COESAM/PDER-87/004
inw r. c rnBf
AD-A201 878
Final Report
The Midden Mound Project
Judith A. Bense, Editor
Prepared under Contract Number I
DACVV01-84C-
For Mobile District, US. Army Corps of Engineers. OTIC
SELECT
OCT 2
^p*
!)
Office of Cultural and Archaeological Research
L •
Report of Investigations Number 6
88 021’
tCeumTV CLASSinCATtON of this page r»h«i Dim BnltnMO
t REPORT DOCUMENTATION PAGE
READ mSTRUenONS
BEFORE COMPLETIMG FORM
3. RECIPIENT’S CATALOG NUMBER
4. TITLE famf SuMItfaJ
THE MIDDEN MOUND PROJECT
B. TYPE OF REPORT A PERIOD COVERED
FINAL
6. PERFORMING ORG. REPORT NUMBER
7. AUTHONf^)
Judith A. Bense, Editor
a. CONTRACT OR GRANT NUMBERf*;
DACW01-84-C-
•. PERFOraUNC ORGANIZATION NAME AND ADDRESS
Office of Cultural and Archaeological Research
University of West Florida
Pensacola, Florida
10. program element, PROJECT. TASK
AREA a WORK UNIT NUMBERS
n. CONTROLLING OFFICE NAME AND ADDRESS
Environmental Resources Planning Section
US Army Corps of Engineers, Mobile District
PO Box Mobile, AL -
12. REPORT DATE
IS. NUMBER OF PAGES
431+xiii
U. monitoring agency name a ADORESSf// Al/terant Am ContnltbH OHIem)
same
IS. SECURITY CLASS, (ot Mo ropott)
Unclassified
lim. 1 jq MT Ml
OtSTRIBUnON STATEMENT (of tfll« Rmpoet)
Unlimited
17. OISTRISUTION STATEMENT (ot th« abalracl anlmYtf In Stock 20, M dHtoiont Inm Ropott)
Unlimited
IS. SUPPLEMENTARY NOTES
If. KEY WORDS fConflnu* on fovoroo old* U nocoooon* mnd idmnUtf by btock rtumbt)
TENNESSEE-TOMBIGBEE WATERWAY, TOMBIGBEE RIVER MULTI-RESOURCE DISTRICT, ALABAMA
MISSISSIPPI, PREHISTORIC ARCHEOLOGY, ARCHAIC, MIDDEN MOUNDS, ACCRETIONAL MOUNDS
BENTON, KIRK, SYKES, WHITE SPRINGS, ARCHAIC BURIALS, GEOARCHEOLOGY,
PALEOENVIRONMENT, SOILS, BIOSILICATES, POLLEN, CERAMICS, LITHICS
20. ARSTR ACT fCwWNin a rooormm .AM tt iimticcmr anN tdoatltp ky block number)
Final report of investigation of 11 sites in the River and Canal Sections of
the Tennessee-Tombigbee Waterway in northeast Mississippi. All work conducted
between January and December . The focus of this project was to obtal
an adequate sample of primarily Archaic Stage deposits from sites in the flood-
plain of the Upper Tomblgbee Valley. Samples were also obtained from single
component of the Middle Gulf Formatlonal and Late Woodland Stages, The project
Included detailed chemical and mlneraloglcal studies of site soils, geomorphol¬
ogy, botany and archaeomaenetlcs. (Continued on reverse.)
JCOIMTY CkAanriCATION or this PAOKOShM M*
20. Detailed archaeological modal analyses were performed on a sample of llthlc
and ceramics. Controlled replication and use wear experiments of Archaic
llthlc systems were also performed. The Early and Middle Archaic components
(ca. 10(000^5,000 B.P.) had well preserved features and middens and provide
one of the few large data bases for this time period in the Southeast. One
factor In this preservation was documented by the soil studies which
Identified a strong mineral bonding between the organics and the annually
deposited sand grains that excluded the sites from leaching and percolation
forces. The effects of the xerlc mld-Holocene climatic episode (Altlthermal)
were well expressed both geomorphologlcally and culturally. The cultural
response to this period was an intense nucleated settlement pattern in
large floodplain base camps and the pedologlcal response was a decrease in
the sedimentation rate and the formation of a well developed soli. The
Information produced In this project is integrated in this report in
two detailed models of the environment and the cultural adaptation to It
during the Archaic and Middle Gulf Formation Stages (ca, 10,000-2,000 B.P.)
In the Upper Tosdilgbee Valley.
SeCUHITr CLASSIFICATION OF THIS PAGEfVh*" Bnftmd)
COESAM/PDER-87/004
IHE MIDDEN MDUtO PBDOECT
Judith A. Bense, Editor
Produced under Contrac± Number nRCW01-84-C-
betwoen
Ihe U.S. Amy Corpe of Engineers, Mobile District
euid
The University of Mest Florida^
Office of Cultural and Archaeological Research
R^Torts of Investigations Nunfcer 6
May
ABsnwcr
THE MIDOQI MDtAlD PROJBIZT FINAL REPORT
J.A. SENSE, EDITOR
0-
This is the £inal report of oontract to investigate 11 sites in the
River and Canal Sections of the Tennessee-Tcnbigbee Waterway in aiitreme -
northeast Mississippi. It is a ocnprehensive account of gll thrpe dTases of
the investigations conducted betMeen January I960 and Deoenter r>tThe
focus of this project ‘was to obtain an adequate sasple of primarily Archaic
Stage deposits from sites in the flooc^lzdn in the Upper Ttnbiciiee Valley.
Samples were edso obtained from single component sites of the Middle Gulf
Formational and Late Woodland Stages. 'Rie project included detailed chemical
and mineralogical studies of site soils, geomarphology, botany and
axohaeomagnetics. Detailed archaeological modal analyses were performed on a
sanple of lithics and ceramics. Controlled replication and use-wear
ejqnriirieants of Archaic lithic systems were also performed. Ohe Early and
Mi^e'Arohaic components (ca. 10,000-5,000 B.P.) had well-preserved features
and middens and pravide on the the few large data bases for this time period
in the Southeast.' - One factor in this preservation was docunented by the soil
studies %hidi identified a strong mineral bonding between the organics and the
annually deposited gand grains that exclxided the sites firom leaching and
percolation forces. |^^The effects of the xeric ndd-Holooene climatic episode
(Altithermal) was well esqiressed both gecmorphologically and culturally. The
cultural response to . ^lis period was an intense nucleated settlement pattern
in large floodplain b^ camps and the pedological respoaise vus a decrease in
sedimentation rate and the formation of a well-developed soil. The
information produced ixy this project is integrated in this report in two
detailed models of the enviroranent and the cultural adaptation to it during
the Archaic and Middle (^f Formational Stages (ca. 10,000 - 2,000 B.P.) in
the Upper Tcnbigbee Valley.
coNmns
LEST OF FIGURES
T.Tgr OF TABLES
LEST OF AFFQDIOES
OHKPTER I INmSUCnON
Jtdith A. Sense
PRXIECT EESCRIPriOW
Lk:i3»ci:^yui^
PHYSICAL EHVHOtiBn’
Jtxiitii A. Sense and D. E. Pettxy
PAST aXDinOMS
PRESENT OOMSmONS
E.S. Sheldon, D.E. Pettry, and J.A. Sense
CULTURAL EMVIRaMaiT
PALBO-INDIAN SIAES
J.A. Sense
ARCHAIC STA3E
J.A. Sense
OUEF FQRHATIONAL SEAGE
D. t)ye, S.H. Ensor, and J.A. Sense
WXDLAMD STAGE
D. Dye, S.H. Ensor, and J.A. Sense
MISSISSIPPIAN STAGE
D. Dye, S.H. Ensor, and J.A. Sense
SUMARY
CHAPTER III RESEARCH DESIGN
INITIAL RESEARCH DESIGN EEVELOtMBTO
J.A. Sense, D. Dye, S.H. Ensor, and E. Futato
DESIGN
MODEL
D.E. Pettxy and J.A. Sense
ARCHABGILOGICAL MODEL
J.A. Sense
OORREUVnON OF PAEEOCNVIRONCNT AM)
ARCHABOIOGICAL/CULTURAL PATIEEN
J.A. Sense
APraESSABLE CULIURAL PROCESSES
J.A. Sense
OTHER ACraESSABLE CULTURAL ISSUES
AEDRESSAB[£ EMVIROMaNIAL PROCESSES
HYPOTHESIS BEPMffiNT
J.A. Sense and R.L. Lurie
SUMARY OF RESEARCH I^IGN
OONUNTS
(continued)
CHAPTER IV ARCHAEX3L0GICAL PROCEDURES 33
FTRrn PRDCHXIRES 33
J.A. Sense and M.J. Rodeffer
LABORAapRY PROCEDURES 36
J.A. Sense and E. Futato
CHAPTER V SITE EXCAVATIONS 39
J.A. Sense, C.H. Lee, and Staff
INTBODOCnON 39
EXCAVATIONS AT THE POPIAR SITE (22It576) 39
EXCAVATIC»1S AT THE WAUATT SHE (22It539) 54
EXCAVAHCNS AT THE HEX SITE (22It590) 74
EXCAVATIONS AT THE HICKORY SHE (22It621) 86
EXCAVATIONS AT THE EEBCH AND OAK SITES
(22It623/22It624) 96
EXCAVATIONS AT THE ARALIA SITE (22It563) 107
EXCAVATIONS AT SITE 22IT606 116
EXCAVATIONS AT THE MID CREEK SITE (22lt622) 124
EXCAVATIONS AT THE SMILAX SITE (22Mo675) 129
EXCAVATIONS AT THE DOGNDOD MXJND SITE
(22M)531) 134
SUMARY OF EXCAVATIONS 141
CHAPTER VI SOILS AND GBOMDRPHODOGY 143
D.E. Pettry
nmODOCTION 143
MBmPDOLOGY 143
THE WALNUT SITE (22It539) 144
THE POPIAR SITE (22lt576) 158
THE HEX SHE (22It590) 173
SITE 22It606 180
THE MUD CREEK SITE (22It622) 187
THE BEECH AND OAK SITES (22lt623/624) 189
PEIXXZNIC INFERENCES 193
SOH DEVELOPMENT IN ARCHABOLOGICAL SITES 193
GEOMORPHIC SETTING AND SOH PARENT MATERIAL 194
ANTHROPIC EPIPEDONS 196
PAIBOSOLS 199
SUWARY EOR ARCHAEOLOGISTS 201
J.A. Sense
DISCISSION AM) INTERPRETATION OF THE SOH ANALYSIS 205
CHAPTER VII BOTANICAL STUDIES 209
VEXarCATIONAL HISTOTY ^ MID-SOLTIH 209
E.S. Sheldon
BOTANICAL ANALYSIS OT ARCHAEOLOGICAL MATERIAL 210
E.S. Sheldon
SUM1ARY OF MACROBOTANICAL INFORMATION
224
LIST OF FIGURES
Nisiber Page
1 Regional map vdth location of archaeological sites investigated. 2
2 Regional vegetation map. 8
3 Topogreqphic msp and excavation plan, 22It576. 40
4 Stra:^graphic profile of Block D, 22It576. 49
5 Stratigraphic cross-eecticns, 22It576. 50
6 Topographic map and excavation plan, 22It539. 55
7 Prepared Area: Feature 120 in Block C, 22It539. 63
8 Plan of burials in Blocks A and D, 22It539. 64
9 Burial 1 in Block A, 22lt539. 65
10 Cremation with stcxie effigy beads in Block D, 22It539. 65
11 Stratigraphic profile of Block D, 22It539. 67
12 Topogra^Mc map and excavation plan, 22It590. 75
13 Prepared Area: Feature 22, 22It590. 80
14 Correlation of stratigrap^c pxjfiles, 22It590. 82
15 Topographic map and excavation plan, 22It621. 88
16 Stratigraphic profiles, 22It621. 93
17 Stratigraphy of Blocks D and E, 22It621. 94
18 Topographic map and excavation plan, 22lt623 and 22It624. 98
19 Stratigraphy of Blocks A and B, 22It623. 103
20 Stratigrapt^ of Blocks A and B, 22It624. 104
21 Topographic nep and excavation plan, 22lt563. 108
22 Stratigra^iiic profile of Block A, 22It563. 114
23 Topographic map and excavation plan, 22It606. 117
24 Stratigraphic profiles, 22lt606. 122
25 Topographic map and excavaticn plan, 22It622. 125
26 Stratigraphic profile of block C, 22It622. 128
27 Topographic map and excavation plan, 22Mo675. 130
28 Stratigra^^c profile, 22Mo675. 133
29 Topogra^^c map and excavation plan, 22No531. 135
30 Stratigra^ic profile, 22Mo531. 140
31 Soils in the vicinity of 22It539. 145
32 Constant sand fabric, 22It539. 152
33 Soils in the vicinity of 22It576. 159
34 Soil profile, 22It576. 161
35 Constant sand fabric, 22It576. 166
36 Soils in the vicinity of 22It590. 174
37 Site location of 22It606 and 22lt622. 181
38 Model of medium-sized cobble reduction sequence. 249
39 Model of small-sized cobble reduction sequence. 250
40 Model of Fort Payne reduction sequence, flake reduction
sequence and bipolar reduction sequence. 251
41 Flake blank for Greenbriar biface manufaKrture. 256
42 First thinning of flake blank for Greenbriar biface. 256
43 Second thinning of flake blank for Grea^riar biface. 257
44 Trinming of flake blank for Greenbriar bifaoe. 257
45 Finished Qreenbriar bifaoe. 258
46 Hafted end-scraper vised in use-wear eiqjeriments. 268
47 Hafted drill used in use-wear experiments. 268
LIST OF FIGURES
48 Diversity and evenness gra{:h for lithic technological class
and time category. 316
49 Distiributlon of flakes fron cobbles and refit pieces. 340
50 Geranlc types at 22It563. 353
51 Bar chart of patterns of Alexander Incised. 355
52 Bar chart of patterns of Alexander Pinched. 356
53 Bar chart of mode of punctaticn of oerandc type
Alexander Pinched. 359
54 SAS "i^ike" plots of counted artifacts. Level 11,
Block C, 22It539. 388
55a Contaur maps of counted specimens in Levels 9-12,
Block C, 22It539. 389
55b OontGur waps of counted specimens in Level 13, Block C,
22It539. 390
57 Trend surface maps of all counted material. Levels 9-13,
22It539. 392
57 Contour map of projectile point/knives. Levels 9 -12,
Block C, 22It539. 394
58 Sequence of events of floodplain midden mound sites. 395
V
LIST OF TABLES
Nmber Page
1 Hoo^ species ocnpositicn of the fonnal vegetation plots,
22It590 and 22It576. 9
2 Oircnological chart for the Ojpper Tbnbi^Pee Valley. 11
3 Gerandc frequencies by ten|)er, 221t576. 42
4 Chipped stone frequencies by type, 22It576. 43
5 Size-grade frequencies of ddsitage, 22It576. 44
6 Ground stane tool frequencies by type, 22It576. 44
7 Burial analysis, 22It576. 47
8 Radiocarbon and archaeoniagnetic dates, 22It576. 48
9 Ceramic frequencies by tenper, 22It539. 57
10 Chipped stone tool frequencies by type, 22It539. 58
11 Size-grade frequencies of debitage, 22It539. 58
12 Ground stone tool frequencies by type, 22It539. 59
13 Frequencies of species of faunal remains, 22It539. 60
14 Burial analysis, 22It539. 62
15 Radiocarbon dates, 22It539. 66
16 Archaecmagnetic ard correlating radiocarbon dates, 22It539. 69
17 Ceramic frequencies by tenper, 22It590. 76
18 Chipped stone tool frequencies by type, 22It590. 77
19 Size-grade frequencies of d^ltage, 22It590. 78
20 Ground stone tool frequencies by t]^, 22It590. 78
21 Radiocarbon dates, 22It590. 83
22 Ceramic frequencies by tenper, 22It621. 89
23 Chipped stone tool frequencies by type, 22It621. 90
24 Size*-grade frequencies of debitage, 22It621. 91
25 Ground stone tool frequencies by type, 22It621. 91
26 Ceramic frequencies by temper, 22It623 and 22It624. 99
27 Chipped stone tool frequencies by type, 22lt623 and 22It624. 99
28 Size-grade frecpjencies of d^ita^, 22It623 and 22It624. 100
29 Ground stone tool frequencies by t^pe, 22It623 and 22It624. 101
30 Radiocarbon dates, 22It623 and 22It624. 105
31 Ceramic frequencies by tenper, 22It563. 110
32 Chipped stone tool frequencies by type, 22It563. 110
33 Size-grade frecjuencies of debitage, 22It563. Ill
34 Ground stone tool frequencies by type, 22It563. Ill
35 Ceramic frequencies by tenper, 22It606. 118
36 Chipped stme tool frecjuencies by type, 22lt606. 119
37 Ground stone tool frequencies by type, 22It606. 120
38 Radiocaudxn dates, 22It606. 121
39 Ceramic frec[uencies by tenper, 22It622. 126
40 Chipped stone tool frequencies by type, 22It622. 126
41 Size-grade frequencies of debitage, 221t622. 127
42 Ceramic frequencies by tenper, 22Mo675. 131
43 Chipped stone tool frequencies by type, 22Mo675. 131
44 Size-grade frequencies of d^itage, 22Mo675. 132
45 Ceramic frequencies by tenper, 22Mo531. 136
LIST OF 13^BLES
(cGntinued)
NuDDber
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
Chipped stone tool frequencies by type, 22Mo531.
Size-grade firequencies of debitage, 22^531.
Classification of soils fron the site 2und vicinity, 221t539.
Mansell color of selected horizons of representative soils
in the adjacent floodplain, 22It539.
Pedon description of representative profile, 22It539.
Peurticle size distribution of selected soil sarnples, 221t539.
