marksville Stamped, var.
Marksville is applied to the Bynum specimen by Jenkins (1979:178, 180).
33 The assumption that these specimens from Bynum and Pharr are imports from the lower Mississippi Valley is reiterated by Jenkins (1979:180): “Interaction with nearby cultures is documented primarily by the nonlocal ceramics found in the mounds and villages at both Pharr and Bynum. The occurrence of Marksville Stamped var.
Marksville at Bynum, and Marksville Incised var. Marksville . . . at the Pharr site indicate contact with the Marksville Culture to the west.”
While the Marksville-type specimens from Bynum and Pharr are relatively well known due to frequent citation in archaeological literature, small amounts of additional grog-tempered broadline incised/stamped pottery, classified using various Marksville varieties and thought to be nonlocal by the report authors, have been yielded by a few additional sites farther south in the upper Tombigbee drainage, including the Vaughn
Mound site (22LO538) in Lowndes County, Mississippi; the Kellogg Village site
(22CL527) in Clay County, Mississippi; and the Craig’s Landing site (1GR2) in Greene
County, Alabama.
In the Vaughn Mound site report, a partial vessel excavated from a trash pit is assigned to a type defined by Phillips (1970:117) for the lower Yazoo Basin: Marksville
Incised, var. Yokena (Atkinson 1974:128; Rucker 1974:27). Like the Marksville pottery from Bynum and Pharr, the Vaughn specimen is deemed by the report authors to be imported from outside the survey region: “The paste of this vessel is clearly non-local, being quite distinct from any other sherds discovered within the same pit or at any other site [visited by the survey]” (Rucker 1974:27). Although the source area of the specimen is not speculated upon, the application of a Phillips (1970) type/variety implies a Lower
Yazoo Basin/Lower Mississippi Valley origin.
34 Like the Vaughn Mound site, the Kellogg Village site yielded a partial vessel also
typed as Marksville Incised, var. Yokena (Atkinson et al. 1980:132-133, 278). Also found were a dozen sherds classified generically as Marksville Incised (no varieties were assigned), plus one sherd of Marksville Stamped, var. Manny (Atkinson et al. 1980:132-
133). The Kellogg site excavation report unspecifically designates these specimens
“exotic trade materials from chronologically defined cultures in other areas” (Atkinson et al. 1980:138), although, as with the Vaughn site Marksville pottery, the report’s use of the Phillips (1970) lower Yazoo Basin typology in effect implies that the Lower
Mississippi Valley is the source region.
Other than Pharr, Bynum, Vaughn, and Kellogg, another site in the upper
Tombigbee basin – Craig’s Landing (1GR2) in Greene County, Alabama – has yielded
several Marksville pottery specimens claimed to be imported. This material (classified as
Marksville Stamped, var. Manny and Marksville Incised, var. unspecified) is termed
“trade sherds” (Jenkins 1975:181, 1982:70; Jenkins and Krause 1986:64), characterized as “evidence of interregional interaction” (Jenkins 1982:81). While the putative geographic origin of the Marksville potsherds from Craig’s Landing is not explicitly stated as with the Marksville material from Bynum and Pharr, it has been implied that the material was also produced in the Lower Mississippi Valley (Jenkins and Krause
1986:64), the core area of the Marksville culture (Toth 1979; 1988).
In addition to the Marksville-style material from the above sites, Peacock
(2003:49-50) notes that grog-tempered broadline incised/stamped sherds conforming to
Marksville stylistic criteria have been reported from a few additional widely scattered sites in Mississippi located in the upper Tombigbee drainage and in immediately adjacent
35 areas to the west. Unlike in the reports above, the reported descriptions of the Marksville
pottery from these sites do not employ variety names. The sites are Nanih Waiya
(22WI500) in the upper Pearl River drainage in Winston County (Carleton 1999:145;
Ford 1936); Ingomar Mounds (22UN500) in Union County, in the upper Little
Tallahatchie River valley (Rafferty 1990:99, 100); site 22OK888 in Oktibbeha County
(Peacock 1995:2); the Brogan site (22CL501) in the Tibbee Creek drainage in Clay
County (Baca and Peacock 1996:20); the Wildcat site (22WI557), Winston County
(Peacock 1994, 1997:251); and the Stinking Water site (22WI516) on the Noxubee River
in Winston County (Peacock 2003:26-29, 49).
Most of these reports do not make claims of extraregional origin of the
Marksville-like pottery found at these sites; however, Baca and Peacock (1996:20) do
tentatively suggest that the Marksville-style material from Brogan could be of nonlocal origin, and Rafferty (1990:100-101) unequivocally states that such material from
Ingomar is exotic, based on the unusually diverse occurrence there of additional decorative styles (as well as macroscopically distinctive paste-tempering materials) characteristic of the Lower Mississippi River Valley, the Tennessee River Valley, northern Florida, and coastal Alabama (see also Rafferty 1987:155, 157). In the case of the Stinking Water site, on the other hand, Peacock (2003:49) invokes stylistic imitation as the most likely possibility: “[The presence of] Marksville types . . . is probably a matter of simple diffusion.”
In contrast to their quite sporadic occurrence in northeast and east central
Mississippi, ceramics classified under the Marksville rubric occur with some regularity in
Porter phase assemblages (which consist predominantly of sand-tempered types) in the
36 Lower Tombigbee River and Mobile Bay areas of southwestern Alabama (Campbell
1988; Jackson et al. 2002:244; Wimberly 1960:59, 109-118; Walthall 1980:156), and in
roughly contemporaneous Middle Woodland contexts in adjacent southeastern
Mississippi (Blitz 1986:22-24; Jackson et al. 2002; Blitz and Mann 2000), although
Marksville types are still in the minority in assemblages from these more southerly regions. This factor has led Walthall (1980:156) to designate such specimens in the
Porter area as “Marksville trade vessels,” which, as with the upper Tombigbee specimens discussed above, implies that they were imported from the Marksville heartland in the
Lower Mississippi Valley. On the other hand, Blitz (1986:24) deems the occurrence of
Marksville-style ceramics in the Porter area “represent[ative of] an indigenous ceramic development from the earlier Tchefuncte ceramic series.” In any case, based on the number of archaeological reports listing them in quantity, Marksville-type ceramics in
both southwestern Alabama (Wimberly 1960) and adjoining southeastern Mississippi
(e.g., Atkinson and Elliott 1979:28, 66, 67,124; Blitz 1986; Brown et al. 1996; Fields
2005; Jackson and Fields 2000; Jackson et al. 2002:234-237, 243-244; Keith 1997;
McClung 1998:13, 2005:8; Schleidt 2002:7) definitely do occur with much greater frequency than in the upper Tombigbee drainage. Such material evidently is even more plentiful in the Mississippi coastal zone (Blitz and Mann 1993, 2000; Marshall 1982:36-
39; Williams 1987:41-43, 52), suggesting the possibility that some Marksville-style material found in the upper Tombigbee area, -- if actually imported and not made locally, as discussed below -- may have come not from the lower Mississippi Valley, but from the
Gulf Coast region.
37 The above-discussed reports that suggest or declare outright that Marksville
specimens in eastern Mississippi and western Alabama are imports may be correct.
However, archaeologists have tended to extend the Marksville classificatory unit,
designed for the Lower Mississippi Valley (Brown 1998; Ford and Willey 1940; Phillips,
Ford and Griffin 1951; Phillips 1970; Williams and Brain 1983) to subsume all northeast
Mississippi ceramics exhibiting grog/clay tempered paste with broad-line incised
decoration (Galaty n.d.:388). This apparently has given rise to a perception that such
material found in the region was generally imported from the LMV (e.g., Galaty
n.d.:388). This perception may be in error, obscuring the possible occurrence in
northeast Mississippi of locally produced grog/clay-tempered broadline incised and
stamped pottery.
That such non-imported pottery may exist is suggested by the fact that some
eastern Mississippi grog-tempered pottery bearing typical Marksville stylistic motifs is
easily distinguishable from Marksville material from the Yazoo Basin in western
Mississippi by the macroscopically noticeable coarser paste texture of the eastern
Mississippi sherds, in contrast to the generally fine texture of the Yazoo Basin material.
This spatial variation in paste texture of Middle Woodland grog-tempered broadline
incised/stamped pottery has been noted in other regions, e.g., between central Missouri
and western Illinois (O’Brien and Wood 1998:199). An early allusion to this apparent
dichotomous pattern in Mississippi is contained in a discussion of the pottery type
Marksville Stamped by Phillips, Ford and Griffin (1951:93), under the heading
“Relationships Outside Survey Area,” regarding stylistically similar but coarser-textured
pottery outside the Lower Mississippi Valley: “Directly east of the Survey Area (the
38 Lower Mississippi Valley/Yazoo Basin) in eastern Mississippi, zoned rocker-stamping on a sandy [emphasis added] paste has been picked up on sites in Oktibbeha and Lauderdale counties . . . in the drainage of the Tombigbee . . .” Griffin (1979:270) further suggests that Marksville material in eastern Mississippi may have been locally made by applying broad-line incised and stamped stylistic motifs copied from Lower Mississippi Valley
Marksville pottery in his remark that “[I]n the Middle and Upper Tombigbee, Marksville period ceramic styles [i.e., not actual imported Marksville pottery] may well have come in from the west.”
The interregional paste-texture differences of pottery classified as Marksville is a pattern readily apparent to anyone who has handled (literally – the difference is discernible to the touch) Marksville material from both the Yazoo Basin and from the upland zones of Mississippi to the east. Geographical textural variability has long been noted for various additional Mississippi ceramic types, as well (Haag 1952; Phillips
1970:54; Ford 1988); Johnson et al. (2002:68) note that “almost all the sherds from the uplands of north Mississippi contain some sand. Some contain quite a lot. Many have argued that this sand may be a result of the sandy nature of the clay in the region.”
However, Neff (n.d.:328-329) discounts textural characteristics as sufficient to differentiate ceramics by region: “[It cannot be assumed] that textural groups correspond neatly to sources; . . . ceramic texture arises out of an interaction between potters and the local [emphasis added] ceramic environment . . . and thus cannot be assumed to provide a straightforward linkage of ceramics to source raw materials.”
39
CHAPTER VII
THE STUDY SAMPLE AND SITE INFORMATION
For this study, the chemical compositional analysis of ceramics has as its ideal goal the correlation of specimens with the source regions of clays, the raw material of pottery. If raw material source regions can be identified on the basis of elemental composition variability in various clays, corresponding compositional variability in ceramic assemblages at a given archaeological site can be used to identify suspected nonlocal specimens, if not their production regions. Alternatively, if suspected imported pottery is found to be compositionally similar to local clays (in the form of archaeological specimens of fired clay/daub or clay samples taken from near archaeological sites), as well as to pottery in the site assemblage that is typical of the region and thus assumed to be of local origin, the purported “exotic” specimen(s) will therefore be revealed to be locally produced items incorporating imitated extraregional traits, whether likely to be stylistic (e.g., surface decoration) and/or functional (e.g., tempering).
Accordingly, as demonstrated in Chapter IX below, I attempt to distinguish nonlocally produced (if present) from locally-made Marksville-style pottery specimens selected from curated collections from archaeological sites discussed in the reports cited above, plus available specimens from additional sites. In particular, I subjected to chemical testing certain specimens that are explicitly asserted in reports and articles to be of extraregional origin, in an effort to determine whether they are in fact exotics. Testing
40 of the National Park Service-owned Marksville vessels from the Bynum and Pharr mound
sites is not feasible due to federal regulatory prohibitions (NAGPRA, etc.) against even
minimally destructive analysis of artifacts from Native American burial mound contexts.
However, other published Marksville specimens declared to be exotics consist of partial vessels and smaller sherds from non-mortuary contexts, and the proposed minimally destructive laser ablation analysis described below was consented to by the controlling agency, the U.S. Army Corps of Engineers, Mobile District. These specimens consist of those from Vaughn Mound/22LO538 (Atkinson 1974:128; Rucker 1974:27) and Kellogg
Village/22CL527 (Atkinson et al. 1980:132-133, 138, 278), both sites in Mississippi; and
Craig’s Landing/1GR2 in Alabama (Jenkins 1975:181, 1982:70; Jenkins and Krause
1986:64). This USACE-owned material is curated at the Cobb Insititute of Archaeology,
Mississippi State University (Vaughn and Kellogg) and at the University of Alabama’s
curation facility at Moundville; permission was granted from the USACE-Mobile District
for research access to these specimens.
As a control, grog-tempered and grog-free plain or cordmarked sherds from each
site collection (commonly found in both the lower Mississippi Valley and eastern
Mississippi and thus assumed to have been made locally) were analyzed, along with
samples of local ceramic resources (clays). If fired clay was not available in a particular
curated site collection, clay was prospected for on public road right-of-ways as near as
possible to the sites from which the ceramics were collected; USDA county soil maps and
Mississippi Geological Survey bulletins were used as guides to possible clay deposit
locations. The raw clay samples were fired to earthenware in a gas kiln for
approximately 13 hours at a temperature range between 998 degrees C (1828 degrees F)
41 to 1013 C (1855 F). This was done in an attempt to convert the soft and friable raw clay
to the approximate hardness of the prehistoric pottery and fired clay samples, hopefully to
minimize differential material extraction by laser ablation due to the obviously marked
consistency differences between ceramics and raw clay.
Ninety-nine specimens from a total of eighteen archaeological sites were analyzed
for this study. The geographic distribution of the sites, spanning the Yazoo Basin in
western Mississippi, the North Central Hills of north-central Mississippi, the central
Tombigbee River drainage of northeastern Mississippi and western Alabama, and the
Southern Pine Hills and coastal zone of southeastern Mississippi (Figure 1), was chosen
in an attempt to represent a geographic cross-section of Marksville ceramic occurrence by physiographic zone throughout Mississippi and western Alabama. Brief site descriptions are presented below (listed in alphanumeric order by site trinomial designations), along with summaries of the ceramic and fired clay specimens acquired from these sites and chosen for analysis.
Each specimen is designated by a code, the first element of which consists of a site identifier composed of two upper case letters – the official county abbreviation – followed by a numeral. The second element of each code consists of two lower case letters representing temper/surface treatment, as follows:
gi = grog tempered broadline incised gs = grog tempered stamped/broadline incised gc = grog tempered cordmarked gp = grog tempered plain si = sand tempered broadline incised ss = sand tempered stamped/broadline incised sc = sand tempered cordmarked sp = sand tempered plain
42 (The designator for non-pottery fired clay, fc, deviates from the above temper/surface
treatment paradigm.) The final element of each specimen code consists of a numeral
identifying each specimen from a given site. As an example, CK21si8 stands for
specimen number 8 from site 1CK21, sand tempered broadline incised.
1CK21 Porter Village, Clarke County, Alabama
The Porter Village, type site for the Middle Woodland period Porter phase of southwestern Alabama (Walthall 1980:155-165), is situated on a natural levee on the east bank of the Tombigbee River in southwestern Clarke County (Figure 1). Excavated under the direction of Harry A. Tourtelot in the spring of 1941, the site consisted of a midden deposit containing human burials, pits filled with domestic debris (including potsherds, animal bone, mussel shell and other material), and some freshwater gastropod shell lenses (Walthall 1980:157; Wimberly 1960:28-30).
The ceramic sherd inventory reported for the site (Wimberly 1960:195) lists types
characteristic of local Porter assemblages (Walthall 1980:157), including the sand-
tempered types Alligator Bayou Stamped, Basin Bayou Incised, Santa Rosa Punctated,
and Franklin Plain. However, Wimberly (1960:109, 114-116, 195) classified some of the
site’s grog-tempered minority types using such Lower Mississippi Valley designations as
Marksville Incised, Churupa Punctated, and Troyville Stamped.