Particle size distribution of typical pedon from Block A,
22It539.
Peurticle size analyses ard pH of soils adjacent to 221t539.
Chendcal characteristics of representative pedon, 22It539.
Organic natter, free iron oxides, total P and 1% citric acid
extractable P^O,. levels of selected pedon, 22It539.
Organic natter, pHT carbon, nitrogen contents, caribon/nitrogen
ratio, and 1% citric acid soluble PjO^ contents of
representative pedon. Block A, 22It539.
Color ratio vedues of NaCn extracts of pedon from Block A,
22It539.
Classification of soils ocnprising Site 22It576 cind vicinity.
Munsell color or selected horizons of rqiresentative soils in
the floodplain adjacent to Site 22It539.
Pedon description of representative profile, 221t576.
Particle description of representative pedon, 22lt576.
Particle size distribution, pH, and organic natter contents of
representative soils adjacent to Site 22lt576.
Chendcal characteristics of representative pedon, 22It576.
Organic natter, carbon, nitrogen contents, carbon/nitrogen
ratio, and 1% citric acid soluble contents of
representative pedon of Block D, 22It576.
Organic natter and ^65^3 typical pedon, 22It576.
Soil phosphorous fractdro of representative profile front
Block D, 22It576.
Soil pho^horus fractions percentage of total phosphorous
cxxTtent of representative profile frcm Blcxk D, 22It576.
Color- ratio values of NaOT extracts of representative pedon
from Block D, 22It576.
Pedon description of representative profile, 22It590.
Particle size distributicm of selecrted soil sanples
representative, 22It590.
Peurticle size distribution of representative soil pedon at
the edge of Site 22It590.
Chemical charac:teristics of representati-ve peckm, 22It590.
list of tables
(cxDntiniied)
Number Page
73 Chanical characteristics of representative soil pedon at the
edge of Site 22It590. 178
74 Organic matter, carbon, nitrogen contents, carbon/nitrogen ratio,
and 1% citric acid soluble P^O_ contents of representative
pedoi Block A-Z, 22It590. ^ 179
75 Organic matter, free iron oxide, total and organic phosphorus
contents and of r^resentative pedon, 22It590. 179
76 Organic matter, carbon, total nitrogen contents, and carbon/
nitrogen ratio of representative pedon, 22It590. 179
77 Pedon description of representative pedon, 22It606. 182
78 Particle size distribution of rqpresentative pedon, Test Unit
106/94, 22It606. 184
79 Particle size distribution of representative pedon, 22It606. 184
80 Chemiced. characteristics of representative pedon. Test Uiit
106/94, 22It606. 185
81 G^emiceil characteristics of representative pedon, 22It606.
82 Organic carbon, nitrogen, caxbon/nitrogen ratio, and 1% citric acid
solxible P^Oc contents of representative pedon, 22It606. 186
83 Particle size Retribution of representative ped^, 22It622. 188
84 Chendc^al characteristics of representative pedon, 22It622. 188
85 Organic matter, carbon, nitrogen contents, ceodton/nitxogen
ratio, and 1% citric acid soluble cxntents of
representative pedcan, 22It622. 189
86 Descnripticn of representative profile, 22It623 and 22It624. 190
87 Particle size distribution of representative pedcxi, 22It623. 191
88 Chemical characteristics of representative pedon, 22It623. 192
89 Organic matter, carbon, nitrogaa contents, carbon/nitrogen
ratio, and 1% citric acid soluble P-O^ cxantents of
representative pedon, 22It623. 192
90 Sunmary of sedimentation rates by cultural cxirponent and
elevation. 195
91 Identified plant remains totaled ty site. 211
92 Floral remains by provenience by site. 213
93 Results of Phase III pollen analysis. 228
94 Phase III midden and feature units. 242
95 Number of items produced in redixrtion experiments. 252
96 Tools procJucxed in the experiment program. 252
97 Mean dimensions of original cdhles and bifacas producad at
stages in reducrtion secjuencas. 259
98 Mean length, width, thickness, and weight for vhole named
bifacas. 265
99 Nurtber of tools and edges used in use-wear expariments. 269
100 Cdiparison of measurement clusters and traditionally named
bifaca types.
293
LIST OF TABLES
(ccHitinued)
Nunber
101 Cross-tabulation of traditionally naned biface types and
time category.
102 Cross-tabulation of traditionally named bifaoe types euid
haft type.
103 Cross-tabulation of haft type and time category for
mor^diological classes 1 through 16.
104 Cross-tabulatic»i of haft type and time category for
morphological classes 1 through 9.
105 Cross-tabulation of moip^logical clcisses 1 through 9 and
cross-section.
106 Cross-tabulation of biface morphological class and pressure
flaking or resharpening pattern.
107 NiXDber of different functions recorded for traditionally named
hafted biface types.
108 Hafted bifaoe tool function by morphological class.
109 Debitage siximary - chert vs. non-chert.
110 Chert debitage sunmary by site, block, level, time category,
and raw material type.
111 Artifact raw material by time category.
112 Ttool chert quadity by time category.
113 Cross-tabulation quality of Fort Payne chert by collapsed
Archaic period.
114 Cross-tabulaticMi of quality of local material by collapsed
Archadc period.
115 Heated and unheated Camden and Tuscedoosa chert by size, site,
and time category.
116 Hatio of heated and unheated Tusccdoosa and Caniden d^itage by
site, time category, and size.
117 Artifact heat treatment by time category.
118 Artifact heat edteration by time category.
119 Diversity indices for technological class by site and time
category.
120 Chipped stone artifact: form time category.
121 Ground stone artifacts time category.
122 Artifact production stage by time category.
123 Nixnber of tool func:tions by site and time category.
124 Tool function by time category and site.
125 Artifact ccnpleteness by time category.
126 Artifact use stage by time category.
127 Frequency of fracture types far the entire sarnple.
128 Raw material quedity and eurtifact form - Fort Payne chert.
129 Cross-tabulation of raw itaterial and artifact form - lcx:al
raw material.
130 Combined fracture types by site and time category.
131 Tool and i.tage sumazy - Phase I and II data. 131
15 Horphologiced. class by time category and site. 132-137
16 Cross-tabulation of all ra»r material quality and
New Time Categories. 138
17 Quality of local material. 139
18 Cross-tabulation of Port Payne quality and New Time
Categories. 140
19 Cross-tabulation of local raw material equality and artifac:t
fcjrm by time category - Early Archaic. 141
20 Cross-tabulation of Icxal raw material equality and artifact
form by time category - Middle Archaic.
Peqeis)
58-62
63-68
69-70
71-72
73
74-78
79-80
81-83
84
85-90
91-92
142
LIST OF APPENDICES
(oontinued)
APPQDIX III
LmHC DKEA
(continued)
e Page(s)
Cross-tabulaticn of local raw material quality and artifact
fcnrni by time category - Late Archaic. 143
Cross-tabulation of heat treatment and Hest Time Categories -
all raw material. 144
Cross-tabulation of heat treatment and New Time Categories -
local raw material. 145
Cross-tabulation of heat treatmMit and New Time Categories -
Fort Payne chert. 146
Cross-tabulation of raw material and tool production by
New Time Categories - Early Archaic. 147-148
Cross-tabulation of raw material and tool production by
New Time Categories - Middle Archaic. 149-150
Cross-tabulation of raw material and tool production by
New Time Categories - Late Archaic. 151-152
Technological class by New Time Category. 153
Nunher and percent of hafbed tools by New Time Category. 154
Sumnary statistics for artifact form by time category. 155
Cross-tabulation of multipurpose tools and New Time Categories
odl raw material. 156
Cross-tabulation of multipurpose tools and New Time Categories
local chert. 157
Cross-tabulation of multipurpose tools and New Time Categories
Port Payne chert. 158
Cross-tabulation of total functional units and New Time
Categories - all raw material. 159
Raw material quality and artifact form - Fort Payns chert. 160-161
Nvitber and percent of vhole and brcdcen pieces by New Time
Category. 162
Cross-tabulation of heat alteraticsn stages and New Time
Categories - all raw material. 163
Cross-tabulation of heat alteraticxi stages and New Time
Categories - Icxal raw material. 164
Cross-tabulation of heat cilteration stages and New Time
Categories - Fort Payne chert.
165
LIST OF APFEMDICES
(ocxitinvied)
AFPBDIX IV
CERAKIC AinLVSIS
^ Pacie(s)
Cerandc attribute analysis. 168-174
Results of Phase III oeranic analysis, 22It563. 175-184
Results of Ftuise III oerandc analysis, 22It606. 185-202
Results of Phase III oezamic analysis, 22Mo531. 203-207
Alexander Incised descriptions of surface treatment
and analysis.
208-244
CHAFEER I nniODUCTIGM
This Is the fiml report of the nordieastem Ifississippi Tennessee-
Tcnbiic^iee Nsterway archaeological investigatlcns conducted betMeen January,
and Deoester, by the Office of Cultural and Archaeological Researdi
of the Qiiversity of West Florida. The tuoric was perfonned in three phases
under contract to the U.S. Amy Corps of Engineers, Mobile District. Interim
reports were written for Phase I (Dense ) and Phase II (White ) .
This is the final report for the project, and infoamation from all pha^ of
the project is presented. The archaeolo^ceJ. investigations consisted of
data recovery of eic^ sites and testing of seven sites between Aberdeen and
Iran's Well, Ms (the Aberdeen Pool throu^ Lock E of the Canal Section of the
Tennesaoc-Tcmbi^aee Waterway) .
The U.S. Amy Corps of Ehgineers was authorized to construct a navigable
waterway between the Tennessee and Tcmbi^bee Rivers vAien Congress passed the
Rivers and Harbors Act of . IVxenty-six years of planning and study
elapsed before construction on the Gainesville Lock and Dam began in .
The waterway (Figure 1) extends north from its southern terminus at
Demopolis, A1 to the East Fork of the Tombi^bee River. It proceeds
Macke's Creek (a tributary of the East Fork of the Tcnibigbee River) , through
Bay firings, over the divide separating the Tcmbi^see and Tennessee River
drainages, and then debouches into Yellow Creek, Which is part of the
Pickwidc Landing Reservoir near the ccinion boundary of AlzJaama, Mississippi,
and Tennessee.
The waterway project has three sections. The southemnost is the River
Sectloh. It consists of the Tcnbig^Dee River made navigable by widening
riverbanks, cutting throu^ oodxws and narrow trends, and oonstructiiig four
artificial lakes, each with a lock-and-dam ccnplex. The next portion of the
waterwey is the Canal Section, a 91 m (300 ft) wide excavation rou^y
paralleling the East Fork of the Totdsigbee River controlled through a series
of five lock-and-dams. The last section is the Divide-Cut, a 43 km (26.7 mi)
oanal through the ridge that divides the Tcnbi^bee Valley from the Tennessee
Valley, with the Bey Springs Lock and Dam at the southern terminus. The
mmai attains a maxiitun depth of 53 m (175 ft) at the peak of the divide.
The northern terminus of the waterway flows into the Yellow Creek, a
tributary of the Tennessee River vhich provides access to the Tennessee River
System.
Management of the archaeologiced. resources in the Tennessee-Tcnbic^bee
Waterway began with the National Park Service survey of the proposed
Gainesville Lake Section. The first reconnaissance surveys covered large
areas in a short period of time. For exanple, the first survey in the
Mississippi portion of the waterway covered a 120 mi (193 km) route from the
Alabama border to the Tennessee River and was conducted in less than two
montlis (McGahey ; Lewis and Caldwell ). Later surv^ divided the
Mississippi portion of the River Section into smaller units vhich permitted a
more detailed investigation (Rucker ; Blakaman , ) . However, the
problems of severe time limitations, large stufy areas, and poorly defin^
boundaries continued. This situation was largely repeated in the Alabama
portion of the River Section with an initial cursory surv^ of the
Gainesville Lake area (Lewis and Caldwell ) , vAiich was followed ly a
limited series of excavations at large sites (Jenkins ; Nielsen and
Moorehead ; Nielsen and Jenkins ; Peebles ) . The problems of
inconsistent surveying were recognized vhen previously neglected portions of
the waterway were surveyed in mid- (Atkinson and Elliott ) , and maiy
new sites were found in the River Section.
1
Mississippi ij ^Alabama
ALCORN
TISHO-i'*
1 M I N G o I / “/Ve/- •
I ^
\ jy
/ i il^E^^2IT590
(p ^^7^^221X539
/ .22IT622CSfv>22IT621
^ / ! Loek^ ^221X563
/ 221X623— »
BoonevOleJ / \ \ / 1* !/
/ WA~~--
PRENT.Ss/y// f
_ i Say Sprino* L4D Russellvilte X
WINSTON
TUSCALOOSA
Tuscaloosa r
Figure 1. Regional map with location of archaeological sites investigated
Construction on the waterway progressed r^idly, and by F^ruary of
it was apparent that a mechanism had to be created vAiich integrated previous
archaeological findings into a mitigation program ce^>able of being carried out
within the established construction schedules. Almost 700 archaeological
sites had been disoovered within the limits of the waterway by . It was
thmi agreed that the establishnent of a Nationed. Register District was the
only feasible way to manage the diverse archaeological resources present
within the waterway that had not been oovered by previous individual
memoranda. The Itambigbee River Multi-Resource District was defined and
enooqpasses cdmost all of the waterway in a oorridor 8 km (4.96 mi) wide and
approodmately 280 km (175 mi) long from Gcdnesville, Al to Paden, Ms. It was
declared eligible for the National Register of Historic Places on Sepbocber
21, .
Hie four-stage mitigation plan for the district had two separate
strategies: preservation and data recovery and it addressed 1) problems of
survey bias, 2) evaluation of site significance and information, 3) excavation
priorities, and 4) previously investigated sites with insufficient information
available.
i^ROJBCT EESCRIPTION
The Midden Mound project reported here developed as a result of a lairge
testing project in the River and Canal Sections (Sense ) . Testing
documented that in the Upper Tonbigbee Valley (UIV) intact Archedc, Gulf
Farmational, and woodland deposits were present in the impact zone both in the
floodplain and on the edge of the first terrace. Most of the significant
sites contained thick, organically stained midden d^)osits that had a
mound-shaped cross-section, locally referred to as "midden mounds." Hiis
project was designed to recover data from eleven sites with Archaic, Gulf
Formaticnal, and WOodland deposits in and near the flooc^lain of the Ipper
Tombi^bee Valley. Prior to this time, little or no information had been
retrieved on the Archaic and Gulf Fornational stages in previous ocrplianoe
work in the waterway. These sites were the focus of this project.
The Midden Mound project had three phases: Phase I (-) initial
research design and data recovery and preliminary analysis of material; Phase
II (-) refinanent of the research design and data recovery from
specific ch3:onological periods; and Eheise III (-) the fin2d refinanent
of the research hypotheses, analysis of specific data sets, and preparation of
the fin2d report.
Construction schedule of the waterway and the location and configuration
of the sites investigated were influential factors in the project
organization. The construction schedule dictated that at least two sites had
to be excavated simultaneously in Phase I. Concurrently, field laboratories
and data management systems had to be established. Although refinements were
introduced for Fhcise II, the same systems of field, laboratory, cuid data
management were used to insure ocnparability.
The project has had one Principal Investigator, however most staff members
changed with the three phases of this seven-year project. The project
included a group of consultants in soil morphology, fluviad gecmorphology,
botany (including macro and micrdbotanics) , archaeozoology, archaeometry,
physical anthropology, and archaeology. The Pheise I staff was the largest
(75-80 persons) . During Phases 1 cuid II headquarters was in Fulton, Ms; Fheise
III was conducted at the university of West Florida campus in Pensacola, FI.
3
-'T
This g^xart is the final isfxurt of the Midden Mound project. It
supersedes the interim reports of the {anvious two phases of the project
(Bense ; Nhite ) . The interim reports, however, contain the bulk of
data recovered fron the sites and druid be used for referenoe purposes,
ram data are stored on ocmputer tape vhich can be ordered from the
versity of Nest Flralda. Susmary tables of the data (artifacts) in each
svatlon block and level are available from the Qniversity of Nest Florida
nicrofiche.
4
CHAPTEIR II THE SETTING
MKSICAL EMyiBOMENT
PAST OONDITICKS
GBOKXIPXSUXSI
stuc^ area is lcx:ated in the Tcmbigbee Hills region of the Gulf
Coastal Plain. Ihe Tcmbigbee Hills that lie within the strdy area are
ocnfirised of vmcxiisolidated marine sediments of l^:per Cretaceous age. Hie
Eutaw and Tuscaloosa formations outcrop in the area and provide the parent
materud for the i^iland soils and edluvial deposits (St^henson and Monroe
) . Hie Tuscaloosa formation is characterized by irregularly bedded sand,
clay, and gravel, while the Eutaw formation is generally ocnfirised of
cross-bedded glauconitic sand and clay. The soils and sediments of the
Tcmbigbee Hills have been eroded and redeposited on the Tcmbi^see River
floodplain during the Pleistocene and perhaps late Pliocene time (Stephenson
and Mcnroe ) . Current Holooene sediments in the active flcxx^lain are
heterogeneous and related to current erosion and deposition processes.
Muto and Gunn () described the gecnorphology of the l^per Tcmbigbee
Valley. Ihe valley Iseg^ forming during late Tertiary times (ca. 30 million
years ago) after iDeing tjplifted as part of eustatic rebound incident to the
recession of the Cretaceaous seas. Continued relative tplift during the
Pliocene and Pleistocene resulted in one Plio-Pleistooene terracx and four
Pleistoc»ie terraces. During the developnent of the valley terraces, the
river channel has generally migrated to the viest in response to the dip in the
imder lying sediments.
During the Pleistocene epcx:h, depositional and erosional cycles were
related primarily to glacial and interglsKzial conditions. Erosion (xxnirred
during the later parts of interglacial ststges as sea level fell and during
glacial periods of low sea level. Depositional cycles occurred during early
and middle interglacial stages when sea levels were hi^. Hie four cycles of
erosion and deposition developed in siKXjessively lower river levels prcxJucing
the four Pleistocene terraces of the Tcmbi^see Valley. Hiese terrac^es are
usually well dissected and cxxiposed of mixed alluvial sands and gravels.
Finer-grained materials oocxir only locally and are asscx;iated with relic
oodxws.
During the Holoc^ene, terrac^e formation has cxxitinued, and two terraces in
the floodplain have been identified. Hie highest is the Early Holcxsene
terraice and cxxxirs between 1-7 m (3.3-23 ft) above the channel. Hie Late
Holooene terrace is the lowest and ocjcurs between 1-3 m (3. 3-9. 9 ft) alxve the
channel. In the Upper Tcmbi^iee Valley, north of smithville, Ms, most of the
Early Holooene terrace deposits appear to have been eroded and reworioed with
only small remnants remaining. The Late Holooene terrace is present
throughout the valley and is usually dominated by fan deposits &cm
hi^i-gradient side-streams.
Comparisons of the Holooene terraces indicate that the Tcmbi^oee River has
not changed significantly during the Holooene. The entire floodpladn area
appears to have renmned essentially the sane during this period.
5
PALBOSCiEf
Within the Holocene terraces Muto and Gunn () and vork associated with
this project have identified three paleosols baaed on the alluvial chronology
and soils associated with dated archaeological materials. Ihese Early,
Middle, and Late Holoc^ene soils formed in overbank and bar d^x>sits. The
Early and Middle Holocsene soils have often been eroded and are usually buried.
Formation of the Early Holcx^ie soil began in the Early Holcxsene and persisted
until approximately 7,000 years ago. At that tiice it was either buried (south
of Colvmbus, Ms) , eroded (north of Coltxnbus, Ms) , or slightly eroded and
overlain by fluvial sediments in which pedogenesis cxsntinued. Formation of
the Middle Holcx:ene soil persisted until ^proximately 3,000 years ago. In
seme instances the Middle and Early Holocene soils form a bisequem, the lower
element of vhich is the degraded Early Holocene soil B horizon. The Late
Holcxiene soil consists of modem {post-3,000 B.P.) depcssits and is poorly
dredned and organically rich.
PRESENT CONDITIONS
The present topcigrapl^ of the Upper Tombi^iee Valley ranges fran nearly
level in the flcxx^lains and terraces to steep in the adjoining uplands.
Elevations range fran 75 m (250 ft) NGVD in the flcxx^lain to 122 m (406 ft)
and greater in the uplands. The flcxadplain cxnmonly excaeds a width of 1.5 km
(0.93 mi) and cxsntains numerous meandering sloughs, abandoned river cutoffs,
and stjaams entering from the uplands. The ac±ive river channel is generally
Icxated in the western part of the flcxx^lain.
The well-disseched uplands bounding the flcxx^lain have steep-sided slcpes
with narrewr ridges and valleys. The streams are deeply incised and form a
dendritic drainage pattern with a relatively low entrance angle into the
TOnbic^see flcodplain. The tributary valleys in the Upper Tcnrbigbee Valley are
usually long and narrow and have steeply slopmg valley walls. Ground slcpe
in the flcxx^leiins is usually gentle eurid often almost nil.
The flcxx^ladn consists of the flocxibasin and Holocene terraces. The
gecmorphic units of the floodbasin include channels, chute cjutoffs, point
bars, levees, splays, marshes, csdxws, and undifferentiated flcxxlb^in areas.
The Holcx:ene terraces include lcx:al fans, levees, fan veneers, and colluvial
units, all of vhich can also cxxur in the flcxdbeisin. The terrace is a
periodically flcxsded depositioisd surface and actually defines the limits of
the flcxx3^1ain. The higher portions of the Holocene terrace are only affected
by hi^-magnitude flcxids (hundred-year intervals) and are semi-relic surfaces.
Hilgard () called the stxjdy area the Northeast Prairie Region. The
prevalent forest trees in antdellum times was shortleaf pile (Pinus
echinata) , blackjack oak (Querces marilandica) , post oak (Q. stellata) , and
chestnut (Castanea dentataTI The narrow bottom of Mackey's Creek and its
gentle slopes possessed a forest of Spanish oak (Q. falc:ata) , other oak
species, and hickories (Carya sp.) , tut lacked pines. East of the Tatbigbee
flcxx^lain, the l^uld surface is very broken with the coarse red-orange soil of
the Thscalcosa and Eutaw formations. South of Fulton, Ms, red loam soil is
more frequent and is exavered by Icurge scerlet oa)cs (Q. eexoinea) , occasional
black oaks (Q. velutina) and white oaks (Q. alba) , as well as hickory and
shortleaf pine.
Just north of Smithville, Ms begins a lew first terrace or high Eeurly
Holocene terrace 3. 2-9. 6 km (2. 0-5. 9 mi) wide bordering the river. This area
is heavily fanned today. In the vegetation consisted largely of Ichlolly
^ igranuB oociaBnTaxxsi , oeeui \raquB sp.j , rxvei: me^ue sp.) , ouacn. gun
(llyssa syivatica) , sweet gun (Liqmdantoar styryiflua) , and cypress.
Associated species included haddierry, recBaud (Cercis can^tensis) ,
great-leaved magnolia (Magnolia maojoi^l^) , silv^sell (Halesia Carolina) ,
storax (St^MX sp. ) , paw-paw (Aslmlna tribola) , and red birdi (Betaila sp. ) . A
nunnter of these rea^ their southern limit of distribution in Mississippi.
In a more recent study^, Zary () found aefveral forest t^pes
intermingled %«ithin Itamosba County. Ihe slopes are occupied by oak-hickory,
oak-pine, and loblolly^shartleaf pine forests. The bottomlands of the
Ibnbigbee River and Mackey's Creek eacB covered by a hardwood forest ocnposed
of tupelo and black guns, sweet gun, oak, and cypress. Ccmaxi trees
associated in this forest included willow, a^, elm (Ulmas sp.) , hackberry,
ma^le, and cottonwood (Populus deltoides) . An ash-elm-oottonwood forest is
intermingled with the alxr^ typ^ Ttiis association includes willow
(Salix sp.) , sycamore, beech, and maple.
To reconstruct the vegetaticnal history of the stuc^ area, flora near four
major archaeological sites wes studied. Plant ocnraunities exanined
quantitatively included a steep oak-hickxsry slope near 22It563 and a
floo^lain levee near 22It590 (Figure 2) .
Plots of 100x100 m (333 ft) were established in areas of honogeneous
vegetation. Eerh tree greater than 0.1 m (3.9 ft) diameter at breast height
(DBH) was recorded. Woody and herbaceous plants less than 0.1 m (3.9 ft) DBH
were counted by taxon over the entire vpland plot and on a strip 30x100 m
(100x333 ft) at the eastern end of the levee plot. Foonal plots were not
established on or near 22It539 or 22It576, because most of the vegetation had
been cut on the sites and the surrounding floodplain levee before the
botanical survey was initiated. However, a list of the ronaining taxa on
these sites was oonpiled.
Ihe woody species cotposition of the two plots is presented in Table 1.
Oak, hickory, sweet gun, red maple, and dogwood were found in both
ocmnunities. Gum, beech, elm, tulip poplar (Liriodendron tulipifera) ,
ironwDod (Carpinus sp.) , and hop-hombeam (Ostrya ^.) appeared only in the
floodplain ocmninity; vAiereas, edder (Alros) , hacldjerry, birdi, sassafras
(Sassafras edbidun) , and sunac (Bhus ^prT~were confined to the upland plot.
The presence of species recognized as sucoessional, i.e., red maple, elm,
tulip poplar, hackberry, and birch, and the large percentage of trees less
than 0.1 m (3.9 ft) DBH is indicative of ooranunities 25-50 years into
secondary succession. The large nuiisers of red maple and hop-honheam trees
were probably the offering of particularly well-adapted individuals or
populaticns.
Although the vegetation which developed in the first one to ten years
after disturbeuice of a habitat nay not be typical of the original climax
forest, later sucoessional stages resemble those vhich were present in the
7
REGIONAL VEGETATION MAP
Fiv^urf »anses of bottanland hardwoods with intermittent codow lakes and streams
provide anple cover for a great variety of species. Deer peculation is
estimated at one per acres. Squirrel populations are high with an
estimated three per acre. Rabbit and quail pqpulations in the area are good,
with the quail being restricted to the more open curecis. Furbearers such as
beaver, muskrat, raccoon, bobcat, and fox are also abundant roanmals. Turkeys
are lew in roxnbers in this region today, but would probably have been numerous
in the past. Ihe abundant oxbow la]ces and sloughs together with hardwood
timber stands, maJee this axea attractive to migrating water &wl and resident
woodcock peculations. In addition, morning doves, red-teil hasdes, great
homed cwls, turkey vultures, and blue and green herons are examples of the
larger bird species of the area.
The ecosystem sicports a wide variety of reptiles and attjMbians, many of
which occur in high nxxnbers. These include turtles, snakes, salamanders,
lizards, and frogs. Fish present in the area inclu^ large populations of
bass, bowfin, carp, catfish, gar, perch, shiners, and sunfish.
CULTORAL IKVIROWENT
The results of the past 15 years of intensive archaeological efforts have
documented a long and continuous human occupation of the Tanbi^bee Valley.
Tbe stages of development reflected in the archaeological record of this
valley are the same as the rest of eastern North America (Table 2) :
Paleo-Indian ca. 15,000-10,000 B.P.; Archaic 10,000-3,000 B.P.; Wbodland
3,000-900 B.P.; Mississippian 900-550 B.P. ; and Historic 550 B.P. to present.
(Alexander a, b; Bense , a, b; Dye and Watrin ; Ensor
, ; Futato , ; Jenkins , ; O'Hear et al. ; Oakley
and Futato ; Peterson n.d. ; Walthall , Weinstein ).
While a ccnplete sutmary and integration of the culture history of the
Torbigbee Valley is beyond the scope of this section, a synthesis of the known
information will be attempted with more attention paid to the Paleo-Indian and
Archaic stages, since the>' were the main focus of this stuefy.
PALEO-INDIAN STAGE
The Paleo-Indian stage is the least xinderstood portion of the aboriginal
occipation in the Tcribigbee Veilley because of the lack of docunented sites.
This low archaeological visibility may stem frem a relatively low population
density or a s\±)sistence/ settlement pattern that resulted in widely scattered
and briefly occvp>ied camps. Alluvial burial or erosion by flood deposits
could mask or eliminate cultural materials. The low frequency of identified
Pcileo-Ix^dian sites likely results from a cembination of all these factors.
Temporally sensitive Paleo-Indian artifacts, primarily projectile
point/knives, have been recovered intermittently from the central valley near
Gainesville, Al to the headwaters near Ryan's Well, Ms through the Divide-CXit
in Tishimingo County, Mississippi to Tennessee. All finds appear to have been
out-of-oontext . There have been surprisingly few specimens recovered in the
Totttoigbee Valley (less than 50 projectile point /knives) catpared to the Icu^je
Date Years
Period
Culture/
Variant
Phase
A.D. B.P.
1,735
265
Historic
Fully Historic
1,540
450
Historic
Protohistoric
Burial Urn
Sunmerville IV
1
400
Middle
Moundville
Simmerville II-III
650
1,340
Woodland
Late
Miller-
Baytotm
Miller III
B.C.
100
2,000
Wbodland
Middle
Miller
Miller I-II
500
2,500
Gulf
Format ional
Late
Alexander
Henson Springs
1,000
3,000
Gulf
Fonnatlonal
Middle
Wheeler
Broken Pimpkin
Qnedc
3,000
4,500
5,000
6,500
Archaic
Archaic
Late
Middle
Little Bear Beech
Creek
Bentcn-S^oes Whlnut
White parings
6,000
8,000
Archaic
Middle
Eva/Mcurrcw
Mountain
Hickory
7,500
9,500
Archaic
Early
Kirk
Poplar
8,000 10,000
Archaic
Early
Dalton
Cochrane
10,000
12,000
Paleo-Indian
Late
Quad-
Beaver Lake-
Greenbriar
11,000
13,000
Paleo-Indian
Middle
Cunberland
Paleo-Indian
Early
Clovis
The three Paleo-Indian stage periods established for the Southeast
(Hillians and Stoltnan ; Dragoo ) are present in the Central and l^per
Tcnhic^Dee Valley (Table 2): Early Clovis (15,000-13,000 B.P.); Middle
CUBfcerland (13,000-12,000 B.P.); axid Late Quad-Beaver Lake-G&reenhriar
(12,000-10,000 B.P.) . The general lithic assenblage associated with these
ocnplexes includes a bi£aoe and a flafce-end-blade technology which produced
many distinctive artifacts such as fluted and eared projectile point/knives,
unifaoe side and and scrapers, and flake-blade knives (Biscr :16) . Clovis
period materials have been xecxjvered from the oantral Tcnbigbee Valley in
Alabama at Clear Lake (1P) near Pickensville (Bense ; Muto and Gunn
), from four sites in the Gadnesville Reservoir (Bnsor :16) and from
the Mann site (22TS565) in the Divide (Dye and Matrin ) . The context of
the centred valley specimens appears to be disturbed. All were surface finds
except Clear Lake. TWo Clovis projectile point/knives were reoovered from the
Mann site in the Yellow Cre^ drainage. One was reoovered from a sand^ zone
beneath the midden, but it eqppears that it was associated with later temporal
markers (Dye and Watrin :7-64, 7-68) . The second Clovis was moved out of
context by an amateur.
Only one site vrith a docuinented Ctnberland period diagnostic marker was
found: the Rickocy site (22It621) in the l^sper Tonbigbee Vall^ (Vihite )
located approadmately 24 kn (15 mi) north of Fulton, Ms in the flooc^lain.
ISie Cttnberland point was recovered from backfill, azxi no assenblage ^rais
raoDvered fron excavations.
The latest Paleo-Indian period, Quad-Beaver La]ce-Greenbriar, has been
docanented at several sites throu^iout the waterwz^. The only site, however,
to produce a Quad projectile point/knife was the Poplar site (22lt576) located
in the Ifper Tcnobic^see Valley e^roximately 9 km (5 mi) north of Fulton, Ms.
This was a single point apparently out of context from a test unit (Bense
, b).
While Beaver Lake and Greenbriar projectile point/knives have been
recovered from sites throu^Kut the waterway, there appears to be a pattern of
increasing frequency in the upper valley, especially north of Fulton. Ms.
Perhaps increcuaed deposition and site burial downstreem and very limited
floodplain surveying and testing account for the dearth of these materials.
Only one site (lPi38) in the Gainesville Reservoir contained a single Beaver
Lake projectile point/knife (Ensor :17) , and no sites contained Greenbriar
projectile point/knives. While only one site (22It590) in the upper valley
produced a Beaver Lake projectile point/knife (Bense b) , Greenbriar
projectile point/knives have been recovered fxx^ ten sites. Three of these
were in the centred waterway between Amory and Aberdeen, Ms [22Mo569 (Brookes
), 22MO710 (Bense ) and 22Mo819 (Rafferty et al. )] and seven
sites were located in the canal section: 22It576, 22It539, 22It590 (Bense
, b; White ), 22lt621, 22Ps542 (Bense b) and 22Ts954
(Alexander ) . It should be noted that Greenbriar and Beaver La]ce
projectile poiiit/Joiife types are similar, and although a distinction has been
made between them in classification, they can be considered to be variations
on the same theme.
It appears that the Central and Ipper Tcmbi^^ee Vall^ were occupied
during the Pcdeo-Indian stage, but the euxhaeological remains are scarce, and
there was an ^parent preference for the rpper V2dl^, especially north of
Fulton, Ms, during the latest period. Although Paleo-lndian sites were
abundant in the Middle Tennessee Vall^, cmly one Paleo-lndian site has been
identified in the Yellow Creek drednage which is adjacent to the Upper
Tcmbi^see Valley and in the Middle Tennessee Valley drainage. People of the
Paleo-lndian stage were floodplain oriented in the Middle Tennessee Vall^,
but little use was made of the tributary valleys such as Yellow Creek.
ARCHAIC STASE
The Archaic stage is well represented in the waterway with many sites
identified in every section. From the initial surveys and testing in the
waterway (Atkinscai ; Blakeraan ; Lewis and Caldwell ; Nielsen and
Ifoorehead ; Rucker ) , it was evident that the Tcmbigbee Valley
contained a hi^ nurrber of Archaic sites. In addition, early on it was
discovered that distinctive types of Archaic sites were present in the
Tcmbic^see Valley; "midden mounds" (Atkinson ; Blakeman ; Rucker ) .
These mounds are de^ly stratified Archaic deposits (1-2 m or 3. 3-6. 6 ft
thick) , seme of which cxxitain burials dating between 6,000-8,000 B.P.
(Atlcinson ; Blakenan ; Rucker ). Studies carried out since the
early s have shown that the Upper Tcmbigbee Valley was a major settlement
area during the Archaic stage and that the midden mound site type extends from
Coluibus vpstream to the headwaters in northeast Mississippi neeir luka.
The three periods of the Archaic stage which have been docvsnented
throu^iout the Southeast are present in the Tcnbigbee Valley: Early
(10,000-8,000 B.P.), Middle (8,000-5,000 B.P.), and Late (5,000-2,500 B.P.).
This division is si^iported in the chixxiologies of Bense (, b); Dye and
Hatrin (); Ensor (, ); Futato (); Jenkins (); Oakl^ and
Futato () ; and O'Hear et al. () , although authors differ in the exact
chronological placement of the periods.