I selected a total of nine ceramic sherds for analysis from the Porter Village site
collection housed at the Alabama Museum of Natural History’s (AMNH) archaeological
curation facility at Moundville, Alabama. The grog-tempered sherds I sampled are
classified in the AMNH catalog as Marksville Incised (n=3), Marksville Stamped (n=2),
43
Figure 1. Map of Mississippi and western Alabama showing physiographic zones and locations of archaeological sites from which the analysis specimens originated.
44 and Marksville Plain (n=2). I also picked for analysis some sand-tempered sherds more
typical of Porter assemblages, listed by the AMNH catalog as Alligator Bayou Stamped
(n=1), Basin Bayou Incised (n=1), Porter Zone Incised (n=2), and Santa Rosa Stamped
(n=1). The AMNH site collection includes no non-pottery fired clay (such amorphous, temporally non-diagnostic material was routinely discarded by excavators during the era in which the Porter Village was investigated), and it was not feasible for me to travel to the Porter site environs to take a local clay-sediment sample.
The specimen codes are CK21gi1 (AMNH No. 1940.18.2332), CK21gi2 (AMNH
No. 1940.18.2334), CK21gi3 (AMNH No. 1940.18.2330), CK21gs4 (AMNH No.
1940.18.2318), CK21gs5 (AMNH No. 1940.18.2317), CK21gp6 (AMNH No.
1940.18.2350), CK21gp7 (AMNH 1940.18.2353), CK21si8 (AMNH No. 1940.18.2301), and CK21ss9 (AMNH No. 1940.18.2208).
1GR2 Craig’s Landing, Greene County, Alabama
Located on the east bank of the Tombigbee River in western Greene County in west-central Alabama (Figure 1), this site was subjected to several episodes of excavation in the 1970s by the University of Alabama, as part of the archaeological salvage work done in the area to be inundated by the construction of the Gainesville Lock and Dam.
The site consisted of a midden consisting of “dark organic soil, dense ceramic and lithic material, and faunal remains” (Jenkins and Ensor 1981:35).
The midden contained a cultural stratum indicative of an intensive occupation of the Miller I phase (100 B.C. – A.D. 300) of the Middle Woodland period; Jenkins and
Krause (1986:64) divide the Miller I occupation at the site into the successive Pharr and
45 Craig’s Landing subphases (estimated at 100 B.C. – A.D. 1 and A.D. 100 - 300,
respectively), based on varying percentage concentrations of sand-tempered Furrs Cord
Marked pottery in the site’s ceramic assemblage. A few grog-tempered sherds from the
site are assigned the designation Marksville Stamped, var. unspecified by Jenkins and
Krause (1986:64), who state that these specimens are:
“trade sherds, . . . indicat[ing] a date between A.D. 200 and 400, based on analogous types dated in the lower Mississippi valley (Phillips 1970:111; Toth 1979:194). Thus, increased amounts of Furrs Cord Marked, from 12 to 20 percent, and cross-dating with Marksville Stamped materials both indicate that a Craig’s Landing subphase date of A.D. 100 to 300 would be reasonable (Jenkins 1982:70)” (Jenkins and Krause 1986:64).
Relevant to this study are some Craig’s Landing ceramic specimens excavated
during the 1974 field season. Although I could not locate the Marksville Stamped sherds
referred to above in the Craig’s Landing site collections curated by the Alabama Museum
of Natural History (AMNH), I did find the three grog-tempered Marksville Incised sherds from Feature 10 (Jenkins 1975:82), specimen codes GR2gi5, GR2gi6, and GR2gi7.
Using Jenkins’s (1981:98) revised type-variety ceramic typology for the Gainesville Lake area, these can be designated Marksville Incised, var. Unspecified, which is roughly contemporaneous with the above-mentioned Marksville Stamped, var. Unspecified,
assumed by Jenkins and Krause (1986:64) to be “trade sherds,” i.e., extraregional
imports. These three “Marksville” Incised sherds were subjected to LA-ICP-MS analysis
to test the idea that these broadline-incised (and broadline-incised/stamped) ceramics
might be of extraregional origin.
For control analysis, seven additional ceramic and fired clay specimens, all
presumed to be of local origin, were selected: 1 sand-tempered broadline-incised; 2 grog-
46 tempered cordmarked; 2 grog-tempered plain (one of these [GR2gp9] is a fired pottery coil piece); and 2 pieces of amorphous fired clay.
The specimen codes are GR2si1 (from Feature 1), GR2gc2 (Feature 1), GR2gp3
(Feature 1), GR2fc4 (Feature 1), GR2gi5 (Feature 10), GR2gi6 (Feature 10), GR2gi7
(Feature 10), GR2gc8 (Feature 10), GR2gp9 (Feature 10), and GR2fc10 (Feature 10).
22CL527 Kellogg Village, Clay County, Mississippi
The Kellogg Village site, located on a low knoll on the west bank of the
Tombigbee River in eastern Clay County, in the Black Prairie physiographic region of eastern Mississippi (Figure 1), was excavated in the summer of 1978 by a Mississippi
State University crew under the direction of James R. Atkinson. The project salvaged materials and information prior to the destruction of most of the site by channel cut-off construction associated with the Tennessee-Tombigbee Waterway. Measuring about 80 by 60 m, with cultural deposits reaching a maximum depth of about 90 cm, the multicomponent site yielded evidence of prehistoric occupations ranging from the Middle
Archaic to Mississippian periods (Atkinson et al. 1980:ix-x, 1-5).
A Kellogg site artifact of prime relevance to my study is a partial grog-tempered broadline-incised vessel found in Feature 44. This specimen is designated by Atkinson et al. (1980:132), using Phillips’s (1970) typology, as Marksville Incised, var. Yokena.
Atkinson et al. (1980:132) discuss this specimen’s chronological context:
“[This partial vessel] . . . and eight [Marksville Incised] variety unclassified sherds were found in Feature 44, a Late Miller II [phase] pit which yielded a radiocarbon date of A.D. 780+/- 205 [uncorrected]. Variety Yokena is diagnostic of the Issaquena phase [in the lower Yazoo Basin of western Mississippi], which has been radiocarbon dated between A.D. 435 and 795 (Greengo 1964:110). These dates indicate
47 contemporaneity of the Issaquena phase with the Late Miller II through Early Miller III periods. . . . The presence at Kellogg of two Marksville Incised body sherds in Feature 123, a Miller III pit, indicates a possible late occurrence for Marksville ceramics in the Tombigbee Valley. Although Jenkins (1979) places Marksville ceramics in a Late Miller I through Late Miller II context, seven out of eight Marksville sherds from Gainesville [Lake area] features occurred in Miller III pits (Jenkins 1979:121-123).”
As discussed in detail elsewhere herein, the Marksville Incised ceramics found at
Kellogg are designated “exotic trade materials” (Atkinson et al. 1980:138), an assertion
tested for this study by using a small portion of the Feature 44 Marksville Incised vessel
(specimen code CL527gi1 for this study) for compositional analysis. In addition, two pieces of fired clay and two grog-tempered plain ceramic sherds, which are commonly found in the region (unlike Marksville material) and therefore assumed to be of local origin, were selected for control analysis. These five specimens were taken from the
Kellogg site collection stored at the Cobb Institute of Archaeology Curation Facility,
Mississippi State University.
The specimen codes are CL527gi1 (from Feature 44), CL527gp2 (Feature 86),
CL527gp3 (Feature 44), CL527fc4 (Feature 44), and CL527fc5 (Feature 86).
22CO502 Dickerson, Coahoma County, Mississippi
This site is located on a natural levee knoll on the west bank of the Sunflower
River, in the upper Yazoo Basin physiographic region, north-central Coahoma County, northwestern Mississippi (Figure 1). Dickerson, also designated by Phillips et al.
(1951:51, 314) as site 15-N-10, is described as a “[l]arge village site with conical mound and small mounds” (Phillips et al. 1951:51) and as a “[l]arge rich village site covering 40 acreas [sic],” with the conical mound measuring 8 ft. high by 100 ft. in diameter (Phillips
48 et al. 1951:314). Based on their frequency seriation of ceramic assemblages from this site and others in the upper Sunflower River drainage area, Phillips et al. (1951: unpaginated foldout Figure 19 between p. 232 and p. 233) assign Dickerson to their chronological period F-E, or early Baytown (early Late Woodland), based on relative assemblage percentages of binomial ceramic types, chiefly Baytown Plain and Mulberry
Creek Cord Marked, and minor amounts of other types. Dickerson is now known to be multicomponent, as evidenced by more recent ceramic collections from the site denoting, especially, Marksville (Middle Woodland) and also Mississippian occupations, among others (Toth 1988:99-105; state archaeological site form on file at the Historic
Preservation Division of the Mississippi Department of Archives and History, Jackson).
The Marksville diagnostic material from the site consists of specimens of the grog-tempered broadline-incised and broadline-incised/stamped types Marksville Incised and Marksville Stamped, respectively. For compositional analysis, I selected from the general surface collection of Dickerson artifacts curated at MDAH in Jackson two grog- tempered stamped/broadline incised (Marksville Stamped) and two grog-tempered cordmarked (Mulberry Creek Cord Marked). Because no non-pottery fired clay is available in the MDAH surface collection for control analysis, on February 27, 2007 I traveled to the Dickerson site vicinity and collected a clay sediment sample from the bank of a slough, a few hundred meters southwest of the archaeological site.
The specimen codes are CO502gs1, CO502gs2, CO502gc3, CO502gc4, and
CO502fc5.
49 22CO666 Dry Bayou 1, Coahoma County, Mississippi
This site is located on a high natural levee overlooking the south bank of Dry
Bayou, in the upper Yazoo Basin physiographic region of north-central Coahoma
County, northwestern Mississippi (Figure 1). The site was recorded in 1978 as a “village site” by MDAH archaeologist John Connaway, who made a random surface collection.
No information is available for this site pertaining to horizontal extent, deposit depth, etc. on the state archaeological site form on file at the Historic Preservation Division, MDAH,
Jackson. Based on the MDAH surface collection and another collection made by amateur informant Van Burnham, Connaway identified three components: Poverty Point (terminal
Late Archaic period), Early Marksville (early Middle Woodland period), and Baytown
(early Late Woodland period).
Middle Woodland/Marksville period ceramic diagnostics listed on the MDAH site form include the grog- tempered types Marksville Stamped and Marksville Incised, several of both were selected from the surface collection at MDAH for compositional analysis. In addition, one grog-tempered plain sherd from this site collection was also selected for analysis. No fired clay was available in this collection, so on February 27,
2007 I traveled to the vicinity of the site to take a clay sediment sample. The site and its immediate environs were inaccessible, so the sample was taken from the nearest suitable roadside location, on the west bank of a slough located about 1 km northeast of the site.
The specimen codes are CO666gs1, CO666gi2, CO666gi3, CO666gi4,
CO666gs5, CO666gi6, CO666gi7, CO666gp8, and CO666fc9.
50 22CO682 County Line Church, Coahoma County, Mississippi
Located on a natural levee on the west bank of Hopson Bayou, in the upper Yazoo
Basin physiographic region, southeastern Coahoma County, northwestern Mississippi
(Figure 1), this site was reported to archaeologist John Connaway of MDAH by a local informant. The MDAH archaeological site form filled out by Connaway in 1983 indicates a site surface area of “5 ac[res],” but no deposit depth is given. Artifact density is designated as “heavy.” Late Archaic, Early Woodland, and Early Mississippian occupations are checked off on the site form, but no diagnostic artifacts are listed.
The general surface collection of ceramic sherds from the County Line Church site curated at MDAH in Jackson contains 3 specimens of grog-tempered stamped/broadline-incised (Marksville Stamped); these were selected for compositional analysis, as were two grog-tempered Mulberry Creek Cord Marked sherds. On February
27, 2007 I traveled as near to the site as possible for a clay sediment sample, which was taken from a field on the west side of a road at a point located about .5 km south of the site.
The specimen codes are CO682gs1, CO682gs2, CO682gs3, CO682gc4,
CO682gc5, and CO682fc6.
22GN685, Greene County, Mississippi
This site is located in southeastern Greene County in the Southern Pine Belt physiographic zone, southeastern Mississippi (Figure 1), on an upland ridge overlooking an intermittent tributary of Maples Creek. The site was initially recorded in 1999 by
Kevin Bruce of the Mississippi Department of Transportation (MDOT), followed in
51 subsequent years by two rounds of excavations (Fields 2005). Measuring a few dozen
meters in diameter, this site had intact cultural deposits beneath a disturbed upper stratum, and numerous prehistoric ceramic sherds, lithics and charred floral remains were recovered. Occupations of the Middle Archaic, Middle Woodland and Late Woodland periods were identified; Middle Woodland ceramics consisted of one grog-tempered broadline incised sherd identified by Fields (2005:270), adopting established Lower
Mississippi Valley type-variety nomenclature (Phillips 1970:117-119), as Marksville
Incised, var. Yokena. The paste of this sherd is described by Fields (2005:270) as containing “small hematite and mica inclusions.” I borrowed this specimen from MDOT for compositional analysis, plus one grog-tempered plain sherd from the site. Both items are from excavation unit 69 (MDOT catalog number 197),
22GN687 Mossy Ridge, Greene County, Mississippi
This site, similar in size, cultural deposit depth, and topographical setting to
22GN685, is also located in southeastern Greene County, southeastern Mississippi
(Figure 1). First recorded in 1999 by the Mississippi Department of Transportation, the site was subsequently excavated, as reported by Fields (2005). Chronologically diagnostic artifacts identified in the site report include projectile points and ceramics of the Late Archaic, Middle Woodland and Late Woodland periods.
Of relevance to this thesis are Middle Woodland ceramics in the site inventory: three grog-tempered broadline-incised sherds (MDOT catalog no. 1853) were chosen for analysis, all designated by Fields (2005:269, 270) as Marksville Incised var. Yokena.
Fields (2005:270) notes that unlike the Marksville Incised var. Yokena sherd from
52 22GN685, the specimens from 22GN687 do not contain hematite or mica inclusions. In
addition to the aforementioned decorated pottery, three grog-tempered plain sherds
(MDOT catalog no. 1886) and two pieces of fired clay (MDOT catalog no. 2048) were selected for analysis.
The specimen codes are GN687gi1, GN687gi2, GN687gi3, GN687gp4,
GN687gp5, GN687gp6, GN687fc7, and GN687fc8.
22HR591 Godsey, Harrison County, Mississippi
As reported by Blitz and Mann (2000:29-32), the Godsey site is a prehistoric
earth-and-shell midden located close to the coastline of southern Harrison County,
southeastern Mississippi, on a relict beach ridge 2-3 m above mean sea level, about 0.1
km north of Mississippi Sound (Figure 1). Originally at least 100 m long east-west by 20
m wide north-south, the site lies in an area of heavy urban development and as a result
has been greatly disturbed, although intact prehistoric cultural deposits reached a depth of
about 1 m in one spot. Excavation here recovered marine shells, animal bones, fired clay
daub, charcoal fragments, potsherds, and a small amount of lithic tools and debitage.
Following archaeological investigation, the remaining portion of the site was sealed
under a casino parking lot (Blitz and Mann 2000:29).