The Early Archaic period has two sequential horizons: Dalton and Kirk-Big
Sandy vAiidh are characterized by different projectile point/knife styles.
This period is first signaled by the appearance of the Dalton assenhlage.
Dalton projectile point/knives have been docisnented throu^iout the waterway,
but most contexts were disturbed. Four sites with intact ccnpcxients have been
identified and investigated: the Hester site (22Mo569) near Amory, Ms, IGrlXl
and lGr2 near Gainesville, A1 and the Colbert site (22C) (Muto ) . The
Hester site provided the most information (Brxxdces :30-31) . It apparently
was a hunting-butchering station as indicated by the large number of
projectile point/lcnives and scrapers found (Brookes :113-114). The site
is located in the floodplain of the central Tcmbi^see River and was used at
least twice during this period. The Dalton sites near Gainesville were only
sanpled (Ensor :20-21) and though smaller than Hester, inforroation from
than oonfirmed the use of the bipolar reduction technique and thermal
alteration for stone tool manufacture. All the in situ Dalton ocmponents were
buried beneath edluvial sands on or near the present river channel. All
oocvpations were interpreted as reflecting low density hunting and gathering
grotps, similar to the settlement pattern Interpreted for the Paleo-Indian
stage. The Dalton oocipation of the Tcmbi^see Valley could well be part of
the Red Hill |hase defined by Walthall (:48) for northern Alabama.
The best known of the Early Archaic horizons is the Kirk-Big San^, and it
has been docimented for all areas of the watervmy. While there are more sites
identified in the vpper valley north of Fulton, Ms, the sedimentation and
survey problems previously noted IDcely effected the low nurber of sites
identified downstream. It is interesting that the Hester site stratigraphy
indicated that the Big Sandy horizon followed the Dalton and preceded the Kirk
occvpation (Brookes :51-54,109). In eiddition, the Kirk (or Pine Tree) was
separated from the Big Sandy by a series of stenmed point types. This has not
been confirmed in subsequent investigations in the waterway. In the Kirk and
Big Sandy oonponents that have been investigated in the centred vall^ the
numbers of Big Sandy projectile point/knives has been too low, or the
stratigraphic relationship to the comer-notched types h2is not been
determinable. Many Early Archaic cenpenents have been investigated in the
vpper valley including seven of the sites reported here along with many sites
in the Divide-Cut (O'Hear et al. ) . In 2dl these investigaticxis , there
has been no clear documentation of a Big Sandy horizcxi below the Kirk horizon
as was the case at the Hester site.
In the flooc^lain of the ipper valley there is an association between the
Kirk and Greenbriar projectile point/)aufe types early in the Kirk horizon,
although Kirk quickly becomes the preferred style (Bense a, b; White
) . There appears to have been an erosional episode at places in the vpper
valley just prior to the Kirk ooc\:pation, possibly dated at 9,030±340 B.P.
(Alexander b) . This could explain the lack of Paleo-Indian, Dalton, and
Big San^ oonponents in this area.
The Kirk oocv?>ation as seen in these vpper valley sites reflects
short-term occipation associated with hunting, stone tool manufacturing, and
food processing. Features include pits, possible postmolds, ai^ clusters of
13
lithic debris frcm manufacture and repair, nie use of the i:9:per valley and
divide during the Early Archaic seens to have been heavy, as indicated by the
hig^ densil^ of Kirk sites in the Canal and Divide-Cut Sections (Sense ,
a; O'Hear et al. ) .
The Middle Archaic period (8,000-5,000 B.P.) has be^ well studied in the
Ttnbig^bee Valley. Alexander (b: 214-230) sutnnarized much of the
information. However, there has been no tharou^ siitmary or interpretation of
these works, and the infonnation is available only in govenxnent r^orts.
Therefore, a brief suimary and interpretation of the infonnation on the Middle
Archaic in the Tcnhi^see Valley is included here. A much more detailed
integration of the information is possible and should be performed.
The primary temporal markers for the Middle Archaic (Eva/Morxow Mountain,
Sykes, White parings, Benton and possibly Ledbetter) have been identified
throu^iaut the waterway. The Middle Archaic assemblages south of Aberdeen, Ms
appear to be linked to the Coastal Plain Archaic tradition in both stylistic
markers as well as lithic raw material. The assemblages north of Aberdeen
appear to be tied to the Tennessee Valley Archadc tradition (Bnsor
:24-25) . The most important aspec:t of the Middle Archaic is that the
adaptation to the valley changed during this time pericxi. The settlement
pattern changed from a series of relatively homogeneous snail encanpmoits at
many locaticxis to large base camps with small satellite cmnps.
The organization of Middle Archaic settlements throughout the waterway is
characterized by base canps which are in or overlcxoking the flcxx%>lain with
smaller sites Icxrated in the surrounding flcxj^lain, terraces, and inlands.
Waterway Investigations of the Micidle Archaic period have centered on the base
canps (Alexander a; Atkinson , ; Bense b; Binkl^ ;
Blakeman ; Dye and Watrin ; Otinger et al. ; Rafferty et al. ;
White ) . Other such base canps have been identified in the Tennessee
Valley drainage (Parker ) as well as in the Tuscuihia drainage (Alexander
b; Weinstein ) . The analler sites received less attention and also
generally had less integrity, with most being Icxated in plcpwed fields (Bense
a; Blakanan , ; Bla)ceman et al. ; Rucker ). Therefore,
most of vhat is known about the Middle Archadc is frcm base cmnp
investigations .
The Middle Archaic shift in settlement pattern appears during the
initiaticxi of the Eva-Morrow Mcunt£dn/Vaughn-^3emcpolis horizon ca. 8,000 -
7,500 B.P., vhen ocxnpation intensified at cartain previcxisly cxx:v:pi^
floc3dplain sites, and the use changed from temporary canps to Icxig-term base
camps. Seventeen such sites have been identified in the 150 km (100 mi) Icxig
area from Columbus north to luka and Corinth, Ms. The northemnost fear are
within a 22 km (15 mi) radius in the headwaters curea of Mackey's Creek, Yellow
Creek, and the Tuscanhia River in extreme northeast Mississippi. Five
additional midden mounds have been identified in the 30 km (20 mi) stretch
down Mackey's Creek and the East Fork of the Tcmbi^ae to Fulton, Ms. These
sites are 22Ts954 (Alexander b) , 22It590, 22It539 and 22It576 (Bense
a, b) , and 22It621 (White ) . There cure no miclden mounds
identified in the 38 km (25 mi) stretch of the Tcmbi^Dee between Fulton and
Amory. 22Mo710 (Bense ) is likely a midden mound, and is located just
below Amory. Three midden meuixis have been identified near Aberdeen, Ms, ca.
22 km (15 mi) downstream frcm Amory. One hcis been excavated (22Mo819:
Rafferty et al. ) , one has been tested (22Mo752: Bense ) , and one has
been surveyed only (22Mo747: Atkinscxi and Elliot ; Blakeman ). Four
miciden ncunds have been identified in the next 35 km (25 mi) stretch of the
Tombigbee between Aberdeen and Columbus, Ns and cue the southemnost midden
mrmmam identified. The Vauc^ tiound (22Le:^38) mbs tested by Atkinscn ()
and pcodooed the first early date (66^±95 B.P.) &qb the adddni aoiaid site
type. The KeUogg Itound (22C) mbs tested by Blahnwi and pcoduoad another
evly date ( B.P.) (BLakman :96). The KelXopg Village (22a527)
Mae tested by Blakanan (:26-39) and later eecevated by AtJdnson et al.
() . The Barnes Mound (22Lo564) , the southemnost ncund, vies also tested
by Blakanan (:75-94).
The change in settlanent pattern during the Middle Archaic at the midden
mound base canps is reflected in the build-ip of a dark, organic midden, an
increase in site features and site fztcilities. Possible structural remains
have been identified in several of these sites: Brinlcley (22T) and Mann
(22TS565) in the upper valley and East Aberdeon (22Mo819) in the central
valley. The inves^gation reported here docixnented ocnplex prepared areas
which were centers of activity and had at least one and often several hearths
at tliree sites in the upper vzdley: Ilex (22It590) , Walnut (22lt539) , and
Pcplar (22It576) , as well as an additional site neiu±y (22Ts943) . Burials
were interred at sane of these sites in both the flamed and extended
positions. Several cremations have also been docuaented. A "oemetery" with
graves ordered in rows and more than one extended bof^ in each grave mbs also
encountered in the vq;per valley (22It539) by the investigations. Not all
midden mounds have all these features or facilities. This is likely due both
to differing levels of investigation and the range of vitiation in aboriginal
site use. Fran all evidence, these sites beczne the foal points of activil^
early in the 3,000-year period of the Middle Archaic. This trend climaxed in
the Benton culture.
As noted above, there is little known of the satellite camps. One was
investigated in this stu^ (22lt623/22It624) , and the results indicate that
the site was used more intensively during the Middle Archaic than in previous
periods. Ebr example, the use of large storage and refuse pits was initiated
at the site during tMs time. Despite increase in use, evidence indicates
fewer inhabitants and a narrower range of activities than at the midden mound
base canpe in the vicinity (22It539 and 22It621) .
Differing tradition affilitations (Coastal and Tennessee Valley) of the
populations in the valley and divide has pemdtted the developaent of an
accurate chronology, but there are sane differences in the stylistic markers
and assemblages within the Tcmbi^see Vall^. l^pstreem &an Aberdeen, the
settlement change occurred during the Bva/Morrow horizcn ca. 7,000 B.P. and by
the of the sykes-White parings culture permanent site facilities have
been documented. Downstream fron Aberdeen, the Vauo^ horizon settlement
pattern changes. Typical midden mound site facilities and burials are
docunented by the following Sykes horizcn. Throughout the entire area, the
Benton culture marks the f loresenoe of this period and the use of the midden
ncund site type.
Little detailed information is available about the Late Archaic period
(5,000-3,000 B.P.) in the Tcnbi^see Valley. This cxxitrasts greatly with vhat
is known of the Middle Archaic. Few intact Late Archaic ccaponents have been
identified and investigated in the waterway, primarily because of disturbanca.
Markers for this period (Ledbetter, Pickwick, and Little Bear Creek projectile
polnt/lcnives) have been found throu^cut the valley and divide, indicating
cxxitinued use of the area. However, the pattern of settlement appeains to
shift fron the midden mound base canps to a more dispersed settlanent pattern
aJcin to that which preceded the Middle Archaic phenomenon. At least one
intact corponent of the Late Archaic period has been investigated,
22It623/22It624, and it is reported in this dcxnment. This period is better
I
knom in the Tennessee drainage, e^aecially in the Little Bear Creek and Cedar
Credc valleys (Futato ; Oakley and Futato ) . Cultural oontinuity
between the Middle and Late Archaic periods is seen in all investigations.
Hewever, large base caops, with a plethora of pits, hearths, prepared areas,
and burials, appear no more in the Tccbigbee Valley. The thidt midden I
build-up during the Late Arcdiaic in the midden mounds is evidence of heavy
occupancy. Although no undisturbed Late Archaic midden mounds were found,
there were no hearths, prepared areas, or burials in the midden mounds during
this time.
GULF FOBMATZGMAL STAGE I
Approodmately 3,000 B.P. ceramics were introduced to the mid-South.
Wheeler fiber-tenpeied ceramics came first, and they were soon followed by the
Alexander sand-teapered series (Jenkins , , ). Sane znithors
include the period of the appearance of these ceramics in the Late Archaic
period (Alexander b) , but most aco^t JenJeins and Walthall's () Gulf
Foonational stage as an intermediate one between the Archaic and Wbodlind in
the Coastal Plain. The Middle and Late periods of the Gulf Farmational stage
have iDeen consistently docixnented throughout the waterway and surrounding
areas. The stage is initiated in the waterway by the appeeuanoe and exclusive
use of Gulf Tradition oeramdes and ooncludes with the appearance of ceramics
derived from the northern, middle eastern, and southern Appalachian traditions
(Jen]dns :49) .
The Wheeler culture (3,000-2,500 B.P.) is found throughout the western
portion of the Southeast, but the majority of the sites and the apparent
development area is in the Western Tennessee Valley and is termed the Bluff
Creek phase (Jerdcins , ) . Sites of this culture are frequenty in the
Upper and Central Torbi^h^ Valley where the local phase is called the Broken
Penpkin Creek phase (Jeidcins , a, b, ; Walthall axxi Jenldns
) . Although Wheeler sites are found throu^teut the Centrzd and tapper
Ttnbi^see Valleys, the frequency increases in the northern portion.
Wheeler ceramics are characterized by fiber tenpering and a sisple bowl
farm vhich is either plain or punctated. Sisple stanped and dentate-stasped
designs were added between 2,800 and 2,600 B.P. (Jenidns , ; Jen)dn8
and Krause ; Walthall :89-91). The stylistic projectile points are
characterized by broad blades and incurvate, horizontal shoulders vhich show
similarities with Late Archaic types and are associated with bifacially
chipped Sterne tools, expanded base drills, and a variety of bone and antler
isplements. Ornaments include ground stene euid e;q)anded center-perforated bar
gorgets (Ensor ; Walthall , ; DeJamette, Walthall, and Winterly
) . The Central and U^jper Tcinbi^»e Valley Wheeler populaticns traded for
Tadlahatta quartzite cuid ceramics from the Bayou la Batre culture to the south
and steatite and sandstone from the Bluff Creek phase to the north.
Wcilthall () and Dye (, ) characterize the Wheeler settlement
pattern by floodpladn occupation in the warm months and vpland hills
occupation in the cooler months. Subsistence remains excavated from the
western Middle Tennessee Valley (Dye :228-231) include: vhite-tailed deer,
rabbit, squirrel, and other small masmals, box and soft-shelled tmrtle,
snakes, freshwater drum, catfish, hickory nut, weed seeds (inclvding
chenepod) , grape, walnut, and aoom.
The following Alexander culture (2,500-2,000 B.P.) also developed in the
western Middle Tennessee Valley and the headwaters of the Tarbi^see River. In
the Tennessee Valley the local expression of Alexander is called the Hardin
I
n
I
16
phase (Dye ) , and in the Upper and Central Tbwbi^jee Valleys it is known
as the Henacn ^ps^igs phase (DeJamette et al. ; Jenkins , , ;
Walthall ).
The separation of Alexander fxan the earlier Vtieeletr is based on changes
in oeraBd.cs. 1!his includes design motifs, aeramdc teoper (fton fiber to
sand) , and vessel form changes. New vessel shapes included globular and
vertical-sided txwls, flat-4)ased beakers or cips and sene exotic "boat* shapes
(Atkinson et al. ) . Decorations included incising, zone stanping, and an
elaboration of punctation (Jenkins and Krause :35) . The Alexander oeranic
attributes are similar to other oontCBporary Gulf Coast ceramic coaplexes such
as Tchefuncte, Orange, and Batyou la Batre, and to the earlier Wheeler horizon.
A detailed mo&l attribute study of the Alexander oeranic assenblage was
performed by the investigators on the isolated assenblage at the Aredia site.
Vessels ocnncnly have podal supports or annular notched bases and a variety of
rim treatnents, including fabric inpeessing, incising, punctating, notching,
stanping, and nodes. The lithic araenblage continues relatively undmnged
&an Wheeler and has been Isolated only at the Aredia site (Sense b) .
In the Tbnbi^hee Valley, Alexander sites have been recorded from the
flooc^lain and uplands (open sites and bluff shelters) (Sucker ; Atkinson
et al. ; Bense , b; DeJamette et al. ; O'Hear et al. ) .
Subsistence evidence is meager, but the presence of freshHater dnin, hickory
nut, walnut, acorn, grape, persimnon, and weed seeds (Dye ) suggests the
continuation of previous subsistence pursuits.
Most of the many sites of the Wheeler and Alexander cultures identified in
the %«atezway have been disturbed. Three intact cooponents in the Ipper
Tombi^aee Valley and Divide have been investigated: Aralia (22It563) (Bense
b) , Turtle Bond (22It643) (Thonas et al. ) , and the Mann site
(22C) (Dye and Watrin ) . In addition, the Yarborough site (22C)
in the central vzdley (Solis and Whiling ) also contained an intact
Alexander component. However, the hi^ nunber of both Wheeler and Alexander
sherds in the mixed deposits of the other sites, eqecially the midden mounds,
points to a full but unknown pattern of use in the valley.
WOCOLAN3 STAGE
The Woodland stage is well represented in the waterway. In fact, most
sites encountered contained materials from this stage. In the upper vedley
and divide-cut these dqioBits usucLLly had been disturbed by cultivation or
amateur digging. Most of the cultural in&srmation from this stage in the
valley has come from the central valley south of Aberdeen, Ms (Atkinson et al.
; Jenkins , ; Jenkins and Krause ; Bnsor , ). The
beginning of this stage is ca. 100 B.C. vben the Gulf Ceramic Tradition rather
ahrxqptly ended and was replaced by the northern-derived styles such as cord
marking and f^bric marking. The middle period (100 B.C. - A.D. 650) is
designated as the Miller I phase of the Miller culture (Jenkins , ;
Jennings , ; Walthall ) and has been divided by Jenkins (,
) into three sutphases based cn ceramic attributes. Jenkins () gives
an excellent description of the current knowledge of this phase that will not
be repeated heire. Instead, the primary characteristics and distribution of
this phase throu^ncut the Upper Tcnbigbee V^dley waterway will be briefly
presented.
The Miller culture in the Upper Tembi^see Valley began ca. 100 B.C. aixi
continued for about 1,000 years. Mortuary oerenonialiam in the upper valley
is represented by several burial mound sites, such as the Bymm Mounds (Cotter
17
and C3o]d)et± ) / the Iharr Mcunds (Bohannan ; Kardwedsley ) , and the
Dogwood Mound (Bense b) located both on anall crocXs and tte main strean.
The I^num site was a village associated %d.th six nounds in the headwaters of
the Ttoabi^oee. It contained nine houses (17>18.5 m or 56-61 ft in diameter)
asaociatsd %«ith flexed burialSf fire pits, and ahallow storage, or trash pits.
Another Miller I mound center, the Iharr Mound, also in the Tcnbigbee
headwaters included excavated and prepared pits to receive the cremation or
extended/ flexed inhiination. Charnel houses were constructed over these
excavated pits, and small logs were thmi placed around the grave. After the
charnel house or hut was burned, mortuary offerings were placed with the body,
and a mound was then built over the grave. Many Bopewellian ritual goods have
been associated with the cermonialiam at these sites docnxnenting a strong
participation in the Bopewellian Interaction sphere.
Several Middle Woodland middens have been investigated including the
Strickland site (22Ts765) (O'Hear and Conn ) , (Bense a) , the
Brinkley site (Otinger et al. ) the Itann site (Dye and Watrin ) , and
several other sites in the divide (O'Hear et al. ) . It appears that anall
villages or camps were scattered throughout the Tombi^^see drainage. Many of
the Miller I sites were established on earlier VSieeler and Alexander sites.
Village middens from this period often include refuse pits, hearths, and
even earth ovens. Structures were recovered at the Miller site (Jennings
) and viere either oval or elliptical (4. 5x5. 4 m or 15-18 ft) or
subrecrtangular (5. 8x6. 4 m or 19-21 ft) , and storage or refuse pits were often
found within than. One flexed burial was recovered from inside one of the
structures. Numerous shallow pits were encountered in the site midden.
Subsistence evidence includes hlckcny nut, acorn, and walnut (Caddell
:56) . Otie subsistenoe pattern appears to represent a ocntinuation of
Archaic and Gulf FoomtiGnal stage tamting, gathering, and fishing.
During the Hiller II phase of the Middle Nbodland sites beccme more
exxioentrated in the (central valley Black Prairie Belt (Jenkins , ) .
Burial mounds oontinued to be constructed, and burials were no Icmg^ in the
village middens. The absence of burial goods indicates a cessation of act±VB
participation in the Bopewellian Interaction Sphere. Excavated sites in the
central valley include 1(311x1 (Nielsen and Moordiead :29-44), 2GR2
(Jenkins :56-158; Nielsen and Jenkins :54-88), and 1PI61. A late
Miller II structure, measuring 8x11 m (26.5-36 ft) in diameter, has been
excavated at Site l(3dxl with a central oven 1.5 ro (5 ft) in diameter.
Nhile the Late Woodland period (A.D. 650-1,100) has been encountered in
all areas of the Tcnbigbee Valley and Divi(3e, it appears that populations were
concentrated in the central valley below (joluonbus and that the vpper vall^
was peripherally used. Jenkins (, ) has (iescribed this period in
detail in the central valley and divides the Miller III phase into four
subphases based an (cerandc typology. This Late Woodland phase is marked by
the introduction of clay eis a dominant cjerandc-taipering agent and by the
presence of a bow-and-arrcw technology. A ndcrotool assemblage was also
established at this time. Small chert flakes were frequently used as knives,
and pebbles were often chipped into scrapers. Other lithic tools include
flake perforators and drills.
Subsistenc^e practic^es of hunting, gathering, and fishing continued,
however, the use of cam is first substantiated during this period (Caddell
:56-57) . Hie frequency of 186, ) .
During the entire Mississippian sta^ the predominant arrow point style is
the moell triangular type (Bnaor ) , particMlarly the Madison type. A
variety of ahell artifacts were manufactured during this stage pciswily for
decorative nmasmnts. Bcxie artifacts edso oentimsd to be manufactured from
turkey and mmanals, for both decorative and utilitarian uses.
Faunal remains associated with Mississippian sites in the Central
Ttsnhlgbee Valley axe deminatad by the vbitsrialled deer, but also include
nammla, turtles, and fish (Wbodrick :157) . Woodrick also notes daring
Mieelesipplan tisms the oentinuing decline of the white-tailed deer and the
conocmltant increase of other faunal food souroes increased.
Com remains increased, but hickory nuts are still prominent. Cbm ««as
probably a main, if not the main, carbohydrate base of the diet during the
19
I
Mississippian pericxl, but hidoory nuts and aoams were still a part of the
diet (Caddell :67) .
l!te U^per Tcnbi^aee Valley seeoB to have either been i^arsely occupied by
Mlssissipptan people, or was well ooojpied by late Woodland culture practicii^
people. MissiMippian villages existed in the area outside the Tcafcic^bee
Valley in the Tipelo Hills, but they have not been investigated as well as
those in the Tennessee Vall^.
SlMAPSr
This project was conducted in the headwater area of the Upper Tonbi^see
Valley in northeast Mississippi. TMs area is daracterized deeply
dissected, san^ terrain and a wide, swanpy floodplain. Much recent
archaeological work in the Tcmbi^^aee Valley was pawned by the Tennessee-
Tonbi^aee Waterway. Most infarmation produced has been from the Woodland and
Missiuippian stages. A detailed chrofxilogy has been established, and
knowledge of the lifeway of these stages is also quite detailed. Mudi less is
known of t2)e Archaic and Gulf Fonnatlon£d stages. Ihis was the focus of these
investigations. The culture history presented here served as the foundation
of the project research design, vhich is enunciated in the next chapter.
20
CHAPTER III RESEARCH DESIGN
The csoRtinuing focus of this researdi has been to detendne the nature of
honter-gatherer adaptations to the Upper Tcnbigbee Valley during the Archaic
stage. The re search design has had two major stages of developnent. The
f ix^ stage vias the establishnent of the theoreticzd perspective and methods
for data recovery (Phases I and II) based on information from testing four of
the sites (Sense ) and the infonnation available from other similar midden
mound sites. The second stage of research design mbs developed after
preliminarily analysis was performed on the recovered material. Models of the
paleoenvironnent and archaeological/cultural patterns in the Upper Tcnbigbee
Valley Mere developed based on this analysis, and certain hypotteses Mere
tested in the fined phase of the project.
INITIAL RESEARCH EESICM DEVELC3Bq}laining the
actual patterns documented in the archaeological recx>id.
The regularities observed by most researchers of hunting and gathering
acx:ieties are as follows:
1. Ecxsnomic b^iavior is the result of oonscious choice. Selection of usable
resources, declsicn as to their proportionad xise and time of utilization,
and demographic and spatial arrangements chosen in order to aooompliah the
exploitation, all xxse human time and energy. These decisions structure
subsistence and settlemient patterns. Hunters and gatherers often esqpend
small amounts of energy in the focsd quest; allotment of the expenditures
depends on the avedlable choic^es amcmg ocmpeting or nutuadly exclusive
activities.
21
I
I
2. Resource selection is deliberate rather than a random or opportunistic
utilization of resources. Local, tenporal, and spatial variations of
resources are present in all hunting and gathering societies, but it
appears that opportunistic utilization is a conscious decision to alter
the usual patterned activities.
3. Decision-malcing process is rational and is appropriate for laiderstanding
the roles of choices and decisions made by hunters and gatherers.
4. tfticertain outocme probabilities must be estimated, because the exact
probabilities of the oonseguenoes of economic choices are not known. At
best they eue estimated from previous experience and new information.
This reduces risk - the decision to a partial uncertainty.
5. Choices are made to satisfy predetermined eviration' levels. Alternative
choices or ccnpeting objectives are considered, and an order or preference
is established. This is inportant, because it incorporates decisions
vhidi include procurement of generally nonedible items (hides, antler, and
bone) euid deals with conflicting goeds or objectives.
6. Resource scheduling uses a mixed strategy solution to ocnpeting resource
availability, 'this conbines severed options, such eis sinultaneous
performance of more than one activity, simultaneous exploitation of more
than one location or area, or sequentied change of activities and
locations.
Ihe desire to limit effort underlies cdl economic decisions and is an
inportant goed that guides the economic behavior of hunters and gatherers.
Minimization of effort (mini-max theory) , or the keeping of effort within a
predefined range, crosscuts all studied grcips.
The decision-making processes of hunters and gatherers are a result of
resolving specific interrelated problems between man and the natured
envirtanment. This relationship is the roost inportant factor conditioning the
economic bdiavior of hunters and gatherers. vRien these relationships are
considered in a systemic framework, it is cedled the ecological approach.
Human ecology considers a human population as part of the ecosystem (Steward
) and focuses on the structural relationship of a groip to its natural
environment. The epproach of ecology provides a structure for the focus and
priority of exploitive activities, and it uses the conceits of ecological
theory such as ad£ptation, stability, diversity, and trophic level.
With this knowledge of hunter-gatherer economies, organizationad
principles, and resource use scheduling developed by anthropological studies,
this reseaixh was designed to apply it to this investigation of prehistoric
hunters and gatherers in the Upper Tarnbi^:?ee Valley. Therefore, it is within
the framework of cultural ecology that the collection strategies, analyses,
and interpretations were designed.
REFINED RESEARCH DESIGN
The above general principles and orientation guided data recovery and
analyses in the first two phases of the project. The collection strategy,
vdiich will be described in the subsequent chapter, was bcised on eooncmic
principles, and chronology was established by utilizing flotation to address
subsistence, det£dled soil analysis for traces of organic matter, large block
excavations to reveal site facilities, and frequent dating of the cxdtural
features and deposits. Preliminsuy analyses, ccxpled with additional new
information from othier geoscience and archaeological studies in the the
Torbi^see Valley, provided multiple lines of evidence for first-stage models
22
of both the paleoenvizonnent and cultural patterns of the Tapper Tcnbi^see
Valley. These models inccnnporated apparent oorrelaticns betMsen the Holocene
enndxcxinent2d shifts and cultural b^iavior as seen in site use, material
asseifclage, and depositional units. These models suggested research
questions, and suibsequent hypotheses were constructed and tested. These
first-stage models obstructed in frcm the results of Phases I and II are
described in the following sections.
PAI£QQ«VIRGN1Ein»L MODE!.
The geoscience data from this and other projects provided iaportant
information on the evolving landscape in the U^per Tombi^iee Valley.
Preliminary analysis indicated the following Holooeb sequence of climatic and
depositlonid. events.
A (S^^namic environment with unstable land surfaces characterized the Late
Pleistocene and Early Holocene (16,000-8,000 B.P.). The Tcnbi^jee was
probably a braided or coarse-graini meandering stream, and floods were
probably of a greater magnitude but with less &equency than during the
previous Pleistocene period because of a greater su^iended sediment load.
There was little surface stability in the flooc^lain with more open ground and
less vegetation in the first part of this period. Typical floodplain
topographic features included distributary channels, braided channels,
streams, and gravel-sand bars.
The tenperature appears to have been cooler than during the remainder of
post-glacial climate, and there vms more episodic rainfall of greater
magnitude than before or since with drought conditions probably occurring in
winter and late sunmer. The presence of maples in the pollen record of the
forests indicated to Muto and Gunn (:6-9) that adeqbte growing season
moisture (spring-summer) was available.
A period of stability followed during vhich the basal Early Holocene
sediments developed in some areas into a midHEiolooene soil, vdiich is locally
preserved. Radiocarbon and archaecmagnetic dating place this period of
landscape stability between 5,000 and 8,000 B.P. Deposition in the floodplain
was reduced towards a xeric peak between 5,000 and 6,000 B.P. This
mid-Holocene period corresponds to the Altithexmal or Hypsithermal climatic
^isode originally defined Antevs () and later identified throu^iout
the Northern Hemi^here.
More mesic conditions ensued approximately 5,000 B.P., and deposition
resumed in the floo^lain. The Toibi^see initiated down-cutting during this
period in response to a drop in sea level during a period of cooling. A mesic
hardwood forest was established during this period (Muto and Gunn :6.11)
In sim, the paleoenvlrormental model of the Late Pleistocene and Holocene
for the Thnbi^see Valley had three bemiic episodes: cooler and nxsre moist
(16,000-8,000 B.P.), warmer and probably drier (8,000-4,000 B.P.), and present
conations (4,000-0 B.P) . This model has been ev^viated in this stu^ of the
past cultuTeil systems of the Upper Tcmbigbee Valley.
ARCHAECfLOGICAL MCDEL
During the E2u:ly and initial Middle Archaic (Kirk through Eva-Morrcw
Mountain: pre-7,000 B.P.) , the sites appeared to have been interroittarttly
occupied. The archaeological material had a low cubic volizne ratio and
ocxisisted of chipped stone tools and debitage. Assemblages were occasion2dly
separated by several centimeters of sterile soil. Site features of this
period included lithic ddsitage ocncentraticns and only three pits. No site
facilities, such as structures or hea3±hs, were encountered in these d^xssits
and short-tenn oocvpaticns were inferred. The Ilex site (22lt590) had the
darkest midden for this period, as well as the hi^iest nunber and density of
cultural material, and it could have be^ a longer term or more intensely
oocqpied locedity. Ihe primary lithic raw materials were heat-treated cobbles
of local dierts modified by bifacial reduction. A few projectile
point/knives, drills, and other small, refined tools were made of a better
quality inpoited chert. One Eva/Hbrrow Mountain burial was also encountered.
Between 7,000 and 5,000 B.P. (Sykes-Mhite Springs and Benton) use of these
sites changed significantly. Ihe material record consists of abundant
charcoal and broken fired clay, increased densi^ of cultural material,
ti^tly clustered residential areas with ocnplexes of features includi]^
prepared clay areas with multiple hearths associated with residential
activity, storage pits, refuse pits, and many burials. One site (22lt539) of
the S^ces-lfhite Springs/Benton occupation had two distinct "cemetery" areas
with graves ordejred in rows, often with multiple individuals (2-3) in the same
grave in the extended, prone position.
The lithic raw material, stone tool assemblage, and technology \ised during
this intense occupation of the floo(f^lain was also different frcm the
preceding oocipations. The primary lithic raw material used for tools was
Fort Payne chert, vdiich was imported in the form of prepared bifaoes. This
pattern of importation contrasts sharply with the preceding pattern of
utilisation of local cherts. The Benton occupation also appeared to have a
narrower range of tools, dominated by the large projectile point/knife, which
were apparently, multipurpose tools.
Severzd of evidence suggested that the Benton sites were used as
year-round or at least as semipermanent base camps. Dating of the four
residential areas indicated that they were used between ca. 6,300 and
5,300 B.P., probedtily during the fizrst part of the Altithermal period.
Occupation of the sites continued after the highly concentrated
S^ces'4 se^s an understanding of
the post-Benton occupations in the Upper Tombi^see Valley. While all sites
investigated had been occupied during the Late Archaic, Gulf Formational,
Wbodland, and Mississippian stages, only two intact ccnponents were present
from these periods. It does appear, howaver, that the Late Archaic cultural
tradition persisted in the UIV.
The middle Gulf Formationad (Henscn Springs) culture (ca. 3,000-2,000
B.P.) had distinctive ceramics, and it apparently interacted with the Middle
Tennessee Valley people. Production of these ceramics appears to have begun
and ended rather abruptly. The fabric- and card-marking tradition which
followed during the Middle Woodland persisted for a relatively long time (ca.
2,000-700 B.P.) . The Late Woodland stage apparently extended well into the
cicoepted time p)eriod of the Mississippian stage documented in the nearby aueais
of the Tupelo Hills and Central Tcmbigbee Valley.
OORRELATIOJ IN PALEXMA/IPOlWENT AND AICHABOIXCIOVL/OIiTURAL PMTEPN
The archaeological record at most of the sites consisted of three
recognizable catponents of the Archaic stage: 1) Eaurly to initiad Middle
2
Ardialc (ca. 12,000-7,000 B.P.) %d.th lic^t and separate oanoantzatians of
lithlc tools and ddaitage, only a fear features, and little organic staining,
2) Middle Arch^ (7,000-5,000 B.P.) of dense, clustered artifacts
with residential features, cemeteries, hearths, storage and refuse features,
and dark, organically stained charooal-rich soil, and 3) Late Archaic
(5,000-3,000 B.P.) deposits with fewer artifacts, refuse, storage, and
occasional burial features. Occupation types appeared to correlate with the
depositional and soil formation episodes: the paleosol contained Early Archaic
(Greenhriar, Big Sand/, Kirk) and soae initial Middle Archaic (Morrow
Mountain, Eva) ooaponents. The terminal Middle Archaic (S^ces-White
Springs/Benton) ocaaponenL %«as in a deak midden zone resting uncmformably on
the paleosol. The disturbed post-Benton, terminal Late Archaic deposits were
situated at the base of the dense, organic epipedon.
AEDBESSABTE CULTORAL PIOCESSES
Three cultural processes supported models of environnental d/namics,
cultural behavior, the eoonoroic regularities of observed hunters and gatheiers
and the data could be addressed in the final phase of the project:
1. Initial settlement dynamics
2. Cultural evolution
3. Transition to Late Archaic
PROCESS 1: INITIAL SEITUmENT DYNAMICS
The Upper Toahi^bee Vall^ was apparently initially occupied by the late
Paleo-Indim popula^ons after initial settlement of Midi^ Tennessee
Valley flood|pIain in the Early Paleo-Indian (Clovis) period. The settlement
of the Tonbigbee Valley apparently progressed in a north-to-south direction
daring the Late Paleo-Indim and Early Archaic periods. The materials
recovered from four floo^laln sites in this pr^ect (22It539, 22It576,
22It590, and 22It621) contained asaemblagee of these periods. These
assesblages, identified by tenporally sensitive projectile point/knives, are
of the Late Paleo-Indian, Dalton, and Early Archaic periods. The asssehlages
consist of a wide variety of lar^ and small chipped stone tools, debitage,
and groundstone. In addition, three pit featuree and six chipped stone
clusters were associated with these ocnponsnts. The integrity of these
deposits was good, and meaningful aasociatlona appaared to have been preaerved
in the cultural material. Prelimincury analyaia indicated that there were
similarities in features and assenblages which croaacut the sites, as well as
differences among them. Fran these preliminary observations the following
first-generation hypotheses and test expectations of the early occupation of
the upper valley were ocnstructed.