The main component identified at the site “is a late Middle Woodland period
(A.D. 200-400) occupation characterized by a homogeneous assemblage of the
Marksville (Issaquena) ceramic series” (Blitz and Mann 2000:31). Diagnostic ceramics
identified in the site report include several varieties of the grog-tempered broadline-
incised type Marksville Incised. Blitz and Mann (2000:98) report that these Godsey site
53 specimens “closely replicate” Lower Mississippi Valley styles, but that “[t]here is no evidence of . . . ceramic imports.” For this thesis, three Marksville Incised sherds (no varieties specified), one grog-tempered plain sherd, and a piece of fired clay daub were selected for analysis from the Godsey site collection curated at the Department of
Anthropology and Sociology, University of Southern Mississippi (USM), in Hattiesburg.
The specimen codes are HR591gi1 (USM No. 60/41-42), HR591gi2 (USM No.
59), HR591gi3 (USM No. 57/6), HR591gp4 (no USM catalog No.), and HR591fc5 (no
USM catalog No.).
22LO538 Vaughn Mound, Lowndes County, Mississippi
As reported by Atkinson (1974), this site is a multicomponent earth and shell midden mound, located on a floodplain about 300 m from the west (right) bank of the
Tombigbee River in the Black Prairie physiographic region, central Lowndes County, eastern Mississippi (Figure 1). The dimensions of the mound are ca. 2 to 2.5 m high by
70 to 75 m in basal diameter (Atkinson 1974:115). Test excavations by Mississippi State
University (MSU) at the site in 1973 revealed a basal stratum of several low shell mounds covering Middle Archaic period human burials. Subsequent layers of occupation midden gradually accumulated over successive millennia, as evidenced by Late Archaic projectile points, plus Woodland and Mississippian ceramics.
Of special relevance to this thesis is the partial grog-tempered broadline-incised vessel found in a trash pit, a Woodland diagnostic artifact classified by Atkinson
(1974:128, 155), using Phillips’s (1970:117-119) Lower Mississippi Valley typology, as
Marksville Incised, var. Yokena. Rucker (1974:27) states that:
54 “[this] vessel [bears] a series of four rather carelessly executed broad- trailed, erect chevron designs on the shoulder of the jar, bordered top and bottom by single, horizontal broad-incised lines. The paste of this vessel is clearly non-local, being quite distinct from any other sherds discovered within the same pit or at any other site [in the upper-central Tombigbee River Valley survey region]. It is comprised primarily of a slightly contorted, laminated clay paste with minute quantities of finely crushed freshwater shell included, probably accidentally. [No mention is made of the sparse but conspicuous small, light-colored grog tempering particles visible in the paste.] All incisions were made while the paste was leather- hard, and are approximately twice as wide as they are deep.”
This partial vessel was reassembled from several sherds in the laboratory following excavation. For the analysis reported in this thesis, one of these glued sherds
(MSU No. D-846/4; specimen code LO538gi1 for this study), seen at lower left of the photograph of the mended-together sherds in Plate 9a (Atkinson 1974:155), was detached from the vessel. For control analysis, two presumably local Middle-to-Late Woodland period grog-tempered plain sherds (labeled as “Tishomingo Plain” on the artifact storage bags) and one piece of fired clay were also selected. All four aforementioned specimens were borrowed from the Vaughn site collection at the curation facility of the Cobb
Institute of Archaeology, Mississippi State University.
The specimen codes are LO538gi1 (MSU No. D-846/4), LO538gp2 (MSU No. B-
286), LO538gp3 (MSU No. C-552), and LO538fc5 (MSU No. D-882-890).
“M MD 13” Holifield, Madison County, Mississippi
A collection of potsherds from this site was donated to the Mississippi
Department of Archives and History’s State Historical Museum in 1962 (MDAH catalog no. 62.37) by a collector/informant named Jimmie Holifield. This site has not been visited by archaeologists; the site designation “M MD 13” was probably assigned to the
55 artifact collection by archaeologist and then-chief curator of the State Historical Museum
Robert S. Neitzel, prior to the adoption of the currently used trinomial site code system, hence the nonstandard site trinomial. The site has not been recorded in the state archaeological site file inventory maintained by MDAH; the only available information about this site is written on the bag in which the artifacts are kept. According to this note, the site is located in the SE ¼ of Section 22, Township 9 North, Range 5 East, on the right (north) bank of the Pearl River; no other information is given in the note. This location is in the North Central Hills physiographic region of northeastern Madison
County, central Mississippi (Figure 1).
For the analysis reported herein, three prehistoric potsherds were selected from the Holifield site collection, which is now kept at the Historic Preservation Division of
MDAH in Jackson. The analyzed sherds consist of 1 grog-tempered stamped/broadline- incised (Marksville Stamped), 1 grog-tempered broadline-incised (Marksville Incised), and 1 grog-tempered cordmarked (Mulberry Creek Cordmarked). In addition, on
February 28, 2007, I traveled as close to the reported location of the site as I could get by public road for a clay sediment sample, which was taken at a point several km to the west of the site, on the right bank of Lake Creek (UTM coordinates: Zone 16, E 234440, N
3610420) in northeastern Madison County. This clay specimen was fired for analysis.
The specimen codes are MD13gs1, MD13gi2, MD13gc3, and MD13fc4.
22NO503/534 S. C. Monk Place, Noxubee County, Mississippi
This site is located in extreme northeastern Noxubee County, in the Black Prairie physiographic region of eastern Mississippi, near the Mississippi/Alabama state line
56 (Figure 1). It was originally recorded on June 15, 1933 by Mississippi Department of
Archives and History field archaeologist Moreau B. C. Chambers as site NX-12, described by him as a “large village site” on property then owned by S. C. Monk. The information provided by Chambers is scant: the site’s size was estimated at “100 yards” in diameter, and the location was given as Section 12, Township 16 North, Range 19
East, on the south side of Broken Pumpkin Creek at its confluence with the Tombigbee
River. This site’s state archaeological inventory code was changed from NX-12 to
22NO503 in the late 1960s by MDAH. In 1973, a Mississippi State University archaeological survey crew directed by Marc D. Rucker rediscovered the site. Unaware that Chambers had previously recorded it, Rucker obtained a new state site inventory number from MDAH (22NO534). The 1973 survey location information for 22NO534 closely matches that of Chambers’ NX-12 (22NO503), showing that the same site was independently recorded twice: Rucker (1974:49) gives the location as “T. 16N, R. 19E,
SE ¼ of NW ¼ Sec. 12. This site is situated on the right bank of the Tombigbee immediately south of its confluence with Broken Pumpkin Creek.” As the designation
22NO503 was issued prior to 22NO534, the former is used hereinafter to refer to the site.
Due to heavy overgrowth, only a few artifacts were recovered from the surface of the site by the 1973 survey. Other than lithic debitage, a few ceramic sherds were collected, but were lost before they could be classified (Rucker 1974:49). Fortunately,
Chambers’s 1933 surface collection is still available for examination: After being stored for decades at Louisiana State University in Baton Rouge (where it was assigned LSU catalog no. 858), the material is now curated at the Historic Preservation Division of
MDAH, Jackson. Consisting mostly of Middle and Late Woodland period sand- and
57 grog-tempered plain and cordmarked sherds, one grog-tempered broadline-incised sherd is present in the Chambers collection, which was selected for quantitative analysis. For control analysis, two grog-tempered cordmarked sherds from the site were also selected.
On February 10, 2007, accompanied by Janet Rafferty and Evan Peacock of the
Cobb Institute of Archaeology, Mississippi State University, I traveled to the reported location of the site to obtain a sediment sample. No prehistoric artifactual material could be seen at the location, due at least in part to the fact that the area is now covered with subdivision lots with waterfront houses and large lawns; these circumstances also made subsurface testing unfeasible. A bucket auger was used to obtain a sample of silty clay from alongside a public road, as near as possible to the center of the reported location of the site. This clay specimen was fired for analysis.
The specimen codes are NO503gi1, NO503gc2, NO503gc3, and NO503fc4.
22OK888, Oktibbeha County, Mississippi
A collector discovered a grog-tempered broadline-incised sherd (Peacock
1995b:2) on the surface of this site in 1994 and donated it to the Cobb Institute of
Archaeology, MSU. The site is located in north-central Oktibbeha County, in the Black
Prairie physiographic province of eastern Mississippi (Figure 1). (In addition to the grog- tempered broadline-incised sherd, the state archaeological site form filled out by Peacock
[1992] lists additional Middle Woodland period diagnostics, including 3 sand-tempered fabric marked, 86 sand-tempered plain, and 4 sand tempered cordmarked.) The broadline-incised sherd and the site are described by Peacock (1995b:2):
“[The donated artifact] is an elaborately incised, grog-tempered rim sherd typical of the Middle Woodland Marksville culture. . . . [The site] was
58 investigated by Evan Peacock and Janet Rafferty last year [1994] after the find was reported. Woodland artifacts were found scattered along a long ridge on the north side of Starkville, with a cluster of sherds at the point where the Marksville sherd was found. Unfortunately, the ridge had been massively impacted by clearance and logging, and the site was too disturbed to figure out what the context of the sherd had been.”
This artifact, which apparently had been found in two pieces, was glued back together.
The smaller of the two sherds (specimen code OK888gi1 for this study) was detached
along the glued joint and borrowed for the analysis reported herein. To obtain additional ceramic material, as well as a local sediment sample, The site was revisited by Rafferty and Peacock, accompanied by me, on February 10, 2007. No additional grog-tempered ceramics were found, but two sand-tempered plain sherds were recovered for analysis, as was a sample of silty clay, which was fired for analysis.
The specimen codes are OK888gi1, OK888sp2, OK888sp3, and OK888fc4.
22SH500 Spanish Fort, Sharkey County, Mississippi
This misnamed site (it is neither Spanish nor a fort) is located on the north (right) bank of the Big Sunflower River in the southern Yazoo Basin physiographic region,
western Mississippi (Figure 1). The “fort” is a much-damaged and eroded semicircular
earthen embankment (maximum height 2.5 m) with a flanking outer borrow ditch; the
earthwork is just over 1 km long and encloses an area of about 18 hectares (45 acres)
(Morgan 1999:95; Phillips 1970:305-308). The precise function of the earthwork is
unknown, although presumably it had ceremonial or other undetermined communal
significance. Moderate amounts of potsherds, mussel shells, and other occupational
debris are scattered within the earthwork, concentrated mostly in a midden in the
southeastern quarter of the enclosed area (Phillips 1970:305-308).
59 The prehistoric embankment has not been directly dated, but ceramics from the
adjacent midden are diagnostic of two successive main components (in the midden area,
at least): the Issaquena phase of the late Marksville (late Middle Woodland) period and
the Deasonville phase of the early Baytown (early Late Woodland) period (Phillips
1970:309-315). Issaquena/Marksville diagnostics found at the site include grog-tempered broadline-incised and -stamped Marksville series sherds (Phillips 1970:310, 312, 314). A small surface collection of artifacts from the site made in 1970 by MDAH (Historic
Preservation Division catalog number A79.59.12) contains two Marksville Incised sherds; these were borrowed for the analysis reported herein. A grog-tempered plain
(Baytown Plain) sherd from the site in this collection was also selected for analysis. To complete my analytical inventory from this site, on February 28, 2007, I collected a sediment sample consisting of silty clay from a field just outside the embankment to the west, on the north side of the county road bisecting the site. This clay specimen was fired for analysis.
The specimen codes are SH500gi1, SH500gi2, SH500gp3, and SH500fc4.
22SH520 Leist, Sharkey County, Mississippi
The Leist site is located in the southern Yazoo Basin physiographic province of western Mississippi (Figure 1). Like the Spanish Fort site (see above) 12 km to the north,
Leist has an enigmatic semicircular earthwork and flanking outer borrow ditch whose ends meet a river (in this case, the west/right bank of the Yazoo). This eroded embankment has a maximum height of about 2.5 m, and at about 900 m long it is slightly shorter than the one at Spanish Fort. A truncated mound (Mound B) with a modern
60 house standing on it lies within the embankment enclosure, while a large conical mound
(Mound A) and a large platform mound (Mound C) with a small secondary conical
mound on top of it are located outside the enclosure to the west (Phillips 1970:367-369).
Chronologically diagnostic decorated ceramic sherds recovered from the Leist site
(mostly from Mound C) are dominated by the Marksville series, indicating a primary
occupation in that sector of the site of the late Marksville (late Middle Woodland) period
Issaquena phase (Phillips 1970:370-372). A small collection of artifacts from Leist, made by archaeologist David T. Morgan of MDAH in 1987, consists of material which had eroded out of the west face of Mound C and fallen onto the left (east) bank of the
Little Sunflower River directly below. This collection (MDAH Historic Preservation
Division catalog number A91.75.1) includes a grog-tempered, stamped/broadline-incised
(Marksville Stamped) specimen which I selected for analysis, along with a grog-tempered
Baytown Plain sherd. On Feb. 28, 2007 and took a sediment sample (silty clay loam) from a field a few dozen meters south of Mound C. This sample was fired for analysis.
The specimen codes are SH520gs1, SH520gp2, and SH520fc3.
22UN500 Ingomar Mounds, Union County, Mississippi
The Ingomar Mounds site is located in the Flatwoods physiographic region, northeastern Mississippi (Figure 1; note that the Flatwoods zone is not shown in this figure). The site comprises a large, well-preserved rectangular platform mound about 8 m high by 70 m in diameter, the bulldozed and plowed-down remnants of several conical burial mounds, and a habitation area located southeast of the platform mound (Rafferty
1983, 1987, 1990; Thomas 1985 [1894]:267-278). Chronologically diagnostic potsherds
61 recovered during 1987 test excavations of the mounds are mainly Middle Woodland
sand-tempered plain and cordmarked, and Late Woodland grog-tempered plain and
cordmarked (Rafferty 1990). Two radiocarbon dates, A.D. 210 and A.D. 10, were
obtained from charcoal samples from an excavation unit atop the platform mound
(Mound 14) and in one of the burial mound remnants (Mound 10), respectively (Rafferty
1990:100). These absolute dates conclusively indicate that the mounds were constructed
in the Middle Woodland period, while the grog-tempered sherds found in the upper levels
of the excavations atop Mound 14 and in the habitation area show that the site was also occupied in subsequent Late Woodland times.
A few sherds recovered by Rafferty’s test excavations at Ingomar have broad- and
narrow-line incisions on both sand tempered and grog tempered ware; Rafferty
(1990:100) notes that one grog-tempered broad-line incised specimen excavated from the
ramp of Mound 14 “resembles Marksville types.” This and other incised sherds found at
Ingomar, rarely found in northeast Mississippi, are considered by Rafferty (1990:93, 100-
101) to be likely extraregional imports.
Rafferty’s excavation of the off-mound habitation area recovered mainly grog-
tempered pottery, with plain and cordmarked surfaces occurring in about equal frequency
(Rafferty 1990:95); these diagnostic traits indicate that the occupation there dates to Late
Woodland times (Rafferty 1990:98). Another collection from Ingomar, consisting of
surface material from an unspecified area of the site, was made in 1935 by Moreau B.
Chambers of the Mississippi Department of Archives and History. In 1993, Rafferty
examined the potsherds in the Chambers collection, then curated at Louisiana State
University, Baton Rouge (LSU catalog no. 903). Andras (2004:19-20), summarizing
62 Rafferty’s unpublished data, states that 272 sherds in Chambers’ collection are sand- tempered, while 123 are grog tempered. This contrasts with the ceramic inventory from the habitation area excavated by Rafferty (1990:95), in which the great majority of sherds are grog-tempered. In any case, the grog-tempered sherds in the Chambers collection do indicate a Late Woodland occupation in some sector of the site, probably the same general location as the habitation area documented by Rafferty (Andras 2004:20).