Hypothesis The adaptations of the initied settlers to the environment of
tne Upper~Tcnbi^3ee Valley became increasingly refined thrcu^ time,
reflecting an incre2ising knowledge of the envircrmentad resources available
and methods of exploiting these resources.
Test Expectations
1. The tool assenblage and technology during initial occupation will be
initially hcjorngenous and increase in diversity through time as a result of
loced adaptation, distcmoe from origination, isolation, and evolution£uy
processes.
2. Individual site use should beocroe more differentiated through time, with
increased knowledge of the area due to the differential distribution of
raw materials and other natursd resources.
3. Specialized tools and weapons of high inportance in food acquisiticm and
production of necessities (projectile point/knives, scrapers, drills,
hamnerstones) will be made of the hi0ie8t quality raw material accessible
by quarry or trcide, will be manufactured systematically, and will be
extensively curate^rediazpened.
4. Generalized tools and weapons of lower inportance in food and ^Iter
provisioning (ciioppers, cleavers, etc.) will be made primarily of local
raw materials, be manufactured less systematically, and will be made and
used as disposables.
PROCESS 2: CULTURAL E\m7nQN
•Hie initiation of the Sykes-White Springs portion of the Middle Archaic
(ca. 6,500-7,000 B.P.) appeared to signal a threshold crossing of adjusOnents
to the floodplain that culminated later in the Benton culture
(5,000-6,000 B.P.) . Mew site features which appear^ included prepared areas
^th multiple hearths vhich were probably residential areas, Ixirials vAiich
were either associated with the large prepared areem or in ordered and
separate cemetery areas. Many hearths were also discovered on the general
site surface at this time.
Locations that were once short-term campsites in the early Archaic period
were used as base canps, perhaps on a year-round basis, and they were more
widely and regularly sgaced. The population organization apparently shifted
from small, dispersed groups, utilizing both uplands and lowlands to
nucleated, large groups focusing on the florx^lain. Apparently fewer
locations were occupied by Benton groups than in earlier periods, but they
were oocvpied larger groips engaged in a wider variety of activities.
Several alternative hypotheses can be generated to explain the
eurchaeological record during this period. Eacdi has expectable results vhich
would be present and testable in the archaeological record.
Hypothesis 2: The Kirk, Eva-Morrcw Mountain, Sykes-White Springs, and Benton
occupations represent an evolution of adjustments in subsistence and
settlement strategies.
Test Expectations
1. Smooth evolution (continuity) should be present in tool tecdmology, raw
material use, and stylistic indicators of the Kirk, Eva-Morrcw Mountain,
Sykes-White Springs, and Benton eusseroblages. Abrtpt changes should not be
present.
2. A gradual increase in the amount of cultural material, tool manufeK:turing
debris, brdcen and lost tools, charcoal, organic remains and features, and
initiation of more permanent site facilities is to be anticipated.
26
Hypothesis 2^ Bva-Morrow Mcuntain culture evolved ftan the Big Sand^ and Kiric
cmtures'of the Early Archaic, but the S^ces-Hhite Springs/Benton is an
esqpansion of ideas or people finom elsewhere and is actually a part of another
cultural tradition.
Test Expectations
1. Continuity of evolutionary changes should be present in eissenblages of
Kirk and Eva-4torrow Mountedn in tool technology, raw material tise, and
stylistic indicators. Discontinuity should be present between thm and
the Sy)ces-White Springs and Benton cultures.
2. Tool inventories and uses of tools should differ in Kirk-Eva-Morrow
Mountain and Sykes-Mhite Springs/Bentcn.
3. Abn;^ increases/changes in site use (length of occv^tion, facilities,
and features) should be evident and correlate with the appearance of the
Sjtes-White Springs/Benton occi^iatians.
4. There should be significant honogeneity within and between early
Sykes-fiftiite Springs/Benton settlements in subsistence and technology,
indicating their ocmmon culture and interaction characteristic of
pioneering settlements.
5. The resident population's (Eva-44orrow Mountain) nonfusictioned tr£dts of
material culture, such as style of inplements or burial practices, will
not be incorporated into the inmigrants' material culture.
Hypothesis The process by \diich Sj^ces-White Springs/Benton culture
de^loped into the Upper Tcmbi^3ee Vall^ was by diffusion of ideas from
outside the area.
Test Expectation
1. There will be noticeable continuity throu^ the evolution from Eva-Morrow
Mountain and Sykes-White Springs/Bentcn nonfuixrtional aspects of material
culture such as styles of inplements and burial practices.
Hypothesis 2! leurge flooc^leiin sites of Sj^ces-White Springs/Benton were
occupied throughout the year with other localities being utilized fca:
specialized extractive purposes.
Test E^qaectations
1. The seasonal indicators (flora axd fauna) in the large floo<%>l2dn sites
will be from all seasons of the year.
2. There will be a difference in site facilities and range of activities
between the floodpledn sites.
PROCESS 3: TRANSITIC»J TO THE TERMINAL LATE ARCHAIC
In the archaeological sites investigated in the Upper Tcmbi^aee Valley,
there appeared to be a cessation of the Benton "way-of-life." Notioeable
27
.A Xi' -J 1. .u . . ‘. |^^ I j -
dianges in site use, tool nanufacture, tool inventory, and raav Baterials were
indicated. Preliminary analyses indicated there mig^ have been a return to
the "EvarMorrcw Mountain way-of-life* rather than an evolution out of Benton
into Little Bear Cre^. While the issue can be framed into testable
hypotheses, there is a problem with the ardiaeological record, sinoe many of
the Little Bear Creek deposits were disturbed. Only one intact Little Bear
Credc cultural deposit was available, and it was not on a former large Benton
floodplain site. Therefore, the test eaq)ectaticns of the hypotheses below are
tailored to the specific project data.
Hypothesis 6: The Sykes-White Springs/Benton culture was a tenporary
intrusion into the Upper TOmbigbee Vall^, and the local
population was never integrated. After the Benton exit, the
local population resumed use of their former sites.
Test Expectations
1 . There should be measurable similarities between Kirk/Eva/Morrow Mountain
and the Xiittle Bear Creek assemblages in tool tedinology, raw materials,
stylistic patterns, tool inventory, site use, and facilities.
2. There should be measurable differ^ces between Benton and Little Bear
Creek in these same a^)ects.
Hypothesis 2* Sykes-White Springs/Benton culture evolved into the Little
Bear Cre^ culture.
Test Expectation
1. There should be measurable similarities and continuities between the
material remains and gradual change in site use and facilities.
OTHER ADDRESSABLE CULTURftL ISSUES
Cultxire history and lifeway level issues of the Gulf Foxroationed and
Woodland stages could also be tested in this project. These research
questions were conoemed with the Woodland ceramic style/horizon sequence and
the design methods and technological variability among Alexander (middle Gulf
Fornational) ceramics.
One ocxtponent of the Henson Springs phase was a single-oonponent site with
both midden and features. This culture v^s difficult to stucty because only a
small nunber of undisturbed sites were found and excavated (DeJamette,
Walthall, and Wimberly ; AtJunscn et al. ; Bense , b; Joikins
, , ) . The large sanple of Henson Springs ceramics recovered from
excellent context at one site offered a good opportunity to define the ceramic
assenblage in the ITIV with those of the mid-South and Coastal Plain. The
following research question can be addressed:
Reseaixh Question 2^ What designs and technologies characterize the Alexander
ceramic assemblage at 22It563, what is their range of veuriability, and what is
the relaticn^^ip of the Alexander ceramic assenblage to that of the
contenporaneous Tchefuncte, Bayou la Batre, Wheeler, and St. Johns cultures?
28
TWO ocnfxsnents of the Woodland stage were enoountered which had sufficient
integrity to provide further information. Site 22It606 had well-^xreserved
Miller /BaytoMn and terminal Miller Woodland features. The Woodland ceramic
sequence for the U^per Tcnbigbee Vall^ was not well understood, and it could
be addressed throu^ ceramic analysis similar to that utilized by Jenkins
() and Stepcnaitis () in the Central Tcnbi^jee Valley.
It appeared from preliminary analysis that the grog^tenpered ware was
utilized oontenporaneously with the shell-tenpered ware. If these ware groups
were used at the same time in prehistory, overlap between the various modes of
form and deooraticn within them should be c^ious. A moded. analysis similar
to that outlined for 22lt563 was inplemented to test the degree of attribute
overlap between the shell- and grog-tenpered wares.
The other Woodland ccnponent with cultural history potential was 22Mo531,
a burial mound of the Miller I (Middle Woodland) culture. These ceramics were
used in ccnparison with those from 22It563 and other sites to refine the
ceramic sequence of the Woodland stage in the tapper TCmbigbee Vall^. The
researdi questicn was as follows:
Research Question 2: What refinements of the Woodlai^ ceramic sequence can be
made based on the material recovered frcm Sites 22Mo531, 22It606, and 22It563.
ACDBESSABLE EMVIROtXENTAL PROCESSES
In this project paleobotanical and gecmorphological data were recovered
vhich can be used to test parts of the paleoclimatic model. Paleobotaniccd.
data, consisting of pollen and charred plant fragments, were recovered frcm
22It590 (Ilex Site) in an Early Archaic context (ca. 9,000-7,500 B.P.) and
reflect a boreal-type forest. This finding challenges the traditional
paleoclimatic model for the Southeast, which postulates the disappearance of
boreal forests ca. 12,000-13,000 B.P. years ago. Althou^ extensive efforts
were made to obtain radiocarton dates on the deposit vhich contained this
suite of boreal pollen, none were successful. However, the tenporal
projectile point/loiife markers are well dated (Kirk and Greehtoiar) . A date
range of 9,000-7,500 B.P. is likely for the deposits which contained the
pollen. It was proposed that pollen in sediments of other Early Archaic
components (if preserved) were to be ccnpared to the 22It590 information.
This comparison could help resolve the issue of whether the boreal forest
indicated in the Ilex site euea during the Early Archaic was an isolated relic
stand or a indicator of a genereilly cooler climate.
An additional line of evidence can be called to this question of Esurly
Holocene climate: charred plant freigments. While there were insufficient
amounts of the charred plant material for dating purposes, a botanical
identification could be used to detemdne the nature of the envirorroent ne2u:
and OR the sites during the Early Holocene.
Preliminary analyses of the sediments of the Early Holocene in this
project and by Muto and Gunn () have indicated that the flooc^lain surface
was relatively unstable, and the Tcrabigbee possibly was a braided or early
meandering stream during this time. Sediment analysis of the Early Archaic
d^xjsits could also be used to identify the depositional processes reflective
of the environmental conditions of the Vpper Tcnibi^Dee Veilley.
The hypotheses and test expectation that structured the investigations
into the Early Holocene environment in the l^per Tcmbic^aee Valley were as
follows:
29
and moi
stream.
isis 8; The Early Holocene period (ca. 12,000-8,000 B.P.) vms cooler
Bter" than present, and the Tbnbi^see was a braided or early meandering
Test Expectations
1. The pollen preserved in the sediments of this age will be boreal or
coniferous species.
2. The macrobotanical remains will consist of plants which reflect cooler
than present conditions.
3. The sediments of the deposits will be coarse and in lenticular or massive
deposits.
The second environmented process vhich could be addressed was the
nddr-Holocene climatic episode known as the Altithermal. This episode has been
characterized as the episode of maximzn Holocene warmth, climatic stability,
and possible dryness. The episode has been documented in pollen studies from
the Appalachians to Missouri, Arkansas, and Tennessee (Delcourt ; Delcourt
and Delcourt ) . Pollen data from the tqpper Central Tcnbighee Valley near
Columbus, Ms. also indicated maximun xeric conditions ca. 5,000 B.P. (Muto and
Gunn ) . Six sites which contained intact alluvial deposits of the
appropriate age t^ch could be examined to determine the nature of the
Altithermal episode in the Upper Ttnbi^^see Valley.
The unexpected discovery of well-developed paleosols with argillic
horizons in the Upper Tonhi^oee Valley floodplain sites raises mai^ questions
which have direct bearing on understanding environmental settings. The
paleosols developed in sediments of the Late Pleistocene and Early Holocene
and contained Early and Middle Archaic cultural material. Distinctive
morphology and perhc^ chemiczil and minera logical domains could be examined to
gain better factual understanding of past landscape evolution.
Hypothesis 9; Nid-Holooene climate (ca. 7,500-5,000 B.P.) was characterized
by an incre^lse in warmth and dryness vhich xreached maxiimzn levels between
6,500 and 5,000 B.P. This was also a period of landscape stability. More
mesic ccxiditions began ca. 5,000 B.P. and extended to those of the present
between 4,000 and 3,000 B.P.
Test Expectation
1. The pedological data will indicate that the mid-Holooene period Wcis stable
and reached a xeric maximum after the soil development.
The third environmental/gecinorphic process vdiich could be examined was the
unique dark, humic-stained soil epipedons. These are specific to
archaeological sites and are the most visible pedogenic feature. These
culturally produced features are apparently reflective of specific ocnditions
of the past occupations and are different than the soil-fonning processes
operating today. The carbon/nitrogen ratio, organic acid cotplexes,
l^qhorus ccmponents, and other features may be unique. The epipedons have
persisted for 5,000-6,000 years of weathering in a very humid climate, but the
rate and extent of change are unknown.
30
The environnental conditions and teaqporal factors necessary for the
formation of the dark epipedons were an enigma and offered an excellent
opportunity better to understand the past. An understanding of the mechanisms
of their formatian and taiporal changes was critical. Ihe epipedons may
represent dynanic bcdances of the past and consequently may be undergoing a
terminal change in our existing environnent. Neither their basic nature nor
the extent to \idu.ch leaching, oxidation, erosion, weathering, lack: of organic
additions, perttuhation, and volatization was influencing these epipedons was
understood. Ikiderstanding these ruiique features produced by past cultures
could e3q>and the knowledge of the past and create a better awareness of the
dynamic changes of the pedosEhere of cultural sites.
Research Question 3t What is the {hysical, chemiced, and mineradogical nature
oi organic-stainai soil epipedons vhich are definitive of past habitation
sites of the southeeistem Uhited States, and how is the development of these
epipedons related to cultural and environmental activities?
tHPOIHESIS BKFINEMEMT
Ihe lithic analyses constitutes the bulk of the Phase III cultural
material and just prior to the initiation of the lithic anadysis in Phase III
the hypotheses vdiich pertained to the Archaic were reevaluated. At tliis point
the lithic specialist and replicator had been selected, and which
operationalized the hypotheses in detailed final analysis was to be designed.
Ihe focus of the Archaic stuty was centered on the difference of the
Sykes-White Springs/Benton ccnponents in the l^per Tcmbi^see Valley midden
mounds. Of the six first-generation hypotheses ccaioeming the Archaic, five
dealt with this phencmenon based on three observed differences or chants.
1. Local Camden and Tuscaloosa chei± cobbles were used in the non-Benton
assennblages, and Fort Payne chert uiported to the sites as bifacial
preforms were found in Benton assemblages. Fort Payne chert was probably
imported from the Middle Tennessee Valley 30-70 km (18.6-43.4 mi) to the
north.
2. A difference in the technology was used to manufacture chipped stone tools
from cobble reduction in non-Bentcn times for the retouching of the
bifacial preforms of Fort Payne into finished tools in the Benton
assemblage.
3. Site use from activities that produced low densities of cultural material,
no organically stained midden, and few features in pre-Benton ccnponents
changed to activities that resulted in extreme density of cultur6d.
material, dark-stained midden, and a plethora of features including
prepeured eueeis with multiple hearths, single hearths, refuse pits, etc. in
Benton ccnponents.
A second generation of hypotheses were then generated that cculd address
these observations through lithic analyses.
Hypothesis 1_: The change in lithic raw material use occurred because of
depletion of the local raw materials (gravel ccbbles) ^diich
were buried in the floodplain of the l^per Tcrobi^3ee Valley.
31
Test E»qpectatiops
Sykes^White Sparinm/Benton
l^s waste ((^lantity)
inferior quality
feMer large cobbles
technological changes
iKnne curaticn of tools
more multipurpose tools
more inported racu materials
feMer tools lost or discarded
Hypothesis 2: The change in site use is due to a decrease in mobility of the
resident populations during the mid-Holooene.
Test Expectations
Indices of mobility used in past research included hafting implements and
tool ccmplexity. Hafting of implements e9q>ands their task utility and more
portability. Ttool ocmplexit^ is inversely related to population mobility. If
Hypothesis 2 is correct, then the Test Ibqpectations au?e:
1. Hoifted tools will increase with decreased mobility.
2. Tools will be more ccmplex with decreased mobility.
StMftRy ^ ^ RESEMCH DESIGN
In the six and one*'half years of this project, there have been two
developmental stages or refinements of the research design. The initial
research design, developed in , was based on testing infarmation and the
general theory of cultural ecology. This guided the data recovery and
preliminary analysis of Phases I and II. The second stage of the research
design was fomulated in with the development of the first-stage models
of paleoenvirorment and cultural adaptations based on the infarmation
recovered in Phases I and II as well as from other projects. Fran these
models, hypotheses were generated, and others were refined, to allow testing
with the data recovered in portions of this project. These hypotheses were
further refined as th^ were operationalized, especially those utilizing
lithic studies for hypotheses testing.