Chambers’ Ingomar site collection is now housed at the Historic Preservation
Division of the Mississippi Department of Archives and History in Jackson, where I selected a sample of eight grog-tempered potsherds from the collection for compositional analysis. Five of these sherds are broadline-incised; one of these is a square-cornered basal sherd (specimen code UN500gi4 for this analysis, see list below), which also has a regionally unusual fine paste texture much more typical of Lower Mississippi
Valley/Yazoo Basin wares. For control analysis, three presumably locally made grog- tempered plain sherds also were selected from the Chambers collection; in addition, a silty clay sediment sample from 2003 fieldwork at Ingomar was selected from a collection at the Cobb Institute of Archaeology, MSU (soil sample from shovel test unit
30N260W, dated 6-24-03, from Box 12, Bag 441). This sample was fired for analysis.
As stated above, the habitation area at the Ingomar site has been considered to be of Late Woodland age based on the preponderance of grog-tempered plain and cordmarked sherds there. However, the presence of the above-enumerated Marksville- like broadline-incised sherds in Chambers’ Ingomar collection, although a small minority, may signify an earlier Middle Woodland presence somewhere in the off-mound area, although it should be kept in mind that, depending on when the arbitrary division
63 between the two culture-historical periods is set, diminishing production of Marksville and Marksville-like ceramics probably lingered for some time into the Late Woodland period in many areas.
The specimen codes are UN500gi1, UN500gi2, UN500gi3, UN500gi4,
UN500gi5, UN500gp6, UN500gp7, UN500gp8, and UN500fc9.
22WI516 Stinking Water, Winston County, Mississippi
This site is located in central Mississippi, in the North Central Hills physiographic
region (Figure 1). The state trinomial code 22WI516 designates what is now recognized
as the southernmost portion of a very large site (collectively designated 22WI515/516)
covering some 3.5 hectares, with artifacts found to be continuously spread out along over
½ km of a first terrace overlooking the Noxubee River. 1998 excavations at the site
directed by Evan Peacock revealed numerous intact cultural features, including postmold
patterns and pits containing ceramic sherds, carbonized botanical remains, and lithic
artifacts (Peacock 2003).
Diagnostic projectile points and pottery recovered from the site indicate a long
occupational sequence ranging from the Early Archaic through Early Mississippian
periods; ceramic chronological markers range from the Gulf Formational (temporally
equivalent to the Early Woodland designation used elsewhere in eastern North America)
period through the Middle and Late Woodland periods to the Early Mississippian period
(Peacock 2003:36,44-45, 47).
Chosen for the compositional analysis reported herein were two ceramic specimens designated Vessel I (grog-tempered broadline-incised) and Vessel V (grog-
64 tempered rocker-stamped/broadline-incised), both excavated from Feature 3 (Peacock
2003:16-17). These specimens were borrowed from the Stinking Water site collection stored at the Tombigbee National Forest headquarters in Ackerman, Mississippi. The grog-tempered broadline-incised and stamped/broadline-incised pottery in the Stinking
Water assemblage is considered to be of Middle Woodland age based on their temper/stylistic combinations, which are designated in adjacent regions to the west and southeast as Marksville Incised and Marksville Stamped, common Middle Woodland types in both areas (Peacock 2003:45, 48-50).
On February 10, 2007, I obtained a sediment sample from a point somewhat south of the Stinking Water site limits, on the north bank of the Noxubee River. This silty clay was fired and analyzed as a local control specimen.
The specimen codes are WI516gi1, WI516gs2, and WI516fc3.
22YZ515 Milner, Yazoo County, Mississippi
This site is located in the southern Yazoo Basin physiographic region of western
Mississippi, on the right (west) bank of O’Neil Creek (Figure 1), about 3 km above its confluence with the Yazoo River. The site is reported by Wynn (1980:23) as a midden
“cover[ing] an area of more than two acres. The midden was very black and about two feet deep with a great deal of shell, lithic debris and Baytown [period] pottery.” Based on this surface-collected pottery, dominated by the grog tempered types Baytown Plain,
Mulberry Creek Cordmarked, var. Edwards, Larto Red, and others, Wynn (1980:23-24) assigned the site to the Deasonville phase of the early Baytown (early Late Woodland) period. However, another small surface collection curated at the Historic Preservation
65 Division of MDAH in Jackson (catalog no. A91.62) contains three probable Middle
Woodland period grog-tempered broadline-incised sherds classifiable as Marksville
Incised, var. unpsecified), which I selected for analysis. For a control sample, I also selected one grog-tempered plain sherd from this collection. In addition, on February 28,
2007 I took a sediment sample from about 1m northwest of the reported location of the
Milner site, which was as close to the site as I could get by vehicle on an unpaved farm equipment road in a cultivated field. This silty clay specimen was fired for analysis.
The specimen codes are YZ515gi1, YZ525gi2, YZ515gi3, YZ515gp4, and
YZ515fc5.
66
CHAPTER VIII
ANALYTICAL METHODS
Determination of Elemental Composition of Samples by LA-ICP-MS
Elemental composition of all ceramic and fired clay specimens (sample sizes vary
based on the quantity of suitable specimens per site collection) was quantitatively
characterized using laser-ablation inductively coupled plasma- mass spectrometry (LA-
ICP-MS). A relatively new method for quantitative compositional analysis in
archaeological sourcing studies (Herz and Garrison1998:225), the advantages of ICP-MS over the currently predominant instrumental neutron activation analysis (INAA) are evident due to several major disadvantages of the latter: INAA can be conducted at only a few institutions possessing the nuclear reactors adapted for this kind of research, and the future of the technique is uncertain due to plans for decommissioning nuclear reactors in coming years (Kennett et al. 2002:444; Rice 1987:396); and, at several hundred dollars per sample analyzed, INAA is costly (Rice 1987:396). In contrast, with ICP-MS there are no regulatory restrictions on the institutional acquisition and use of the equipment, and the cost per sample analyzed is much lower than INAA (Kennett et al. 2002:444).
The usefulness of the ICP-MS for compositional analysis to multiple disciplines (e.g., biology, chemistry, geology, as well as archaeology) has led to the acquisition of ICP-MS apparatus by research institutions nationwide, including the Institute for Clean Energy
Technology (ICET) at Mississippi State University (MSU). As a result, archaeological
67 researchers at MSU now have conveniently at hand the ability to conduct sourcing
studies with advanced materials-characterization equipment at a comparatively low cost.
Solid sample preparation for ICP-MS analysis has commonly involved chemical
digestion through the use of highly corrosive acids. The hazardous and time-consuming
nature of this digestion process has been reduced through the application of microwave-
generated heat in a closed vessel (Kennett et al. 2002:444), but a newer, acid-free and
therefore safer alternative for solid-sample preparation is available, and is used in
conjunction with the ICP-MS apparatus at MSU: vaporization with a laser beam, i.e.,
laser ablation, or LA (Kennett 2002:444-445; Speakman and Neff 2005:2).
A chief advantage of LA-ICP-MS is that a microscopic sample -- generally less than 1,000 microns across and 30 microns deep -- is removed from each specimen, making LA extraction virtually invisible macroscopically, and therefore minimally destructive (Speakman and Neff 2005:5). One practical drawback, however, is the restricted size of the laser ablation chamber, necessitating the removal of small pieces from objects too large to fit in the chamber (Speakman and Neff 2002:144). This can preclude LA-ICP-MS analysis of such artifacts as whole pots or large potsherds,
depending on whether a sample can be removed, how the sample is removed, and how
much; many individuals and agencies controlling research access to curated artifacts may
consider any visible loss of material unacceptable, even if only from an aesthetic point of
view. However, in the case of the proposed analysis of the partial vessels from the
Kellogg and Vaughn sites (see pp. 12-13 above), each of which are composed of several
mended- together (glued) sherds, one individual sherd small enough to fit within the
68 ablation chamber was detached from each vessel along existing break lines. These were restored to the vessels following analysis.
Another potential drawback of LA-ICP-MS is that, depending on the nature of the
material analyzed, the microscopic sample size might raise concern of how well laser-
ablated samples represent the composition of the entire specimen (cf. Harbottle 1982:20).
In homogeneous archaeological materials such as lithics, shell, and metals, this is not
much of an issue, but ceramics, which typically are composed of sediment particles of
widely varying size and mineral composition, are inherently heterogeneous in their paste
composition. As a result, bulk sampling traditionally has been preferred for ceramics to increase confidence in analysis. In contrast to LA-ICP-MS, which extracts samples on the order of nanograms (Hector Neff, personal communication, October 25, 2007), samples are much larger with INAA and MD-ICP-MS: typically 100 mg of powder are removed from sherds with a drill or burr for analysis (e.g., Steponaitis et al. 1996:556;
Kennett et al. 2002:445). However, concern over small-sample laser ablation is obviated by a comparative study by Larson et al. (2005), who report that elemental constituent identification of a sample of Anasazi ceramics and local clays through LA-ICP-MS essentially replicated the data previously derived using the bulk-sample INAA and MD-
ICP-MS techniques.
Regarding ceramic fabric heterogeneity, Rice (1987:415) notes that “chemical
methods are less suited to characterization of very coarse, sandy pastes”; however, the
grog tempered ceramics to be analyzed in my study, even those with relatively sandy clay
matrix as occurs naturally in northeast Mississippi, are not very coarse, hence the
69 potential unsuitability due to extreme sandiness referred to by Rice is not seen as an
insurmountable obstacle to satisfactory analysis.
Nevertheless, with the LA-ICP-MS apparatus used for the present study, the
uncertainty factor represented by ceramic paste heterogeneity was mitigated by ablation
in lines, in an attempt to crosscut compositional variability by avoiding sampling any
particular spot. Speakman and Neff (2005:9) advocate this approach: “. . . ablating along
lines and raster patterns [can] accommodate some of the variation that results from
heterogeneity in the sample matrix” (See also Hirata 1997; Perkins et al. 1997.) Other
advantages of line-ablation noted by Speakman and Neff (2005:9) are “significantly
higher count rates and better signal stability,” both deemed key factors in accurate data quantification, as well as reduction in laser-induced elemental fractionation, i.e., non- representative sampling of the laser target.
Elemental Data Analysis
As demonstrated by others (e.g., Neff n.d.:334; Steponaitis et al. 1996; Cogswell
1998), various pattern-recognition techniques for defining compositionally homogeneous groups, “which are assumed to represent geographically restricted sources or source zones [based on the ‘provenance postulate’]” (Neff n.d.:333; see also Wigand et al. 1977;
Neff 2000) can be employed to analyze ceramic elemental composition data. For this study, parts-per-million (ppm) elemental concentration data for 46 elements were acquired via LA-ICP-MS from the ceramic and sediment samples, and were recorded in a matrix format on an Excel spreadsheet, by site/specimen identity codes. These data were subjected to two kinds of analysis, performed using the software application PC-ORD,
70 version 4: Bray-Curtis correspondence analysis (CA) and, to (analyze levels of two
elements), principal components analysis (PCA).
In CA and PCA, the elemental concentration values for each sample are
represented graphically as a single point on an ordination diagram. Ordination is a means
of condensing massive multivariate data sets and simplifying them by representing each
sample visually on a graph in two or three dimensions, along coordinate axes (Ludwig
and Reynolds 1988:205; Peacock 1998:243; Pielou 1977:332; Smith and Neiman
2007:47). The variance accounted for on each axis is expressed as an eigenvalue; the
eigenvalues are derived from the data and represented on the ordination in a manner in
which the first axis represents the most variation, while the second and third axes
represent much of the remaining variation (e.g., Ortmann and Kidder 2004:198-204;
Ludwig and Reynolds 1988:205; Peacock 1998:243-244). The distribution of sample unit points on the diagram in relation to each other reflects relative compositional similarity and differences between them; i.e., the closer sample plots are together along the axes of the ordination, the more similar the samples are in their elemental concentrations; while the farther apart the sample plots in the ordination, the less similar they are in their elemental composition. It must be emphasized that graphic displays of data using CA and PCA “provide no statistical measure of the strength of correlation between assemblages or of the magnitude of change through . . . space” (Peacock
1998:237). Rather, the observer of ordinations interprets and defines patterns which may emerge (Gower 1987; Ludwig and Reynolds 1998; Peacock 1998:276).
CA is used herein due to its ability to ordinate not only the sherds, but also their constituent elements. Plotting the elements first is done to show which elements are
71 causing separation in the subsequent sherd ordinations. PCA was used only to analyze concentrations of two elements for which CA would not produce valid ordinations (see below).
In this study, patterns formed by sample ordinations may reflect spatial elemental variation of ceramic clays. Under the provenance postulate, sherd samples from a given site assemblage and local fired clay samples exhibiting relatively closely spaced points allow the interpretation that the sherds were locally produced from local clay sediments, reflecting the assumed differential composition of geological deposits by region. Sample points in the ordination outlying the main site sample grouping can be suspected to be of nonlocal geological provenance, especially if a Marksville-style sherd from a northeastern or east-central Mississippi site assemblage is represented. Conversely, should sample plots of assumed local sherds (plain or cordmarked) and local fired clay from a site be widely scattered along the ordination axes, high elemental heterogeneity in local sediment deposits is suggested. However, identification of ceramic sherd provenance by chemical comparison with local clays is complicated by the fact that “raw clays do not share the same compositional history as pottery that has experienced effects of manufacture, use, and burial” (O’Brien et al. 1995:182); cf. Neff 1998:117-118.
72
CHAPTER IX
RESULTS
The correspondence analysis was performed on the sample data at an intersite scale. Intersite ordination diagrams enable the viewer to perceive compositional similarity and differences among all samples. Ordinations of specimens claimed to be nonlocal imports may then be compared with those of locally common material, i.e., plain and cordmarked sherds. Specimens graphed as isolates from main clusters were checked against the elemental concentration spreadsheet (Appendix A1), as well as the elemental ordinations derived from correspondence analysis, in an attempt to identify which elements may account for the isolated specimen ordinations.
Correspondence Analysis
To begin, correspondence analysis was performed on the entire sample of potsherds and fired clay samples. The CA ordination diagrams (Figures 2-4) display most samples as a relatively undifferentiated mass concentrated in and around the upper right quadrants formed by all three axes. The exceptions are generally arrayed along the left and lower half of axes 1 and 2, respectively; these specimens are mostly those of sites
22GN687 in Greene County, Mississippi and 1CK2/Porter Village in southwestern
Alabama. This apparent pattern might be construed, in part, as a gross reflection of
73
CA ordination of all samples
GN685gi1 2
GN687gp5 GN685gp2 s CO502fc5 i CO502gc3 CL527f c4 CO682fc6 GN687fc8 CL52 7gi1 CO502gcCO4 502gCsL521 7f c5
LOAx 538 gp 2 CO682gc4 LO53 8g p3 CK21gi2OK888gi1CO682gc5 CL527gp3LOC53K821fMDcgi5 313MDgi113gc3 CO682gs2 GN687gp4 UN500gpMD7 13fc4UN500gi1 Y Z515f c5 OK888sp2 CO502gs2YZ515gi3 GN687gi1 GN687gi3 GN687gi2 WI516gi1OK888sp3 UN500OfcK9888fc4 CHRO5669 16g pfc4 UN500gi5 Y Z515gp 4 NO503fc4LO53 8g i1CO682gs1 WI516fc3 GCROC2gO6668p96g2gi4SsH3520gp2SH520f c3 GN687GNfc7687gp6 UN500gi3 UN500gi2 HR 59 1g i 2GR2gc8 Axis 1 UN500HRgp659 1g i 3 SH500f c4Y Z515gi2 WI516gs2MD13gi2 Y Z515gi1 SH500f c4 CK21gs5 HR 59 1f c5CKCK212s1ss9i8 SH500gi1 GR2fcC10O666gCKi7 21gs4GR2gc2 GR2gp3 CSHO566206ggsp81 CK21gp6 GRCO266si16gi6 CO666gs1 CO666gs5 SH500gi2 UN500gi4 HR59 1CgOi1666gi2 GR2gi5 NO503gc2 SH500gp3 GR2fc4 CK21gp7 CK21gi1 NO503gi1
CO666gi3
GR2gi7 CL527g p2 GR2gi6
NO503gc3
UN500gp8
Figure 2. Correspondence analysis (CA) ordination graph of all samples, axes 1 and 2.
74 5 1 s 00gi i 5 N U Ax 5 c 5f 51 Z 5 6 c c Y f 3 3 02f c 5 682 2 5 O s c O 3 i 2 C 20f 0gp3 i C 02gc 5 0 5 5 4 5g H 5 c 5g 1 O gp4 f i S 682g 682g H 51 1 7 51 C S Z c O O 4 9 1 2 Z 5 c 82gc 5 f 20gp2 s Y C C 4 Y 00g 2 5 6 R c i 5 f H CL H g O H 00f 0 1 S 666 02g 5 S C 3 0 9 5 3 H O 1 5 5 1 O 9 s S 4 C 4 1 2 c 4 C 3gc 6gp8 c SH s HR f 1 3 gi 0gs g 1 3 i 1 D c 2gp 4 66 1 2 2 f 88gi 682gs g s 682g R 5 02g O 0 2 M 6 8 3 g 5 O 1 H 5 O 1 G 666 1 1 C K gi 1 5 MD 8 S C O 00gi i 2 C 2 O O 5 g C 21 CO C 9 WI MD N K 8 2 CK 0gi s 2gc i 1 gp7 2 i U s C 5 0 s gi R 666gs 685 1 1 g 5 c 4 00 5 2 N G 2 O 3 c H 1 5 3 2gc f i 4 2 5 S C G 7 666 N R c 38f g CK CK
f 1 2
5 38gp3 5 p2 U O G 5 YZ 5 2 c p2 7
O 03gc p3 3gc 9 C f g 6 L 5 O 1 c 1 1 0 gi 3 s i Ax i gp8 i CL 5 f 888 L CK 5 38g 9 2g O g p2 0 K 5 5 s 1 O R N 00 68 0 O O 9 666 5 N 38g G N L 5 HR 666gi 5 e O N 3 i G l 888s O O C U UN5 g HR L K 1 C p 9 O 4 1 1 5 0 i c 2 axes 1 and 3. 1 i f c 1 , g i 3 7g gp3 HR f 3gi 0gp6 m s 7 0 0 7 2 0 0 i 4 2 5 2 0 5 5 g L52 5 p a 3 N O les 2 L GR g C p3 NO GR U N 5 6 C p 1 UN5 5 GR 4 s 5 00gi c 2gi l 5 f 888s 1 R YZ 2 N 2 K G gi U 5 p O i GR 21 g 666gs al 7g K 2 1 O i C g C L52 GR 6 2 1 i C 5 g of 5 3 s 1 WI g 1 2 2 h of all sam MD on i p CK gp5 t 6gs 1 5 ra I a 687 g W N n G i d 4 p 3 or g A 687 8 c 666gi N f O C G ordination C gp6 ) 687 2 4 N i 687 gi g G 0 N CA 0 G 687 7 ( c gp6 gp7 N f 7 UN5 G 21 21 8 K K 6 sis C C y GN 3 i g 687 N G ondence anal p 1 i g 687 ure 3. Corres N G g Fi
75 1 2 gi 5 is gp2 3 685 5 gp c gi 4 N x 3 G 685 21 gc 687 4 02f 6 N K A c 5 c N 4 1 C G 2 8 c O 02gc s 9 7f G 682 c 5 c 1 C f 4 5 i O 52 O 5 682f 4 88f L C c 02g C c 7g 00f f 8 O 5 C 687 1 682gs 5 1 7 00gp7 K 2 C 02gc O N p3 82gc 5 2 p2 N O 1 5 O 00f L52 3 3 2 5 gi 1 6 C 5 G N 6gi U C 5 p 8s C O c 4 1 1 2 5 O 3 H U 21 gp4 f 3gi c s 5 CL C 38g c c 888gi I 1 f S C K 8 3gc 7g 3 88 5 1 p2 i 00gi K 3 00gi D 5f 3 C K W 1 5 O c 5 2 1 687 53 D 2 O i 20f 3 51 L 02gs 5g gi O M N L52 N N 5 gi 682g 5 gi M Z 3 6 f gi 51 U LO C 5g 3 H U 4 MD G i 888s c O O 4 Y 4 2 Z c S f i 51 687 1 g K 7 p 3 f
66 C C 00
687 1 687 6 Y g Z 8gp3 c s i5 N 5 0 1 O 1
O f 9
N 4 gp6 N 5g 3 2 g 5 Y 0 G N 5 9 c 5 C 66gi gp4 3 s i Ax G 5 0 G 2 5 51 38gi 1 U 6 O 0 5 WI 2gp9 687 Z 9 687 gp2 HR L 2 682g SH O 03f 5 1 R HR 66gi O N Y 0 N 8 5 6gs 5 C L O 6 R 1 G 2 G G 5 UN5 O 3gi 1 5 C H 5 O I 1 gs 4 c 7 s H N f 2gc 00gi ss9 C D 1 1 W S 21 5 2 gs R 9 2 2 H K M 6 5 G 21 K S 1 5 C axes 2 and 3. i 8 666g 2gc 666gi K i C 0gi p6 , 1 gp6 HR 2gp3 i s O R C O 0 s 5g 1 s R 5 6gs 2 4 C 21 C G 2 2 51 666gi c H G K 00g 666gp8 1 e les Z 66 S O 5 i C l CK O Y GR g C p N O 00gi 1 03g 5 C 0 i 5 U C 1 5 9 p g c N 5 1 O f 2 U 2 N HR gp3 m GR 0 GR 20gs 4 0 5 a c 5 f H H 2 S gp7 S 1 s 21 GR gi l 1 K C 21 K h of all sam 03gi C 3 5 al p O 6gi N ra g 66 of O C 7 2 p 2gi on 7g R i 6 G i t L52 g C 2 a GR n i ordination 3 ) d 3gc 0 CA 5 ( O or N A sis y C ondence anal p 00gp8 ure 4. Corres 5 g N U Fi
76 variability in sediment geology by physiographic zone; note that these sites are located
near the southeastern geographic extremity of the sample zone (Figure 1).
On the other hand, some relatively isolated ordination plots – for example
UN500gp8 and UN500gi4 from Ingomar, CL527gp2 from Kellogg, GR2gi6 and GR2gi7
from Craig’s Landing, and NO503gc3 from the Monk site – stand out from the main
mass of samples from those sites. If these outliers represented solely grog-tempered
broadline incised sherds (UN500gi4, GR2gi6, GR2gi7), the suggestion could be raised
that they might be of nonlocal origin. Note, however, that the grog-tempered plain
specimens UN500gp8 and CL527gp2, as well as grog-tempered cordmarked NO503gc3,
are equally prominent in their isolation.
Furthermore, specific specimens named in their respective site reports using
Lower Mississippi Valley types and type/varieties and hence believed to be likely imports, i.e., Marksville Incised, var. Yokena for CL527gi1 from Kellogg (Atkinson et al.
1980:132-133, 138, 278) and LO538gi1 from Vaughn (Atkinson 1974:128; Rucker
1974:27), are graphed in general proximity to the main distributions of grog-tempered plain (e.g., CL527gp3, LO538gp3 and LO538gp2), and fired clay (e.g., CL527fc4). This compositional evidence weakens the published inferences that these particular
“Marksville Incised, var. Yokena” specimens are nonlocal, either from the Lower
Mississippi Valley or elsewhere.
Similarly, the grog-tempered broadline incised sherd from 22OK888 (specimen
OK888gi1), implied to be a possible extraregional import by Peacock’s (1995:2) description of it as “typical of the Middle Woodland Marksville culture” is shown in the intersite CA ordination diagrams (Figures 2-4) comparatively close to sand-tempered
77 plain sherds (OK888sp2, OK888sp3) and fired clay (OK888fc4) from the site, lending no support to the possibility that it is of nonlocal origin.
In addition, two grog-tempered specimens from Stinking Water, a broadline
incised and a broadline incised/rocker stamped sherd (specimensWI516gi1 and
WI516gs2, respectively) appear among the main mass of graphed specimens in the
intersite CA ordinations in Figures 2-4, weakening any notion that they could have
originated from some distant region.
Somewhat unexpectedly, the most distant and conspicuous outliers in the
ordinations in Figures 2-4 include grog-tempered plain (CL527gp2 and UN500gp8 from
Kellogg and Ingomar, respectively) and grog-tempered cordmarked (NO503gc3 from
Monk) sherds. The plain and cordmarked potsherd surface modes are ubiquitous
throughout all sample zones and hence there are no logical grounds for believing that
these particular specimens are imports. One could, however, drop any preconceptions
regarding what surface treatments are likely to be associated with local as opposed to
imported pottery and adopt the counterintuitive stance that the outlying positions of the
plain and cordmarked sherds in the ordinations might indicate nonlocal material. If so,
the CA ordinations suggest that, among the limited number of specimens selected and
analyzed for this study (n=99), the best candidate specimens for being of extraregional
origin do not bear locally rare surface decoration; instead, they are visually
indistinguishable from locally superabundant plain and cordmarked wares. Nevertheless,
the possibility that these three particular sherds could be long-distance imports is
regarded as implausible, at best.
78 As with the above-discussed material from Mississippi sites, the CA ordination
data for the specimens from the two Alabama sites in the sample, Craig’s Landing
(1GR2) and Porter Village (1CK21), also fail to strongly segregate material classed under
the Marksville series from the rest of the ceramics from both sites (cf. Figures 2-4). This
undermines published suggestions to the effect that the occurrence of Marksville material
in otherwise typical Miller II phase and Porter phase ceramic assemblages like those of
Craig’s Landing and Porter Village, respectively, represent “trade” vessels from the
center of Marksville ceramic production in the Lower Mississippi Valley (Jenkins
1975:181, 1982:70; Jenkins and Krause 1986:64; Walthall 1980:156).
Correspondence analysis of constituent elemental concentrations of the entire
sample (Figures 5-7) show that overall variability among specimens is greatest in calcium
(Ca), strontium (Sr), manganese (Mn), and sodium (Na) (Figures 5-7).
In the case of calcium, it was suggested in Chapter V that locally made sherds, fired non-pottery clay, and raw clays from sites in physiographic zones with sediments derived from chalks (in the study area, the Black Prairie of northeast Mississippi and west-central Alabama) would tend to show comparatively elevated calcium concentrations, while pottery made in regions of non-calcareous sediment geology would exhibit correspondingly low calcium levels. This expectation is partially supported by inspection of the ppm spreadsheet data (Appendix A1), which shows that the grog- tempered broadline incised, grog-tempered plain, and fired clay specimens with the lowest concentrations of calcium (a few hundred ppm) – those from 22GN687 in a non- chalk region – are among those graphed farthest (along axis 1) from the majority of
specimens. In contrast, and also as expected, sherds from Black Prairie sites, whatever
79
CA ordination of elements, all samples
2 Mg 24 s i Na 23 Zr 90
As 75 Ax Hf 18 0 Sb 121 Si-30 Sn 120 Lu 175 U 238 Al 27 Ta 181 Yb 174 Tm169 Cs 133 Axis 1 V 51 Er166Ho 16 5 Tb 159 Th 232 Ti 47 Dy 164 Eu 153 Sc 45 K 39 Y Nb 93 Ni 60 Fe57 Sm 152 Rb 85 Gd158 Cr 52 Li 7 Cu65 Pr 141 Co 59 Nd 142
Pb 208 La 139 Ce 140 Ba 138
Zn 66
Mn 55 Sr 88
Ca 44
Figure 5. Correspondence analysis (CA) ordination graph of sample elements, axes 1 and 2.
80
CA ordination of elements, all samples
Cr 52 3 K 39 s i Fe57 Na 23
V 51 Ax Pb 208 Sc 45 Ni 60 Y Th 232 Nb 93 Ti 47 Ba 138 Si-30 Gd158 U 238 Cs 133 Rb 85 Zn 66 As 75 Zr 90 Axis 1 Nd 142 Li 7 Sm 152 Pr 141 Dy 164 Er166 Tb 159 Ho 16 5 Ce 140 Eu 153 Ta 181 La 13 9 Tm169 Cu65 Yb 174 Hf 18 0 Lu 175 Sb 121 Co 59 Mg 24 Al 27
Sn 120
Ca 44
Sr 88 Mn 55
Figure 6. Correspondence analysis (CA) ordination graph of sample elements, axes 1 and 3.
CA ordination of elements, all samples
Cr 52 K 39 3 s i Fe57 Na 23 Pb 208 V 51 Sc 45 Ax Ni 60 Y Nb 93 Th 232 Ti 47 Ba 138 Si-30 Gd158 Cs 13U3 238 Rb 85 Zn 66 As 75 Zr 90 Axis 2 Nd 142 Li 7 Pr 141 Sm 152 Dy 164 Er166 Tb 15Ho9 16 5 Eu 153 Ta 181 Ce 140 Cu65 Tm169 Hf 18 0 La 139 Yb 174 Lu 175 Sb 121 Co 59 Mg 24 Al 27
Sn 120
Ca 44
Sr 88 Mn 55
Figure 7. Correspondence analysis (CA) ordination graph of sample elements, axes 2 and 3.
81 their temper and surface treatment, exhibit among the highest calcium concentrations
(e.g., grog-tempered broadline incised sherds GR2gi6 and GR2gi7 from Craig’s Landing, at 113,626 ppm and 106,931 ppm, respectively; grog-tempered cordmarked sherd
NO503gc3 from the Monk site, at 147,811 ppm; and grog-tempered plain sherd
CL527gp2 from Kellogg, at 153,445 ppm). However, also as shown in Appendix A1, other specimens from some of the same Black Prairie sites do not show comparably high calcium levels, but differ little in their concentrations of this element from specimens from sites in all other physiographic regions. Hence, there is no reason to suspect that the calcium concentration variability among Black Prairie sherd samples, whatever their temper and surface treatment, represents anything other than local calcium variation in sediment deposits. Black Prairie specimens with very high calcium levels may represent the nature of the sediments used to make the pottery, or may reflect differential post- depositional leaching of calcium from sediments into the potsherds. In any case, the fact that calcium concentrations vary widely among sherds of all surface treatments, as well as among fired clay samples, indicates that calcium content has no implications for identifying nonlocal pottery, whether found at Black Prairie sites or elsewhere.
As for strontium and manganese, as seen in the spreadsheet matrix data in
Appendix A1, the levels of these elements vary widely among sherds from the same sites, even sherds which share the same surface finish modes, whether broadline incised, plain, or cordmarked. Therefore, no extraregional sourcing implications are revealed by concentration variability in these elements.