Non-S^^ces-tft^te Sprinqs/Benton
more wmste (quantity)
hi^ lithic quality
larger local cobbles
initial cobble reduction
less curation of tools
reliance on loced raw
materials
\
32
OAFTER IV ABOAEXDliOGICAL PRXEXXIRES
FTKTJI PROCEDDPES
Data recovery methods and techniques viere designed to be ocnpatible with
previous investigations in the Tennessee-Tcnbi^bee Waterwe^. Excavation
strategies and field recovery farms were developed and standardized prior to
Phase I to provide a mininun standard and ocnparable set of information for
each site investigated. A Field Manual was developed during both phases of
fieldwork. Only minor cilterations, as noted in the sections below, were made
for Phase II.
FIELD RBCCVERY OBCHNIQUES
SITE preparahom and mapping
When necessary, site surfaces were cleared of vegetation by hand utilizing
chainsaws, bushhocdU, and axes. Once sufficient space had been cleared, a
Cartesian grid was laid in at each site for horizontal control. An arbitrary
0-0 point was established off each site and a datixn of lOOS/lOOW weis staked
on-site. Sites vhich had previous grids established during testing of
22It539, 22It563, 22It576, and 22It590 were vised during data recovery. All
units Were designated by the northeast comer coordinates. Baseline grid
states were extended from the lOOS/lOOW point to aid in topographic mapping
and placement of cores and excavation units. Unless topographic or surfEu^e
features interfered, site grids were aligned with magnetic north. Vertical
control was established by setting in one or more bencdmarks at each site. If
an actual N6VD elevation could not be ixmediately established an arbitrary
100 m (330 ft) reference point was established for each site, and it was later
tied into U.S. Amy Corps of Engineers' benchmarks vdien possible.
A detEdled topographic map was prepared for each site using either a
transit or alidade and plane table. As excavation or test units were opened,
these were added to the beise map, or, if the site was ccmplex, a separate
excavation plan was made.
DATA REOOVEFY
SURFACE COLLBCTIONS
Most sites investigated in this project had either been previously surface
collected or there was no material on the surface. IWo of the sites tested
(22It622 and 22Mo675) were in plowed fields and had surface materials. The
surfEu:e collection at these sites was by a randcm stratified sanpling method,
vdiich selected units for 100%-timed collection of eurtifacts for a 20-22%
sairple of each site.
MECHANICAL EXCAV7Vri(»JS
At all of the sites during Phase I except 22Mo675 and 22It606, trenches
were excavated by a backhoe to expose stratigrajhic profiles prior to or
shortly after excavations had begun. The trenches provided an assessment of
site formation processes, means to delimit sites boundcuries, and an added in
the placement of excavation units. At 22Mo675 and 22lt590 a box-end scraper
was used to strip mechanically the plow zone in selected areas to determine if
features were present.
33
In Phase II, a badchoe was utilized at three sites (22It623, 22It634, and
22It621) to lenoove the overburden and eaqxase the cultural deposits selected
for investigation. Stratigraphic trenches were also excavated by a badchoe in
Phase II.
VISUAL MD OCRING
Systematic cxnes were taken at two sites, 22lt539 and 22It576, to aid in
excatvation unit placement and to locate subsurface anomalies. An Oakfield
3/4-ind\ Tube Saspler with extensions was used to remove 20 cm (7.9 in)
sasples frcm surftxm to the estimated base of the cultural deposits at
pacific intervals. Visual cxxres were examined in the field fcsr cultural
(xxitent and the presence of anomalies. Detailed notes concerning soil type,
texture, color, and OGapaction, as well as cultural content, were kept fcx
each core segpoent taken. Chemical cores were also taken at specified
intervals, visual information was recorded, the soil was bagged, and sent to
the field lab to be tested for pH, phoqhate, and calcitm carbonate levels.
Field lab chemical testing was performed only on samples &om the Poplar
(22It576) and Walnut (22It539) sites. The results bean the chemical test at
these sites did not warrant continued xise at other sites. The results of the
visual cores and backhoe trenches proved to be the best methods of locating
subsurface anomalies and determining xmit placement.
EXCAVATION IMITS
Excavation unit location was usually determined by information gathered
from research projects, surface collections, stratigraphic trenches, and
coring along vdth any topographic site features. Ar^tional units at the
larger sites were located primarily on information obtained in the field as
well as from stratigraphic information. At one site, 22Mo675, a oonbination
of randomly selected test units was ocaplemented by a judgementally placed
unit.
The standard excavation and recording unit was 2mx2mxl0cm
(6.6 ft x 6.6 ft X 3.9 in) in size and was removed by shovel. These units
were placed individually or in groups (blocks) as desired. Smaller units were
also excavated (1x2 in/3.3x6.6 ft, 1x1 in/3.3x3.3 ft, or 50x50 an/19. 7x19. 7 in).
The vertical dimension was occasionally divided into 5 cm (2 in) levels. At
one site, 22It606, during excavation, no esumvation units were used, rather
the middm was removed, and only features were eocamined.
Excavation blocks were lettered sequentially on the site. One adjustmmit
made in Phase II was that within each 2x2 m (6. 6x6. 6 ft) unit the four 1x1 m
(3. 3x3. 3 ft) divisions were excavated separately, and 5 cm (2 in) levels were
standard.
Features were sequentially nuilsered and handled as separate entitles.
They were mapped and photographed prior to, during, and eifter excavation.
Features other than burials were generally bisected and removed by trowel with
each hedf being separately processed by water flotation. Burials were neaped
and {hotogre^hed, and the bones were then carefully wrepped and moved to the
field lab for special studies.
SPECIAL SAMPIES
Severed types of special sanples were recovered in the project. The most
Sequent types were materials found i^ situ during the general excavation,
designated plotted specimens.
34
Radiocarbon samples were also taken from select ^ sitai proveniences.
These were removed with a clean trowel, placed in clean aluminum foil, then
placed in a plastic bag.
One or more 50x50 cm (19.7x19.7 in) "control blocks" were located in each
block of units. These were subdivided into four 25x25 cm (9. 9x9. 9 in)
sections. At every 10 cm (3.9 in) level a four^liter nacrobotanical sample
and two-liter perpetuity/soil sanple were taken. One quandrant vas also
fine-scneened thnough 1/16-inch (.015 cm) roe^. Originally, there were three
additional one liter soil sanples taken from the fourth cjua^ant of each
control block fen: pollen, biosilic:ate, and lipid analyses. Early in Phase I
the decision was made that these sanples cxxild be obtained from the two-liter
perpetuity/ soil sanples, if desired, so the quadreuit was reincorporated into
the general unit level fill. Two-liter perpetuity soil sanples were also
taken from the features and cu:ea and were plac^ in storage for future
studies.
Archaeoroagnetic dating sevrples were taken by archaecmcignetic cxmsultant
Dr. Robert DuBois at 22It539 and 22It576. Additional sanples were later taken
at 22It539 and 22It590 '^y Phase I staff personnel, who had been instructed by
DuBois. All sanples were dated at the Eaurth Science Observatesry at the
university of Oklahoma.
WATERSCREEN AND FLOTanON PROCESSING
All fill frem general 10 cm (3.9 in) levels wets processed through 0.63-cm
(0.25 in) hardware mesh. Fine-scn:een samples from cxmtrol blexhs and features
were passed throu^ a 0.15 cm (0.06 in) mesh as well.
Most feature fill and other special samples were prexsessed by flotation to
recover macrobotanical materials. The flotation machine used followed the
SMAP design after Watson () . One or two machines separated materials by
water agitation into three fracticxis: 0.63 cm (0.25 in) "A" fraction, 0.15 cm
(0.06 in) "B" fraction, and 0.5 cm (0.02 in or No. 35 sieve) "C" fraction.
FIEID RECORDING TECHNIQUES
MAPS AND SCALE DRAWINGS
The maps prepaired for each site included a topcsgrajhic map, a site
excavation plan itap, detailed drawings of stratigrajhic profiles, and flxxir
plan drawings of the base of each 10 cm (3.9 in) level in all units. The
flcxir plan drawings were attached to the apprepriate field form and traced in
ein on-site cenposite drawing of flcor plans for each block. Pre- and
post-excavation plan maps, as well as cross-seertion drawings, were cxitpleted
for features.
PHOTOGRAPHS
Photograjhic logs were kept at each site for both black-and-vhite prints
and cx>lor slides. General site (hotogra^s recorded all steps frem
pre-clearing, through excavation, to post-excavation views of each site.
Features were jhotogrsphed from pre- through post-excavation. Flcxsr plans
vhich exhibited unusual characteristics were photografhed, as were all drawn
stratigrajhic profiles, and all blcx:k unit profiles. A permanent catalcjg and
cuDss-index of all black-and-»hite prints and slides by subjeert and nunfcer wzis
made for each site. Each slide was catalogued individually and placed in
35
plastic sheets in notdsooks ordered site and slide nvx±ier. The
black-and-white contact prints were mounted an 5x7 inch cards with a copy of
the photograph log informatiGn and ordered by subject. Black-ancHuhite
negatives were stored in glassine holders in notdDooks by site and photograph
nixnber.
IDEMTIFICATiaJ WMBER SYSTEM
An identification nxirber (ID) system was used during both phases of data
recovery in vdiich each provenience and special sanple was given a sequential
site-^)ecific identification nunter. Tvo minor modifications of this system
were made in Phcise II: the assigiment of Master Identification Numbers System
which grocped multiple nurbers from the same general provenience (e.g.
feature) was dropped and ID nunhers were assigned in the laboratory rather
than the field.
FTETD POEMS
The most frequent form was the Level/Stratum form vhich was completed for
each 2mx2mxl0cin (6.6 ft x 6.6 ft x 3.9 in) level. This form recorded
beisic unit infonration (e.g. , site; block; unit; elevation) , associated ID
nunbers and types, soil information, and tlie excavator's observations and
corroents for each 2mx2mxl0om (6.6 ft x 6.6 ft x 3.9 in) provenience.
Bie feature form was similcur to the Level Form and recorded similar
information about features. If the feature was a burial, a burial number was
assigned, and a Burial Record form was oonpleted vhich detailed field
observations on body orientation, position, preservation, age and sex
determination, and component affiliaticm.
FIEID AND LABORATORY INTERFACE
All laboratory processing and classification for Ph£ise I and most of
Ihcise II was performed concurrently with excavation. The excavation of a
minimum of two sites simultaneously plus large field crew, produced a large
volvme of material each day. A critical point in processing and organizing
this volume of matericil was the deiily transfer frtm the field to the lab, and
a formal check-in procedure was developed to insvire an accurate accounting of
materials and acccnpanying paperwork.
LABORATORY PROCEDURES
METHODOLOGY
PROCESSING OT MATERIALS
Material from completed proveniences was tracked through the laboratory
through a series of check-out*". Washing was usually done outdoors behind the
laboratory but was also done in the lalx»ratory during inclement weather. The
wadiing process was l>asically a continuation of the waterscreening procedure.
A geurden liose with nozzle was used to spray water over each specimen until it
was cleaned sufficiently for classification. The specimens were then dried
outside or in a specially constructed heated dryer in the laboratory.
SOBariNG, CLASSIFICATION, AND CATALOGING
All cultural material vias initially rough sorted into four material
classes: ceramics, lithics (modified and debitage) , introduced rock, and other
(bone, shell, charcoal, and historic) . Each of these rough-sorted classes was
then processed s^>arately.
CERAMICS; (Ceramics were size-graded through 0.5 inch (1.7 cm) mesh. Those
greater than 0.5 inch (1.7 an) were sorted into types by tenper and surface
treatment. The ceramic types used in Phase I are defined in i^pendix I of
this report. Sherds exhibiting characteristics of vessel sh^je (rims, bases,
poded svpposts, ani handles) were separated and designated diagnostic. Sherds
in each type were counted, weired, and cataloged. Ceramics passing through
the 0.5 inch (1.7 cm) screen were classified as "sherdlets" and were
collectively weighed, and a 20% sample by wei^t was cataloged. Daub and
fired clay were included in the ceramic category, and they also were only
weighed, and a 20% sample was cataloged.
LITHICS; TlTe lithics in this investigation included modified lithics and
debitage and were classified into grotps by morphology, technology, and
function. The Phase I and II classification definitions are presented in
appendix II of this report. All modified lithics were classified, counted,
and cataloged. Lithic debitage included all flaked and fire-cracked chert or
chunks and was size graded through 1.0, 0.5, and 0.25 inch (2.5, 1.7, and 0.8
on) meidi screens. Each size grade Wcis sort^ by chert type and utilization.
The count and weight of each category were recorded, and a 20% sample by
wei^t was cataloged.
IWnoxxZED ROCK; Introduced rock was used to refer to rock vdiich did not
naturally occur on the site but did not exhibit any observable modification.
The specimens in these groups were identified lithologically, weighed, and 20%
were cataloged.
OTHER; This group included items v^ch were "other" than ceramics, lithics, or
introduced rock, such as bone, shell, charcoal, or historic material. The lew
amount of historic material, plus the lack of a developed historic
clcissification system in the earlier phases, necessitated a rough sorting of
this material, eind it was not computerized. Bone, shell, and charcoal was
wei^ied and cataloged. The historic material was counted, weighed, and
cataloged.
BOXING AND BAGGING
Most specimens were placed in sealed coin envelopes with provenience and
clcissification information labeled on each. These were bagged ID number
and placed into plastic-lined cardboeud boxes 1x1x1. 5 ft (30x30x45 cm) in
size. The boxes were orgeinized by artifact group and ID number. The boxes
were labeled on the outside and maintained in an organized manner in the field
headquarters prior to transfer to the University of west Florida
archaeological storage facilities.
The specimens selected for Phase III analysis were pulled from these
collections and are curated separately from than. They are organized by
provenience rather then ID number and cataloged by the most recent and highest
level of analysis.
CHAPTER V SITE EXCAVATIONS
INTRCOOCnON
Dlls chz^>ter describes the archaeologiccd investigations conducted at the
eleven sites included in this project and presents the results of the
pnelindnary analysis of cultural material conducted in I^iases I and II.
Oetedled studies of site soil, gecroorphology, ceramics, lithics, and botanical
information recovered from these sites vd.ll be presented in det^dl in
subsequent chapters and vd.ll be briefly characterized in this chapter.
EXCAVATICXIS AT THE POPLAR SITE (22It576)
The Poplar site was located in the active floot^lzdn of the l^per
Tombi^see Valley, 2.8 km (1.7 mi) northwest of Fulton in Itawanba County in
northeast Mississippi (Figure 1) and was situated approodmately 600 m (1,968
ft) west of the Veilley wall on a local topograidu.c mound 40x50 m (132x165 ft)
in area and 80-100 on (31.2-39 in) above the level of the floodplain (Figure
3) . The site was an isolated floodplain island Ixxmded on all sides lay
second-order streams and abandcxied channel segments of a ccirplex floo(p>ledn
swamp drainage system.
It appears to have originated as a point bar deposit in the flooc^lain
v^ch probably began as a truncated terrace outlier of the nearby valley wall.
The site consisted entirely of fluvial deposits.
Prior to testing in , the site was covered by mixed, second-growth
hardwoods with a thick understory of shrubs and vines. The dominant trees
were poplar (Populas sp.) , oak (Quercus sp.) , and hickory (Carya sp. ) .
FIELD METHODS
The Poplar site was initially recorded during Blakeman's survey (:19)
of the canal section of the Tennessee-Tcmbi^Dee WaterwEVr and the surface
collection indicated the presence of Gulf PormatiOTial and Woodland (Miller I
and II) carponents. The site was recommended for testing to evaluate its
information potential.
Testing conducted in January of (Dense ) consisted of three 2x2 m
(6. 6x6. 6 ft) units which were excavated to 140 cm (55 in) below the surface.
The excavation revealed a dark organic midden zone 90-100 can (35.5-39.4 in)
thick and a yellow-brown stratum directly below it. The cultural material
recovered indicated that the site liad been ocxnjpied from the Late Paleo-Indian
through the Mississippian cultural stages. Intacrt cxxiponents included
examples of Paleo-Indian (Quad) and Early through Middle Archaic (Benton) .
The upper 70 cm (26.7 in) appeared to be mixed. A fired clay heeirth
asscxriated with the Benton component, a pit cxjntaining Wheeler and Alexander
sherds, five human teeth (possibly asscx:iated with the Wheeler cxmpcxient) , and
a dog tjurial were also recovered.
Subsecjuent excavations were conducjted from February through September of
. Field methods used included: topographic mapping, visual and chemical
coring, bac)choe trencliing, Icirge and small excavation blocks, and ezpansicsn of
the test pits. Soils and sediments were augered and samples, and
monoliths were removed.
Seven hundred and twenty visual cores ta)cen on a 2 m (6.6 ft) grid over
the entire site surface provided the first data. Core depths varied from
60 cm (23.6 in) to c3ver 220 cm (86.7 in) , but always penetrated well talcw the
39
dark midden zone. The caring program was designed to identify subsurface
anonalies and strata to ensure that excavation units would be placed in
representative areas of the site. Hie visual core infooiation was hand
recorded and identified the location of charcoal^ fired cla^, artifacts, and
strata. Hie visual cores indicated that the base of the dark organic midden
was relatively level about 1 m (3.3 ft) below the surface and was underlain by
a yellow-brown silty clay zone. Cultural material (charcoal, flakes, silty
clay to sanfy clay, fired clay, ash, etc.) dropped out at the interface of
these strata.
Chemical cores were talcen ot a 8 m (26.4 ft) grid with two control
transects samples at 4 m (13.2 ft) intervals across the major axis and deepest
d^xjsits of the site (112S and 108. 5W) . Samples in the chemical cores were
taken every 20 cm (7.9 in) and were processed in the laboratory to determine
phosphate, and ccu±onate.
Hie pH and carbonate tests revealed no identifiable ancroalies within the
site. Phosphate content increased with depth and toward the center and
northwest portion of the site. Site readings were lower, and the phosphate
readings were higher on-site than off-site.