The only other element outstanding in the CA elemental concentration ordinations is sodium (Na), shown in marked isolation along Axis 1 in Figure 5, and, especially, in
82 Figure 6. As discussed in Chapter III, Steponaitis et al. (1996:565-566), Lynott et al.
(2000), O’Brien et al. (1995), and Neff (n.d.:337, 346, 348, 353) have presented findings
revealing that, in contrast to sediments of adjacent upland regions, the alluvial sediments
of the Mississippi Valley contain relatively high concentrations of sodium (as well as
potassium). To see whether the compositional data collected for this study conform to
the above-noted general geographical pattern, it was decided to create an ordination diagram of all specimens by analyzing the elemental concentration data for sodium, and for potassium as well. (Although potassium is not segregated in the ordination diagrams in Figures 5-7, it is included with sodium in the analysis because the PC-ORD program will not ordinate data from a single-dimension variable, in this case, a single element.)
Principal Components Analysis
CA was judged unsuitable for analysis of the sodium and potassium data, because
PC-ORD produced artificially skewed CA ordination patterns (not included herein) in the
form of parabolas and sine curves (cf. Ludwig and Reynolds 1988:257; Peacock
1998:244-245). As an alternative to CA, principal components analysis (PCA) was done
to ordinate the sodium and potassium concentration data; this type of analysis produced a
suitable ordination (albeit only for axes 1-2; see Figure 8), without the spurious linear
distortions in the CA ordinations.
The resulting PCA specimen plots in Figure 8 are quite interesting in that they
largely conform to the geographically differentiated pattern noted in the above-cited
studies: With the exception of only one specimen (grog-tempered plain sherd
YZ515gp4), all pottery and fired clay from sites in the Yazoo Basin of the Lower
83 Mississippi Valley (indicated by specimen prefix codes CO, SH, and YZ for Coahoma,
Sharkey, and Yazoo counties, respectively) are confined to the right side of the diagram
(right of Axis 2).
In contrast, most specimens from sites located in physiographic zones east of the
Yazoo Basin are concentrated along the left one-third of Axis 1, (left of Axis 2). On the other hand, some specimens from non-Yazoo Basin/Lower Mississippi Valley sites (all specimens from the coastal Godsey site [sample prefix HR591]; grog-tempered broadline incised sherd NO503gi3 from Monk; grog-tempered plain sherd LO538gp2 from
Vaughn; grog-tempered broadline incised sherd GR2gi6 from Craig’s Landing; and grog- tempered plain sherd UN500gp7 from Ingomar) are plotted near the middle of Axis 1 along with a number of Yazoo Basin specimens. However, none of these specimens are located along the right one-third of Axis 1 in the ordination, where only Yazoo Basin material is plotted. As such, the overall pattern seen in the PCA ordination of specimens by sodium and potassium concentration shows a generally dichotomous pattern of Yazoo
Basin material and material of other zones sorting separately, even considering some overlap at the center of the diagram. Furthermore, it is noteworthy that both of the northeast Mississippi pottery specimens identified using Marksville-series type-variety stylistic criteria and therefore identified in site reports as possible imports from the Lower
Mississippi Valley – CL527gi1 from Kellogg and LO538gi1 from Vaughn – are plotted not with the Yazoo Basin site material along the center and right of Axis 1 in Figure 8, but instead are concentrated on the left with material almost exclusively from non-Yazoo
Basin sites. Judging from the ordination pattern in Figure 8, it seems unlikely that these
84 CO666gp8 2 s i Ax
CO666gi6 CO666gs5 MD13gi2 CO666gs1 CO666gi4 WI516gs2 CO666gi7 GR2gi5 CO666gi2 SH500gi1 GR2gi7 SH520gp2 OK888sp2 SH500gp3 Y Z515gi2 NO503gc3 NO503gi1 OK888sp3 Y Z515g i3 GR2gc8 UN500gi3 Y Z515gi1 CK21ss9 CK21gi1 GR2gp9 NO503gc2 LO538 gp 2 GR2gc2 CK21gs5 SH500gi2 CK21gs4 Y Z515gp 4 UN500gp8 GR2gi6 CO666gi3 GR2gp3 UN500gi1 MD13gi1 MD13gc3 UN500gp7 GR2fc10 UN500gi5 HR 59 1g i3 Axis 1 GR2fc4 OK888gi1 SH520gs1 WI516fc3 CK21si8 UN500gp6 CL527f c5CK21gp7 NO503fc4 LO53 8gp3 GN687gi2 LO53 8f c5 CO666 fc CK21gi3 UN500gi4UN500gi2 CO682gs1CO682gc4 GN687gp6 WI516gi1 687gi3 CK21gp6 CO682gs3 GN685gp2 CO502gc4 7gi1 CO502gs2 CO682gc5 7 GN687gp4 GR2si1 UN500fc9 CO502gs1 CO682gs2 GN687gp5 GN685gi1 OK888fc4 HR59 1g p 4 N687fc8 CL527g p2 MD13fc4 HR 59 1g i2 CK21gi2 CL527gp 3 CL527f c4 HR59 1f c5 CO502gc3 CL527gi1 LO53 8gi1 HR 59 1g i1 SH500f c4
SH520f c3 Y Z515f c5 SH500f c4
CO502fc5 CO682fc6
Figure 8. Principal components analysis (PCA) ordination graph of all samples by sodium (Na) and potassium (K) concentration.
two specimens were imported to northeast Mississippi from the Yazoo Basin (see also following paragraph).
The aggregate elemental concentration data for sodium and potassium accounting for the above-discussed PCA ordination are consistent with the pattern reported bySteponaitis et al. (1996:565-566) and Neff (n.d. 337, 346, 348, 353) that levels of both elements are higher in Yazoo Basin sediments, and pottery made from them, than in physiographic zones to the east.
85 As can be seen from the ppm elemental concentration data in Appendix A1, the
Yazoo Basin specimens exhibit levels of sodium ranging from a low of 1,521 ppm for grog-tempered plain sherd SH520gp2 to a high of 11,926 ppm for fired clay specimen
CO682fc6; the average sodium concentration for all Yazoo Basin specimens is 4,826 ppm. These data contrast with the generally lower sodium figures for non-Yazoo Basin material, ranging from a low of only 106 ppm for fired clay specimen GN687fc7, to a high of 5,012 ppm for grog-tempered broadline incised sherd HR591gi1; the average sodium concentration for all non-Yazoo Basin specimens is 1,696 ppm.
The same pattern is seen with potassium: Appendix A1 shows that levels of this element range from a low of 15,417 ppm for fired clay specimen CO502fc6, to a high of
49,420 ppm for grog-tempered stamped sherd CO666gs5, with an average potassium concentration for Yazoo Basin material of 27,925 ppm. This is in sharp contrast to non-
Yazoo Basin material, which shows potassium levels from a low of 3,388 ppm for fired clay specimen GN687fc7, to a high of 31,706 ppm for grog-tempered broadline incised sherd MD13gi2; the average concentration of potassium for non-Yazoo Basin samples is
15,640 ppm.
It can be seen in Appendix A1 that the sodium and potassium concentrations for northeast Mississippi and Alabama grog-tempered broadline incised sherds, some of which (as mentioned in the previous paragraph) have been considered nonlocal, differ little from plain and cordmarked material and archaeological fired clay from their respective sites; this, in combination with the ordination data discussed above, further reinforces a conclusion that they are not extraregional imports. (Almost across the board in Appendix A1, the Na and K concentrations of the non-archaeological fired clay
86 samples made from raw sediments collected for this study [see Chapter VII for
identification codes] are markedly higher than for potsherds and archaeological fired clay, suggesting post-depositional leaching of these elements out of both potsherds and archaeological non-pottery fired clay over millennia.) These broadline incised specimens
and their respective sodium (Na) and potassium (K) concentrations are:
CL527gi1 (Kellogg): Na=1,759 ppm; K=4,872 ppm
LO538gi1 (Vaughn): Na=3,327 ppm; K=10,688 ppm
OK888gi1: Na=990 ppm; K=13,875 ppm
UN500gi1 (Ingomar): Na=1,358 ppm; K=17,879 ppm
87
CHAPTER X
SUMMARY AND CONCLUSIONS
The primary purpose of this thesis was to test, using elemental characterization evidence, published assertions that sparsely occurring Marksville-style prehistoric ceramics found in northeastern Mississippi and western Alabama were imported from the
Lower Mississippi Valley (the core area of the Marksville culture) or other regions.
Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry was used to determine elemental concentrations for 99 artifacts, consisting of Marksville-style sherds, locally abundant plain and cordmarked sherds, archaeological non-pottery fired clay, and fired clay prepared from newly collected raw clay samples. The LA-ICP-MS data were then subjected to correspondence analysis (CA) and principal components analysis (PCA) to generate patterns of elemental compositional similarity and differences. The analyzed artifacts represent a sample of archaeological sites located in major physiographic zones across Mississippi and western Alabama, to document interregional variability in ceramics as well as corresponding interregional variability in geological sediments, which should be reflected in both pottery and the clays used to produce it. By this means, specimens exhibiting elemental concentration profiles at marked variance with other specimens from their respective site assemblages could be identified as of possible nonlocal origin.
88 As presented in Chapter IX, the composition results are uniformly negative: none
of the Marksville-style specimens declared or suggested in various publications to be long-distance imports displayed elemental concentration profiles markedly different from abundant plain and cordmarked pottery and archaeological fired clay from the same sites.
Therefore, however much some northeastern Mississippi and western Alabama specimens visually or texturally resemble Marksville-series material in the Lower
Mississippi Valley or other regions, the elemental composition evidence supports the conclusion that they actually were made locally, i.e., within the region, if not at or near
the respective sites at which they were found.
In addition to the findings that the Marksville-style pottery in the study sample
generally exhibited no significant variance from local pottery and archaeological fired
clay, principal components analysis of the elemental concentration data for the entire
sample of specimens showed that pottery and fired clay found at sites in the Yazoo Basin
subdivision of the Lower Mississippi Valley exhibited consistently higher concentrations
of both sodium and potassium than did specimens from sites in any of the upland
physiographic zones to the east. These findings corroborate those of Steponaits et al.
(1996:565-566), Neff (n.d.:337, 346, 348, 353), and others (see Chapter III) which
revealed the same dichotomous pattern of higher sodium and potassium concentrations in
the Lower Mississippi Valley vs. lower levels for both elements in upland zones to the
east. As pointed out by Neff (n.d.:337), these data reveal that “that the compositional
patterns discovered in the archaeological ceramics reflect underlying variation in the proportions of clay minerals present in the sediments of the various regions.”
Accordingly, the correspondingly dichotomous pattern of the sodium and potassium data
89 as reported herein support a further conclusion that there was no interregional transport of
any of the Marksville-style material in the sample from the Yazoo Basin to the other
regions, or vice-versa.
In Chapter IV, it was suggested that the production and possible interregional
distribution of elaborately decorated ceramics could be considered from an evolutionary
perspective, using the concept of “costly signaling” (Bird and O’Connell 2006:163-164;
Neiman 1997). A possible case in point may be seen in Blomster et al.’s (2005)
identification via compositional analysis of locally made and imported Olmec pottery in
Mexico, which exhibits surface decoration quite similar in its high degree of elaboration
to Marksville-style ceramics. In many areas of Mesoamerica, the occurrence in ceramic
assemblages of low quantities of imported Olmec pottery, and also of locally made ware
with copied Olmec-style motifs, is associated by Blomster et al. (2005:1071) with “elites”
and “leaders.” From the viewpoint of the costly signaling concept, apparent wasteful
elaboration in artifacts may actually serve to signify the ability of powerful individuals to aquire quantities of elaborate goods; both the affluent consumer (i.e., in prehistoric cultures elites or leaders) benefits, as well as the potters, whose extra time and energy expended in producing elaborate pottery is worth the effort. In effect, elaborate pottery decoration may be selected for its ability to advertise the consumer’s competitive fitness, in both the political and biological sense.
A roughly analogous scenario in which costly signaling may apply can be
observed in the case of Marksville-style ceramics found in northeastern Mississippi, but
perhaps not in the context of the pottery analyzed for this study. The sample analyzed for
this thesis is biased in that it includes no Marksville-style material from burial mounds in
90 northeast Mississippi. This, along with the fact that the amount of Marksville material
from sites in northeast Mississippi is minuscule, may mean that the negligible residual representation of such pottery at non-mound sites might be regarded as mere
“background noise,” with no value for addressing questions of evolutionary selection of
artifact traits.
In contrast, as discussed in Chapter VI, vessels which have been classified as
Marksville Incised, var. Marksville and Marksville Stamped, var. Marksville vessels have
been found in the region at, respectively, at the Pharr and Bynum sites. In addition,
artifacts made of exotic minerals – mica, copper, galena – have been found in these
mounds and at others in the region; these are known to be long distance imports, as these
minerals do not occur naturally in Mississippi (Baca and Peacock 1996:16-17; Bohannon
1972; Cotter and Corbett 1951:8). Only a few individuals have been found interred in
burial mounds in the region, which suggests that they possessed higher social status than
the majority of the population (Baca and Peacock 1996:18). The occurrence of elaborate
Marksville-style ceramics in some of these burial mounds, as well as other artifacts made
of exotic minerals, could be construed as evidence of costly signaling, i.e., advertising of
the high status of the dead interred in the mound, and, by extension, the status of the
kinship group of the deceased. It follows that the high expenditure of effort (energy)
involved in the extraction and interregional transport of regionally rare goods was not
wasted, but was to the competitive advantage (fitness) of both trader and consumer,
which can be viewed as selection for material elaboration.
Therefore, the possibility remains that the Marksville-style vessels found at the
Pharr and Bynum sites could be interregional imports, perhaps from the Lower
91 Mississippi Valley. Unfortunately, I was unable to secure samples of Marksville material from Pharr and Bynum for analysis. However, given their prime potential for addressing the persistent question of nonlocal Marksville pottery in northeast Mississippi, it is highly recommended that any future opportunities to conduct elemental sourcing studies with these artifacts be pursued.
92
REFERENCES CITED
Anderson, David G., and Robert C. Mainfort, Jr. 2002 An Introduction to Woodland Archaeology in the Southeast. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 1-19. University of Alabama Press, Tuscaloosa.
Andras, Paula Marie 2004 A Place in Prehistory: The Function of Ingomar Mounds, a Middle Woodland Site, Northeastern Mississippi. Unpublished M.A. thesis, Department of Sociology, Anthropology, and Social Work, Mississippi State University.
Atkinson, James R. 1974 Appendix A: Test Excavations at the Vaughn Mound Site. In Archaeological Survey and Test Excavations in the Upper-Central Tombigbee River Valley: Aliceville-Columbus Lock and Dam and Impoundment Areas, Alabama and Mississippi, by Marc D. Rucker, pp. 115-158. Department of Anthropology, Mississippi State University.
Atkinson, James R., and Jack D. Elliott, Jr. 1979 A Cultural Resources Survey and Evaluation in the Tallahala Creek Lake, Jasper County, Mississippi. Department of Anthropology, Mississippi State University.
Atkinson, James R., John C. Phillips, and Richard Walling 1980 The Kellogg Village Site Investigations, Clay County, Mississippi. Department of Anthropology, Mississippi State University.
Baca, Keith A., and Evan Peacock 1996 The Brogan Mound, a Middle Woodland Site in Clay County, Mississippi. In Mounds, Embankments, and Ceremonialism in the Midsouth, edited by Robert C. Mainfort and Richard Walling, pp. 12-21. Arkansas Archeological Survey Research Series No. 46, Fayetteville.
Bicker, Alvin R., Jr. 1969 Geologic Map of Mississippi. Mississippi Geological Survey (n.p.).
Bird, Douglas W., and James F. O’Connell 2006 Behavioral Ecology and Archaeology. Journal of Archaeological Research 14(2):143-188.
93 Bishop, Ronald L., Robert L. Rands, and George R. Holley 1982 Ceramic Compositional Modeling in Archaeological Perspective. In Advances in Archaeological Method and Theory, Vol. 5, edited by Michael B. Schiffer, pp. 275-331. Academic Press, New York.
Bishop, Ronald L., and Veletta Canouts 1993 Archaeometry. In The Development of Southeastern Archaeology, edited by Jay K. Johnson, pp. 160-183. University of Alabama Press, Tuscaloosa.
Blitz, John H. 1984 A Cultural Resources Survey in the Tombigbee National Forest, Mississippi. USDA Forest Service, Jackson, Mississippi.
1986 The McRae Mound: A Middle Woodland Site in Southeastern Mississippi. Mississippi Archaeology 21(2):11-40.
Blitz, John H., and C. Baxter Mann 1993 Archaeological Investigations in Coastal Jackson County, Mississippi. Interim Report 1, Mississippi Gulf Coast Archaeological Project.
2000 Fisherfolk, Farmers, and Frenchmen: Archaeological Explorations on the Mississippi Gulf Coast. Archaeological Report No. 30, Mississippi Department of Archives and History. Jackson.
Blomster, Jeffrey P., Hector Neff, and Michael D. Glascock 2005 Olmec Pottery Production and Export in Ancient Mexico Determined Through Elemental Analysis. Science 307:1068-1072.
Bohannon, Charles F. 1972 Excavations at the Pharr Mounds, Prentiss and Itawamba Counties, Mississippi. National Park Service, Washington, D.C.
Booth, David C., and Darrel W. Schmitz 1983 Economic Minerals Map of Mississippi. Mississippi Bureau of Geology (n.p.).
Brookes, Samuel O. 1976 The Grand Gulf Mound: Salvage Excavation of an Early Marksville Burial Mound in Claiborne County, Mississippi. Archaeological Report No. 1, Mississippi Department of Archives and History. Jackson.
Brose, David S., and N’omi Greber (editors) 1979 Hopewell Archaeology: The Chillicothe Conference. Kent State University Press, Kent, Ohio.
94 Brown, C.T., C.A. Darby, J.A. Green, C. Davies, M. Williams, G. Gordon, F. Venta, and W.P. Athens 1996 Phase III Data Recovery of Site 22PR533 for the Proposed Florida Gas Transmission Company Phase III Expansion Project, Pearl River County, Mississippi. R. Christopher Goodwin and Associates, Inc., New Orleans.
Brown, Ian W. 1998 Decorated Pottery of the Lower Mississippi Valley: A Sorting Manual. Mississippi Archaeological Association and Mississippi Department of Archives and History.
Campbell, L. Janice, Christopher Hays, Paul V. Heinrich, James H. Mathews, Gregory Mikell, A. Frank Servello, Prentice M. Thomas, Jr., and Eugene Wilson 1988 Cultural Resources Testing on the Mobile Bay Pipeline Project, Choctaw, Washington and Mobile Counties, Alabama, Vol. 1. Report of Investigations No. 167, New World Research, Inc. Prepared for EMANCO, Inc., Fort Walton Beach, Fla.
Carleton, Kenneth H. 1999 Nanih Waiya (22WI500): An Historical and Archaeological Overview. Mississippi Archaeology 34(2):125-155.
Cogswell, James W. 1998 Woodland Sand-Tempered Pottery and Its Distribution Across Dunklin, Mississippi, New Madrid, and Pemiscot Counties, Missouri. Unpublished Ph.D. dissertation, University of Missouri, Columbia.
Conant, Louis C. 1942 Union County Mineral Resources. Bulletin 45, Mississippi State Geological Survey.
Cotter, John L., and John M. Corbett 1951 Archeology of the Bynum Mounds, Mississippi. Archeological Research Series 1. National Park Service, Washington, D.C.
Dunnell, Robert C. 1980 Evolutionary Theory and Archaeology. In Advances in Archaeological Method and Theory, vol. 3, edited by Michael B. Schiffer, pp. 35-99. Academic Press, New York.
1986 Methodological Issues in Americanist Artifact Classification. In Advances in Archaeological Method and Theory, Vol. 9, edited by Michael B. Schiffer, pp. 149-207. Academic Press, New York.
95 1995 What Is It That Actually Evolves? In Evolutionary Archaeology: Methodological Issues, edited by Patrice A. Teltser, pp. 33-50. University of Arizona Press, Tucson.
2002 Systematics in Prehistory. Blackburn Press, Caldwell, N.J.
Earle, Timothy K. 1982 Prehistoric Economies and the Archaeology of Exchange. In Contexts for Prehistoric Exchange, edited by Jonathon E. Ericson and Timothy K. Earle, pp. 1- 12. Academic Press, New York.
Ericson, Jonathon E., and Timothy K. Earle 1982 Preface. In Contexts for Prehistoric Exchange, edited by Jonathon E. Ericson and Timothy K. Earle, pp. xv-xvi. Academic Press, New York.
Ericson, Jonathon E., and Timothy G. Baugh 1994 Systematics of the Study of Prehistoric Exchange in North America. In Prehistoric Exchange Systems in North America, edited by Timothy G. Baugh and Jonathon E. Ericson, pp. 1-15. Plenum Press, New York.
Fields, Rita D. 2005 Archaeological Data Recovery at Sites 22GN680, 22GN685, and 22GN687, Greene County, Mississippi. Report prepared for the Mississippi Department of Transportation, Environmental Division, Jackson.
Fisk, Harold N. 1944 Geological Investigations of the Alluvial Valley of the Lower Mississippi River. U.S. Army Corps of Engineers, Mississippi River Commission, Vicksburg, Mississippi.
Ford, James A. 1936 Analysis of Indian Village Site Collections from Louisiana and Mississippi. Anthropological Study No. 2, Department of Conservation, Louisiana Geological Survey. New Orleans.
Ford, James A., and Gordon R. Willey 1940 Crooks Site, A Marksville Period Burial Mound in La Salle Parish, Louisiana. Anthropological Study No. 3, Department of Conservation, Louisiana Geological Survey. New Orleans.
Ford, Janet L. 1980 Alas, Poor Womack! Mississippi Archaeology 15(2):26-32.
1981 Time and Temper in the North Central Hills of Mississippi. Journal of Alabama Archaeology 27(1):57-71.
96
1988 An Examination of the Twin Lakes Phase. In Middle Woodland Settlement and Ceremonialism in the Mid-South and Lower Mississippi Valley, edited by Robert C. Mainfort, Jr., pp. 61-67. Archaeological Report No. 22, Mississippi Department of Archives and History, Jackson.
1989 Time and Temper Meet Trend and Tradition. Mississippi Archaeology 24(1):1- 16.
Futato, Eugene M. 1989 An Archaeological Overview of the Tombigbee River Basin, Alabama and Mississippi. University of Alabama, Alabama State Museum of Natural History, Division of Archaeology, Report of Investigations 59.
Galaty, Michael L. n.d. Ceramic Petrography and the Classification of Mississippi’s Archaeological Pottery by Fabric: A GIS Approach. In Time’s River: Archaeological Syntheses from the Lower Mississippi River Valley, edited by Janet Rafferty and Evan Peacock, pp. 356-406. Manuscript in preparation for publication by the University of Alabama Press, Tuscaloosa.
Gibson, Jon L. 1993 Ceramics. In The Development of Southeastern Archaeology, edited by Jay K. Johnson, pp. 18-35. University of Alabama Press, Tuscaloosa.
1994 Empirical Characterization of Exchange Systems in Lower Mississippi Valley Prehistory. In Prehistoric Exchange Systems in North America, edited by Timothy G. Baugh and Jonathon E. Ericson, pp. 127-175. Plenum Press, New York.
Gibson, Jon L., and Mark A. Melancon 2004 In the Beginning: Social Contexts of First Pottery in the Lower Mississippi Valley. In Early Pottery: Technology, Function, Style, and Interaction in the Lower Southeast, edited by Rebecca Saunders and Christopher T. Hays, pp. 169- 192. Tuscaloosa: University of Alabama Press.
Gladfelter, Bruce G. 2001 Archaeological Sediments in Humid Alluvial Environments. In Sediments in Archaeological Context, edited by Julie K. Stein and William R. Farrand, pp. 93- 125. University of Utah Press, Salt Lake City.
Goad, Sharon I. 1978 Exchange Networks in the Prehistoric Southeastern United States. Unpublished Ph.D. dissertation, Department of Anthropology, University of Georgia, Athens.
97 1979 Middle Woodland Exchange in the Prehistoric Southeastern United States. In Hopewell Archaeology: The Chillicothe Conference, edited by David S. Brose and N’omi Greber, pp. 239-246. Kent State University Press, Kent, Ohio.
Gower, J. C. 1987 Introduction to Ordination Techniques. Developments in Numerical Ecology, edited by P. Legendre and L. Legendre, pp. 3-64. Springer-Verlag, Berlin.
Greengo, Robert E. 1964 Issaquena: An Archaeological Phase in the Yazoo Basin of the Lower Mississippi Valley. Society for American Archaeology, Memoir No.18.
Griffin, James B. 1979 An Overview of the Chillicothe Hopewell Conference. In Hopewell Archaeology: The Chillicothe Conference, edited by David S. Brose and N’omi Greber, pp. 266-279. Kent State University Press, Kent, Ohio.
Grim, Ralph E. 1936 The Eocene Sediments of Mississippi. Bulletin 30, Mississippi State Geological Survey.
Haag, William G. 1952 Archaeological Survey of the Grenada Reservoir in Mississippi. Unpublished report, on file at the Southeast Archeological Center, National Park Service, Tallahassee, Fla.
Harbottle, Garman 1982 Chemical Characterization in Archaeology. In Contexts for Prehistoric Exchange, edited by Jonathon E. Ericson and Timothy K. Earle, pp. 13-51. Academic Press, New York.
Herz, Norman, and Ervan G. Garrison 1998 Geological Methods for Archaeology. Oxford University Press, New York.
Hirata, T. 1997 Ablation Technique for Laser Ablation Inductively Coupled Plasma Mass Spectrometry. Journal of Analytical and Atomic Spectrometry 12:1337-1342.
Jackson, H. Edwin, and Rita D. Fields 2000 Phase II Testing of 22GN668, Greene County, Mississippi. Report prepared for the Mississippi Department of Transportation, Environmental Division, Jackson.
Jackson, H. Edwin, Melissa L. Higgins, and Robert E. Reams 2002 Woodland Cultural and Chronological Trends on the Southern Gulf Coastal Plain: Recent Research in the Pine Hills of Southeastern Mississippi. In The
98 Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 228-248. University of Alabama Press, Tuscaloosa.
Jenkins, Ned J. 1975 Archaeological Investigations in the Gainesville Lock and Dam Reservoir: 1974. Department of Anthropology, University of Alabama, Tuscaloosa.
1979 Miller Hopewell of the Tombigbee Drainage. In Hopewell Archaeology: The Chillicothe Conference, edited by David S. Brose and N’omi Greber, pp. 171-180. Kent State University Press, Kent, Ohio.
1981 Gainesville Lake Area Ceramic Description and Chronology. Report of Investigations No. 12. Office of Archaeological Research, University of Alabama, Tuscaloosa.
1982 Archaeology of the Gainesville Lake Area: Synthesis. Report of Investigations No. 23. Office of Archaeological Research, University of Alabama, Tuscaloosa.
Jenkins, Ned J., and H. Blaine Ensor 1981 The Gainesville Lake Area Excavations. Report of Investigations No. 11. Office of Archaeological Research, University of Alabama, Tuscaloosa.
Jenkins, Ned J., and Richard A. Krause 1986 The Tombigbee Watershed in Southeastern Prehistory. University of Alabama Press, University.
Jennings, Jesse D. 1941 Chickasaw and Earlier Indian Cultures of Northeast Mississippi. Journal of Mississippi History 3(3):155-226.
Johnson, Jay K. 1984 Prehistoric Settlement in the Upper Yocona Drainage, North Central Mississippi. Mississippi Archaeology 19(2):13-23.
1988 Woodland Settlement in Northeast Mississippi: The Miller Tradition. In Middle Woodland Settlement and Ceremonialism in the Mid-South and Lower Mississippi Valley, edited by Robert C. Mainfort, Jr., pp. 49-59. Archaeological Report 22, Mississippi Department of Archives and History, Jackson.
1994 Prehistoric Exchange in the Southeast. In Prehistoric Exchange Systems in North America, edited by Timothy G. Baugh and Jonathon E. Ericson, pp. 99-125. Plenum Press, New York.
99 Johnson, Jay K., Gena M. Aleo, Rodney T. Stuart, and John Sullivan 2002 The 1996 Excavations at the Batesville Mounds: A Woodland Period Platform Mound Complex in Northwest Mississippi. Archaeological Report No. 32. Mississippi Department of Archives and History, Jackson.
Keady, Donald M. 1962 Geologic Study along Highway 45 from Tennessee Line to Meridian, Mississippi. Bulletin 94, Mississippi Geological Survey.
Keith, Grace F. 1997 A Technological Analysis of Ceramics from the Leaf River Drainage, Southeast Mississippi. Unpublished master’s thesis, Department of Anthropology and Sociology, University of Southern Mississippi, Hattiesburg.
Kelly, Arthell 1973 The Geography. In A History of Mississippi, edited by R.A. McLemore, pp. 3- 23. University and College Press of Mississippi, Hattiesburg.
1974 Topography. In Atlas of Mississippi, edited by Ralph D. Cross and Robert W. Wales, pp. 4-7. University Press of Mississippi, Jackson.
Kennett, Douglas J., Sachiko Sakai, Hector Neff, Richard Gosett, and Daniel O. Larson 2002 Compositional Characterization of Prehistoric Ceramics: A New Approach. Journal of Archaeological Science 29:443-455.
Kidder, Tristram R. 2002 Woodland Period Archaeology of the Lower Mississippi Valley. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 66-90. University of Alabama Press, Tuscaloosa.
Koehler, Thomas H. 1966 Archaeological Excavation of the Womack Mound (22Ya1). Bulletin 1, Mississippi Archaeological Association, University, Mississippi.
Lafferty, Robert H., III 1994 Prehistoric Exchange in the Lower Mississippi Valley. In Prehistoric Exchange Systems in North America, edited by Timothy G. Baugh and Jonathon E. Ericson, pp. 177-213. Plenum Press, New York.
Larson, Daniel O., Sachiko Sakai, and Hector Neff 2005 Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) as a Bulk Chemical Characterization Technique: Comparison of LA-ICP-MS, Digestion-ICP-MS, and INAA Data on Virgin Branch Anasazi Ceramics. In Laser Ablation-ICP-MS in Archaeological Research, edited by Robert J.
100 Speakman and Hector Neff, pp. 95-102. University of New Mexico Press, Albuquerque.
Logan, William N. 1907 Clays of Mississippi. Bulletin No. 2, Mississippi State Geological Survey. Brandon Printing Co., Nashville.
1908 Clays of Mississippi, Part II: Brick Clays and Clay Industry of Southern Mississippi. Bulletin No. 4, Mississippi State Geological Survey. Brandon Printing Co., Nashville.
1909 The Pottery Clays of Mississippi. Bulletin No. 6, Mississippi State Geological Survey. Brandon Printing Co., Nashville.
Lowe, E. N. 1933 Coastal Plain Stratigraphy of Mississippi, Part First: Midway and Wilcox Groups. Bulletin 25, Mississippi State Geological Survey.
Ludwig, J. A., and J. F. Reynolds 1988 Statistical Ecology. John Wiley and Sons, New York.
Lyman, R. Lee, and Michael J. O’Brien 2002 Classification. In Darwin and Archaeology: A Handbook of Key Concepts, edited by John P. Hart and John Edward Terrell, pp. 69-88. Bergin & Garvey, Westport, Conn.
Lyman, R. Lee, Michael J. O’Brien, and Robert C. Dunnell 1997 Introduction. In Americanist Culture History: Fundamentals of Time, Space, and Form, edited by R. Lee Lyman, Michael J. O’Brien, and Robert C. Dunnell, pp. 1- 16. Plenum Press, New York.
Lynott, Mark J., Hector Neff, James E. Price, James W. Cogswell, and Michael D. Glascock 2000 Inferences about Prehistoric Ceramics and People in Southeast Missouri: Results of Ceramic Compositional Analysis. American Antiquity 65:103-126.
Maggetti, M. 1982 Phase Analysis and Its Significance for Technology and Origin. In Archaeological Ceramics, edited by J.S. Olin and A.D. Franklin, pp. 121-133. Smithsonian Institution, Washington, D.C.
Maher, Thomas O. 1983 Ceramic Exchange in the Southeastern United States: An Examination of Three Methods for Mineralogically Characterizing Aboriginal Ceramics from Archaeological Sites in Alabama, Louisiana and Mississippi. Unpublished M.A.
101 thesis, Department of Anthropology, State University of New York at Binghamton.
Marshall, Richard A. 1982 A Report on Archaeological Test Excavations at Goode Lake, Jackson County, Mississippi. Archaeological Report No. 10. Mississippi Department of Archives and History and Jackson County Port Authority.
Matson, Frederick R. 1965 Ceramic Ecology: An Approach to the Study of the Early Cultures of the Near East. In Ceramics and Man, edited by Frederick R. Matson, pp. 161-177. Viking Fund Publications in Archaeology, Chicago.
McClung, Terry 1998 The Central Midden: News from the National Forests in Mississippi – Bienville National Forest. Mississippi Archaeological Association Newsletter 33(2):13-14.
2005 The Central Midden: News from the National Forests in Mississippi – Bienville National Forest. Mississippi Archaeological Association Newsletter 40(1):8.
McNutt, Charles H. 1996 The Upper Yazoo Basin in Northwest Mississippi. In Prehistory of the Central Mississippi Valley, edited by Charles H. McNutt, pp. 155-185. University of Alabama Press, Tuscaloosa.
Morgan, William N. 1999 Precolumbian Architecture in Eastern North America. University Press of Florida, Gainesville.
Neff, Hector 1995 A Role for “Sourcing” in Evolutionary Archaeology. In Evolutionary Archaeology: Methodological Issues, edited by Patrice A. Teltser, pp. 69-112. University of Arizona Press, Tucson.
1998 Units in Chemistry-Based Ceramic Provenance Investigations. In Unit Issues in Archaeology: Measuring Time, Space, and Material, edited by Ann F. Ramenofsky and Anastasia Steffen, pp. 115-127. University of Utah Press, Salt Lake City.
2000 Neutron Activation Analysis for Provenance Determination in Archaeology. In Modern Analytical Methods in Art and Archaeology, edited by Enrico Ciliberto and Giuseppe Spoto, pp. 81-134. John Wiley and Sons, New York.
n.d. Review of Ceramic Compositional Studies from In and Around the Mississippi Valley. In Time’s River: Archaeological Syntheses from the Lower Mississippi
102 River Valley, edited by Janet Rafferty and Evan Peacock, pp. 326-355. Manuscript in preparation for publication by the University of Alabama Press, Tuscaloosa.
Neiman, Fraser D. 2007 Conspicuous Consumption as Wasteful Advertising: a Darwinian Perspective on Spatial Patterns in Classic Maya Terminal Monument Dates. In Rediscovering Darwin: Evolutionary Theory and Archeological Explanation, edited by C. Michael Barton and Geoffrey A. Clark, pp. 267-290. Archeological Papers of the American Anthropological Association No. 7.
O’Brien, Michael J., James W. Cogswell, Robert C. Mainfort, Jr., Hector Neff, and Michael D. Glascock 1995 Neutron-Activation Analysis of Campbell Appliquéd Pottery from Southeastern Missouri and Western Tennessee: Implications for Late Mississippian Intersite Relations. Southeastern Archaeology 14:181-194.
O’Brien, Michael J., and Robert C. Dunnell 1998 A Brief Introduction to the Archaeology of the Central Mississippi River Valley. In Changing Perspectives on the Archaeology of the Central Mississippi River Valley, edited by Michael J. O’Brien and Robert C. Dunnell, pp. 1-30. University of Alabama Press, Tuscaloosa.
O’Brien, Michael J., and R. Lee Lyman 2000 Applying Evolutionary Archaeology: A Systematic Approach. Kluwer Academic/Plenum Publishers, New York.
O’Brien, Michael J., R. Lee Lyman, and James W. Cogswell 2002 Culture-Historical Units and the Woodland Southeast: A Case Study from Southeastern Missouri. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 421-443.
O’Brien, Michael J., and W. Raymond Wood 1998 The Prehistory of Missouri. University of Missouri Press, Columbia.
Ortmann, Anthony L., and Tristram R. Kidder 2004 Petrographic Thin-Section Analysis of Poverty Point Pottery. In Early Pottery: Technology, Function, Style, and Interaction in the Lower Southeast, edited by Rebecca Saunders and Christopher T. Hays, pp. 193-209. University of Alabama Press, Tuscaloosa.
Paulson, Oscar L., Jr. 1974 Geology. In Atlas of Mississippi, edited by Ralph D. Cross and Robert W. Wales, pp. 8-11. University Press of Mississippi, Jackson.
103 Peacock, Evan 1992 Archaeological site survey form for the Allen’s Ridge I site (22OK888), Oktibbeha County, Mississippi. On file at the Historic Preservation Division of the Mississippi Department of Archives and History, Jackson.
1994 Archaeological site survey form for the Wildcat site (22WI557), Winston County, Mississippi. On file at the Historic Preservation Division of the Mississippi Department of Archives and History, Jackson.
1995a Shovel-Test Screening and Survey Methodology: A Field Study from the Tombigbee National Forest, North Mississippi. USDA Forest Service, Jackson, Mississippi.
1995b Tombigbee Chapter. Mississippi Archaeological Association Newsletter 30(1):1-2.
1997 Woodland Ceramic Affiliations and Settlement Pattern Change in the North Central Hills of Mississippi. Midcontinental Journal of Archaeology 22(2):237- 261.
1998 Fresh-Water Mussels as Indicators of Prehistoric Human Environmental Impact in the Southeastern United States. Unpublished Ph.D. thesis, Department of Archaeology and Prehistory, University of Sheffield.
2003 Excavations at Stinking Water (22WI515/516), a Prehistoric Habitation Site in the North Central Hills Physiographic Province of Mississippi. Mississippi Archaeology 38(1):3-91.
Peacock, Evan, and David W. Fant 2002 Biomantle Formation and Artifact Translocation in Upland Sandy Soils: An Example from the Holly Springs National Forest, North-Central Mississippi, U.S.A. Geoarchaeology: An International Journal 17(1):92-114.
Perkins, W.T., N.J.G. Pearce, and J.A. Westgate 1997 The Development of Laser Ablation ICP-MS and Calibration Strategies: Examples from the Analysis of Trace Elements in Volcanic Glass Shards and Sulfide Minerals. Geostandards Newsletter 21:175-190.
Phillips, Philip 1970 Archaeological Survey in the Lower Yazoo Basin, Mississippi, 1949-1955. Papers of the Peabody Museum of Archaeology and Ethnology, Vol. 60, Part 1. Harvard University, Cambridge.
104 Phillips, Philip, James A. Ford, and James B. Griffin 1951 Archaeological Survey in the Lower Mississippi Alluvial Valley, 1940-1947. Papers of the Peabody Museum of American Archaeology and Ethnology, Vol. 25. Harvard University, Cambridge.
Pielou, E. C. 1977 Mathematical Ecology. John Wiley and Sons, New York.
Potter, P.E., D. Helling, N.F. Shimp, and W. Van Wie 1975 Clay Mineralogy of Modern Alluvial Muds of the Mississippi River Basin. Bulletin du Centre de Recherches de Pau 9:353-389.
Rafferty, Janet E. 1986 A Critique of the Type-Variety System as Used in Ceramic Analysis. Mississippi Archaeology 21(2):40-50.
1987 The Ingomar Mounds Site: Internal Structure and Chronology. Midcontinental Journal of Archaeology 12(2):147-173.
1990 Test Excavations at Ingomar Mounds, Mississippi. Southeastern Archaeology 9(2):93-102.
2002 Woodland Period Settlement Patterning in the Northern Gulf Coastal Plain of Alabama, Mississippi and Tennessee. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 204-227. University of Alabama Press, Tuscaloosa.
Ramenofsky, Ann F. 1995 Evolutionary Theory and Native American Artifact Change in the Postcontact Period. In Evolutionary Archaeology: Methodological Issues, edited by Patrice Teltser, pp. 129-147. University of Arizona Press, Tucson.
Rapp, G., Jr. 1985 Provenance of Artifactual Raw Materials. In Archaeological Geology, edited by G. Rapp., Jr., and J.A. Gifford, pp. 353-375. Yale University Press, New Haven.
Rice, Prudence M. 1987 Pottery Analysis: A Sourcebook. University of Chicago Press, Chicago.
Rucker, Marc D. 1974 Archaeological Survey and Test Excavations in the Upper-Central Tombigbee River Valley: Aliceville-Columbus Lock and Dam and Impoundment Areas, Alabama and Mississippi. Department of Anthropology, Mississippi State University.
105 Saucier, Roger T. 1994 Geomorphology and Quaternary Geologic History of the Lower Mississippi Valley. U.S. Army Corps of Engineers Waterways Experiment Station, Vicksburg, Mississippi.
Schleidt, Maria L. 2002 The Central Midden: News from the National Forests of Mississippi – De Soto National Forest. Mississippi Archaeological Association Newsletter 37(3):7-8.
Seeman, Mark F. 1979 The Hopewell Interaction Sphere: The Evidence for Interregional Trade and Structural Complexity. Prehistory Research Series No. 5(2). Indiana Historical Society, Indianapolis.
Setzler, Frank M. 1933a Hopewell Type Pottery from Louisiana. Journal of the Washington Academy of Sciences 23(3):149-153.
1933b Pottery of the Hopewell Type from Louisiana. Proceedings of the United States National Museum 82:1-21.
Smith, Karen Y., and Fraser D. Neiman 2007 Frequency Seriation, Corrsespondence Analysis, and Woodland Period Ceramic Assemblage Variation in the Deep South. Southeastern Archaeology 26(1):47-72.
Snowden, J.O., and R.R. Priddy 1968 Loess Investigations in Mississippi: Geology of Mississippi Loess. Bulletin 111, Mississippi Geological, Economic and Topographical Survey.
Speakman, Robert J., and Hector Neff 2002 Evaluation of Painted Pottery from the Mesa Verde Region Using Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry. American Antiquity 67:137-144.
2005 The Application of Laser Ablation-ICP-MS to the Study of Archaeological Materials – An Introduction. In Laser Ablation-ICP-MS in Archaeological Research, edited by Robert J. Speakman and Hector Neff, pp.1-14. University of New Mexico Press, Albuquerque.
Stephenson, Lloyd W., and Watson H. Monroe 1940 The Upper Cretaceous Deposits. Bulletin 40, Mississippi State Geological Survey.
106 Steponaitis, Vincas P., M. James Blackman, and Hector Neff 1996 Large-Scale Patterns in the Chemical Composition of Mississippian Pottery. American Antiquity 61(3):555-572.
Thomas, Cyrus 1985 [1894] Report on the Mound Explorations of the Bureau of Ethnology. Smithsonian Institution Press, Washington, D.C.
Thomas, Emil P. 1942 The Claiborne. Bulletin 48, Mississippi State Geological Survey, University, Mississippi.
Tobia, S.K., and E.V. Sayre 1974 An Analytical Comparison of Various Egyptian Soils, Clays, Shales, and Some Ancient Pottery by Neutron Activation. In Recent Advances in Science and Technology of Materials, Vol. 3, edited by A. Bishop, pp. 99-128. Plenum Press, New York.
Toth, Edwin Alan 1974 Archaeology and Ceramics at the Marksville Site. University of Michigan Museum of Anthropology, Anthropological Papers 56.
1979 The Marksville Connection. In Hopewell Archaeology: The Chillicothe Conference, edited by David S. Brose and N’omi Greber, pp. 188-199. Kent State University Press, Kent, Ohio.
1988 Early Marksville Phases in the Lower Mississippi Valley: A Study of Culture Contact Dynamics. Archaeological Report No. 21. Mississippi Department of Archives and History, Jackson.
Veblen, Thorstein 1994 [1899] The Theory of the Leisure Class. Penguin Books, New York.
Walthall, John A. 1980 Prehistoric Indians of the Southeast: Archaeology of Alabama and the Middle South. University of Alabama Press, University.
1981 Galena and Aboriginal Trade in Eastern North America. Scientific Papers XVII, Illinois State Museum. Springfield, Illinois.
Weaver, David C., and James F. Doster 1982 Historical Geography of the Upper Tombigbee Valley. Center for the Study of Southern History and Culture, University of Alabama, University, Ala.
107 Weigand, Phil C., Garman Harbottle, and Edward V. Sayre 1977 Turquoise Sources and Source Analysis: Mesoamerica and the Southwestern U.S.A. In Exchange Sytstems in Prehistory, edited by Timothy K. Earle and Jonathon E. Ericson, pp. 15-35. Academic Press, New York.
Williams, Mark 1987 Archaeological Excavations at the Jackson Landing/Mulatto Bayou Earthwork. Archeological Report No. 19. Mississippi Department of Archives and History, Jackson.
Williams, Stephen, and Jeffrey P. Brain 1983 Excavations at the Lake George Site, Yazoo County, Mississippi, 1958-1960. Papers of the Peabody Museum of Archaeology and Ethnology, Vol. 74. Harvard University, Cambridge.
Wimberly, Steve B. 1960 Indian Pottery from Clarke County and Mobile County, Southern Alabama. Museum Paper 36. Alabama Museum of Natural History, University.
Wynn, Jack T. 1980 Cultural Resources Survey of Rocky Bayou Project, Yazoo County, Mississippi. Report submitted to the U.S. Army Corps of Engineers, Vicksburg District. Copy on file, Historic Preservation Division of the Mississippi Department of Archives and History, Jackson.
108
APPENDIX A
SAMPLE ELEMENTAL CONCENTRATION DATA
109
centration data. tal con n e m le ele p
Table A1. Sam
110
) continued (
Table A1
111
) continued (
Table A1
112
) continued (
Table A1
113
) continued (
Table A1 114
) continued (
Table A1 115
) continued (
Table A1 116
) continued (
Table A1 117
) continued ( Table A1
118
) continued ( Table A1
119
) continued (
Table A1
120
) continued (
Table A1
121
) continued (
Table A1 122
) continued ( Table A1
123
) continued (
Table A1
124
) continued (
Table A1 125
) continued (
Table A1 126
) continued (
Table A1 127
) continued (
Table A1 128
) continued (
Table A1 129