Four excavation blocks were ocnpleted at 22lt576 (Figure 3) . Blocks A, B,
and C were 4x4 m (13.2x13.2 ft) in size and placed on the inner third or
perimeter of the site aurea to investigate the site settlement pattern. Block
D, the largest unit (12x8 m; 36x24 ft) , was placed in the center of the site
to test the area with thickest deposits and to embrace the test pit vdiich
produced the Quad projectile point/knife. All units were excavated by hand to
sterile soil.
Block A was a 4x4 m (13.2x13.2 ft) unit, although a 20 cm (7.9 in) de^,
2x2 m (6. 6x6. 6 ft) extension was rtade subsequently to include a feature. Hiis
unit was placed in the southeast portion of the site to determine the nature
of the mound periphery both gecmorihologically and culturally. Previous
investigations in this emd other areas suggested that the edges of this type
of site could contain structures and provide information on site
morphogenesis. Proximity to the late Paleo-Indian material recovered in
test unit 114S/100W vas also suggrested this test site choice.
Block B was a 4x4 m (13.2x13.2 ft)unit located in the southern part of the
site. Hie unit was situated in an area of large grinding tools which were
recovered during the excavation of the privy just to the south. Further, the
site periphery questions addressed in Block A were also part of the goals of
Block B excavation.
Block C was 4x4 m (13.2x13.2 ft) unit placed to investigate the area of
high f^iosphate readings. Hie phosphate emcmaly was actually 2 m (6.6 ft)
south of the unit, but the presence of a large sturrp inhibited exact centering
over the anomaly.
Block D vus the largest unit excavated at the site. Hie placement was
based on the following factors; 1) the high phosphate area in the site center,
2) the presence of deepest and most frequent cultural material and strata, and
3) the presence of late Paleo-Indian naterial in the test unit.
Information from testing other similar sites in the eurea also had indicated
that the most informative cultural deposits of the midden mound were in the
areas of highest elevation (Atkinson ; Bense ; Calm , ) . This
block was excavated through Level 23 in the northwest comer 2x2 m
(6. 6x6. 6 ft) unit. The remaining western half of the unit was excavated
through Level 22, and the eastern half was excavated through Level 17.
Differences in excavation depths in this block occurred to ccaifirm the
culturally sterile deposits and to explore for possibly deeply buried cultural
materials .
The three test pits excavated v«ere extended to sterile soil in the
excavations. Test Uhit 100S/112M was extended 1x1 m (3. 3x3. 3 ft) in the
north wall to include a feature for removal.
Five stratigraphic trenches were excavated with a backhoe to investigate
further investigate the site sediments (Figure 3) . Trench placement was based
on information recovered in the excavation units. Trenches were placed
primarily on the edges of the site - a s^isitive area in determining site
formation. Block excavations provided a sufficient view of the interior site
dqxjsits and no backhoe trenches were placed there. Trench 2 was placed
off-site in the edge of the drainage canal to obtain maximun depth and to
e^qx^se the dqxssits of gleyed clay vhich underlie the Holooene deposits in the
Upper Tdrbi^cee Valley floodplain.
CUtTURftL RB1AINS
Cultural material recovered from the Poplar site include ceramics, stone
tools, debitage, ground tools, and animal and plant remains. The following
sections describe material recxjvered by artifact class.
A total of 16,289 sherds were recovered at the Poplar site (Table 3) .
Althouc^ the tarporally sensitive types were vertically mixed at this site, a
few patterns were observed. While sand-tenpered wares dominated all levels,
in Blocks A, B, and C, Wheeler Plain or Eroded Fiber ranked second in
frequency in all levels, with the exception of Level 1, Block C which
contained a greater proportion of Mississippian Plain. In Block D, the
Mississippian and Late Wbodland types Mississippi Plain or Baytown Plain were
second to sanchtenpered types in the \jpper 20 cm (7.9 in) of the site. This
oculd reflect differential site use during these periods.
TMU 3
Oeraiic fteqaencies by taeper, 22lt576.
Shell
1,305
8.0
Shell-Grog
363
2.3
Grog
1,920
11.8
Bone
207
1.3
Limestone
1,501
9.2
Sand
8,638
53.0
Fiber
2,355
14.5
Total
16,289
The relative frequency of the pottery types indicates that the site was
occupied throu^xxit the 4,000 years of the Gulf Formationed, Woodland, and
Mississippian stages. The frequency of tenper groups associated with each
stage may reflect relatively similar site activities during these stages. The
relatively high percentage (14.5%) of fiber-tempered sherds may indicate a
higher site use during the brief 500-year period they were in use.
A total of 5,081 chipped stone tool specimens was recovered from the site
(Table 4). Of these, 1,176 were projectile points/knives (over 23.1%), cores
(0.8%), preforms (3.8%), bifaoes (4.7%), scrapers (3.9%), and other unifaoe
and biface tools (56.9%) .
A total of 5,234 utilized flakes were edso recovered from the Poplcu: site.
Since the frequency of these ready-made, disposable tools is hi^ier than the
number of chipped stone tools there must have beer a steady and frequent need
for them.
42
Benton
Big Sandy
Big Slough
Bradley %>ike
Collins
Cotaoo Creek
Cypress Creek
Dalton
Elora
Eva
Flint River Spike
Gary
Greenhriar
Kirk Comer-Notched
Late Woodland/Mississippian Triangular
Ledbetter/Pickwick
Limestone
Little Bear Creek/Flint Creek
Mclntire
Morrow Mountain
Mud Creek
Residual Stemned
Residual Triangular
Savannah River
Swan Lake
S^ces-Nhite Springs
Tonnbic^bee Stemned
Snail unfinished Triangular
Vau^m
Subtotal
Cores
Preforms
Bifaoes
Scrapers
Drills, Perforators, etc.
Other Uniface and Bifaoe Tools
A total of 105,637 non-utilized flakes were recovered from the site.
Table 5 presents the frequencies l:y size and excavation block. There is a
high correlation between frequency and flake size. The 0.64 an (0.25 in)
debitage doninated (89%) debitage, vAiile 1.27 an (0.5 in) and 1.54 on (1 in)
non-utilized flakes consisted of only 11,308 (10.7%) and 291 (0.3%) specimens,
respectively.
With the relative proportion of chipped stone tools to dd^itage of 1:10
and the flake proportion of 1:39:323 of queu:ter:tmlf:one inch flaJces, chipped
stone tool manufacture and retcuching are docixnented throu^xxit site
occupation.
43
Flake size
1.0 inch
0.5 inch
0.25 inch
Total
291
11,308
94,038
105 637
10.70
89.02
A totel of 1,735 ground stone tools were recovered and include both ground
and polished stone artifacts. As Table 6 shows, most specimens were
unidentifiable fragments (1,386 or 80%). The most frequent identifiable
ground stone categories are flaJces of ground stone tools (n=67) , ground
limonite (n=56) , hanmerstones (n=52) , and pitted anvilstones (n=33) .
22It576.
Fri
Gkound stone tool
Hamnerstone
Pitted Anvilstone
Hatimer /Anvils tone
Abrader
Muller
Mortcur
Pestle
Atlatl weight
Discoidal
Bead
Sandstone sherd
Worked sandstone concretion
Ground limonite
Ground hematite
Edge-ground cobble
Other (ground flake)
Muller/Pitted Anvilstone
Drill Core
Bead Preform
Muller/Hartmerstone
Soapstone
Ground projectile point/knife
Tubuleu: pipe
Mortar /Anvils tone
Pitted Anvilstone /Abrader
Unidentified ground stone fragments
The vast majority (88.4%) of the 741,727 grams of introduced rock from
22It576 consisted of sandstone, followed by fire-cracked chert chunks (4%) ,
and cobbles (1.4%) . Use of the dominant ferruginous sandstone cannot be fully
explained; however, it is the most abundant locally available rock and the
only type in the l^;per Tanbigbee Valley that occurs in large flat slabs.
Ihe few (176) historic specimens recovered from this site include hunting
and fishing inplements (13) , metal container freigments (64) , faisteners (13) ,
44
wire ludls (77) , and ndsoellaneous material (9) . Most of the historic/nodem
material was reooveied in the mixed upper 50 an (19.7 in) of the site deposits
and historic intrusicns such as potholre. The range of reflects the
historic activities which are known to have occurred at the site, i.e. :
logging, cultivation, pig oantaiment, relic collecting, hunting, and fishing.
Fau]^ frarpants ocnslsted primarily of calcined and charred fia^aents and
cnly 609 (12%) of the 4,953 analyzed were identifiable beyond the clus level
of identification , primely because of the emnll size of the firageents.
Ooneequently, the faunal information is tiest viewed as qualitative rather than
quantitative. The identifiable faunal material included both large and snail
manwals, birds, i?eptiles (especially turtles) , and fish. Btost of the faunal
material from the midden cene fron Block D and in the tapper 50-60 cm
(19.7-23.6 in) vhich contained mixed Gulf Formational thrcu^ Mississippian
temporal markers.
FEATURES
The investigations encountered 93 features which were classified into ten
feature types:
8 chipped stone clusters
9 rock clxisters
29 fired aggregates
1 hearth
2 prepared areas
43 pits
1 ceramic cluster
The cultural affiliation of the features was determined by diagnostic
artifacts and/or stratigraphic position of origin. If a feature did not
contain ai^ diagnostic or tenporal markers and the stratigraphic position was
unclear, the cultured eiffiliation was classified as unknown.
Only one oeramic cluster was encountered at this site. It consisted of a
major portion of one plain grog and shell-tempered vessel. The vessel was
lying iprl^t just below the plow zone and the vpper portion, incliKiing the
rim, had been removed by the plow.
Of the ei^t chipped stone clusters, one was a cache of thinned blfaoes
("quarry blades") made of Fort Payne chert. This cache of blades (Feature 10)
contained four bifaoes and one piece of sandstone. The bifaces were similar
in size and shape, approximately 11 cm (4.3 in) long and 1-5 mm thick. While
the context was mixed with markers from the Middle Wbodland and the Middle
Archaic (Benton) , similar features from similar sites have been documented as
associated with the Benton occupation and this is llkaely the case here.
The remaining chipped stone features were oonomntrations of chipping
debris and broken tools. TVio (Features 9 and 16) were located in the upper
levels (5 and 1, respectively) , and their cultural affiliation is unknown.
The remaining five chipped stone clvisters (Features 113, 116, 118, 119, and
120) , were located in the western portion of Block D in the Early Archaic
oemponent. The largest and most dense clusters (116 and 118) were within a
2 m (6.6 ft) area at the same elevation and contained a variety of complete
tools inclviding a side scraper, a flake knife, and a core. Additional
materials consisted of broken bifaces, utilized flakes, and over 200
non-utllized flakes. Both clusters were extremely concentrated and appeared
to be piles. This was not the case for all the chipped stone cluster
features, since 119 contained only 33 specimens, including (Xily one
45
unidentified tool and a utilized flate. iUK)ther (120) was a linear
arranganent (60x20 oo or 24x7.8 in) of a anall nuniaer of tools and flakes. An
additional Early Ardiaic cMpped stone cluster was present in the narthwest
comer of Block D at Level 15 but was not noticed in excavation, hcwever the
5 on (2 in) levels applied sufficient control to reconstruct the feature.
The Early Archaic features appear to be the results of tool production,
maintenance, use, and breakage. Concentration of this activity in the center
of the site is stipported by the absence of these features in other parts of
the site at this level. It is likely that this part of the site was tii^iest
{epixxxttBc) during this occipation.
Nine rock clusters were identified at 22It576. These features were
generally oval and ranged from 2.42 m (8 ft) to 0.17 cm (0.7 in) in diameter.
Th^ were concentrations of unmodified fist-sized pieces of san^tone. Other
cultural material included in the rock clusters consisted of ceramics, chipped
stone tools, ground stone tools, flakes, fire-cracked chert, and fired clay.
The range of material appears to Ise related to size of the feature, i.e. , the
larger the feature, the more diverse the cultural material.
These features were IDcely related to fire liearths, rock ovens, and/or
refuse deposition. Four were associated with other fire-related artifacts,
such cU5 flre-crac)cBd chert. The remaining rock cluster features which
contained a wide variety of tools were likely related to refuse deposition.
TVmnl^-nine fired aggregate features were encountered. Most were
recovered in Block D (23) , Block A (6) , and with only one in Block B. The
fired aggregates were ocnposed of hard fired orange-colored silt loam and
averaged 9 cm (3.8 m) thick, 48 cm (1.6 ft) long, and 36 cm (1.2 ft) wide.
They were devoid of artifacts, charcoal, and ash. Associated cultural
material was quite difficult to docunent because encircling stains or other
mechanisms of association were leaving in the surrounding dark, culture-rich,
organic midden. The silt loam was brou^t to the site from the adjacent
wetlands, and the highly oxidized, burnt orange color of these features
suggests that they were specially pr^>ared hearths.
A basin-shaped hecuth encountered in Block A was 36 cm (1.2 ft) wide,
62 cm (2.8 ft) long, 10 cm (3.9 in) deep and was composed of gray, burned
clay. The dark brown midden fill in the basin contained eight fiber and
sand-tenpered sherds, two ground stone fragments, 13 flakes and sandstone. If
this fill was in situ, it appeared to be associated with the Gulf Formational
stage.
TV? prepared areeis were encountered at 22It576: Features 44 in Block D
and 49 in Block A. Both were characterized by a mottled yellow, clay loam
matrix with areas of oonoentrated clay. These features were actually a mosaic
of fired areas, fired aggregates, and scattered fired clay fragments. Only
Feature 44 was ccmpletely exposed and measured 4. 4x3. 2 m (14.4x10.5 ft) and
was 5-10 cm (2-4 in) thick. The portion of Feature 49 which was exposed
measured 5x4.6 m (16.4x15.1 ft) and was 23 cm (9 in) thick. The two prepared
areas varied in consistency. Feature 44 was not ocmpact, had diffuse edges,
and contained significant quantities of charcoeil. Feature 49 was ocnpact
throughout, had little charcoal, and had the appearance of being cleaned.
Both prepared aureas had fired aggregates, pits, postmolds, aixi burials
associated with them. It is inferred that they were centers of intense
cultural activity.
Forty-five pits were enccuntexed. The cultural material in the pits was
not different from that contained in the surrounding itatrix. Pits were
generally cliaracterized by a dark color of the internal fill and concentration
of cultural materiad. Thirty-two (71%) of the pits could be associated with
46
the S^ces-Whlte ^ings/Benton ocxpxient. Only one pit could be affiliated
with tile Eva Morrow Mountain oonponent, vMle ei^t pits extended into the
Kitic zone, these ei^t pits contained no diagnostic natarial, and it is
postulated that they aze lower portions of pits intruding trcm later
ocsgonents. ^nio additional pits %wze identified in tiie vpper 30 cm (10 in) of
the dark midden, and cultural affiliation could not be dstenained.
Four historic intrusions were identified, all of then in Blodc 0,
including two apan potholes on the surface. •Oao otiwr filled pothole were
identifi^ during excavation. One large filled pothole %fBS 1.4 m (4.6 ft)
dsep, with a burned fenoepost standing upright vdth in it.
One stain feature was recorded at this site. It was circular and
lens-shzged with yellow mottled soil containing material similar to the
airzounding midden. These deposits wes^e so mmerous and undifferentiated fraa
the surrounding midden in this and severed other midden mound sites in the
project that it soon became inefficient to separate those lenses of
yellow-brown soil. Their function %«bs not determined, aixi ^peculation
suggests that they were probably accidentcdly l>rought on site from the
adjacent wetlands.
ElevKi burials and three cremations were recovered at the Poplar site.
The majority were located in Bloc)c D, with one burial ea<^ in Blodc A and test
pits 99S/112W and 119S/113W. Burials were usually discovered liy e]gx»ure of
sloeletal parts, a dar]c stain above the skeleton, or a concentration of fired
silt loam. Burial pits were not discernible in the dark midden and were
poorly defined even in the yellow-brown zones below. The average depth of the
portions of excavated burial pits was 25 can (9.8 in) . The deepest laurial pit
was 73 cm (2.4 ft) . A suninary of the skelet^ information is presented in
Table 7.
Burial analysis, 22It576.
Burial
Position
Side
Facing
Direcrtion
Aqe/Sex
Asscxriated
Materials
3
Unknown
Unknown
Uhknown
Adult
5
Flexed
Left
North
Elderly female.
6
Flexed
Back
unknown
40+
Adult
7
8
Flexed
Unknown
Ri^t
Ihknown
North
Unknown
Adult,
10
Unknown
unknown
Unknown
probable male
Adult
11
Flexed
Right
South
Female 30
12
Extended
Supine
Northwest/
Adult male
C^xhe of
16
Extended
Supine
southwest
North/ south
Ahilt male
tools
Dog ticull
Of the eleven burials, four were flexed, two were extended, and the
position of five was undetermined. The skeletal material was £Q.weys in a
soft, friable cxndition, frequently iapossible to remove intact. Of the eic^
skeletons examined, three were males, two were female, and the sex of the
three remaining was not ascertained. The burials probably occurred during the
Middle Archaic oocn:pations of the site (Eva/Morrow Mountain and Sykes-Mhite
Springs/Benton) . Burial goods , Including a dog skull and a c:ache of chipped
stone blfaoes, were present with two adult males.
47
SIRATIGBftPHY
The Poplar site was ccxoposed of loanv fluvial sediments. The site appears
to have fooned as a point bar of a Holocene stream formed around and on
ocnpact basal eroded Pleistocene terrace outlier remants. The course of the
stream weis then directed euxund the outlier fozming the beise causing
deposition on the side of the bend, creating the point bar. There were eight
relatively flat strata identified in this site (Figures 4 and 5) with tapered
or lens-shaped edges. The strata were sQJ. present and thickest in the center
of the site thinning erratically to the point of extinction near the out
margins. The strata were organized into a dark brown organically stained
midden zone (Strata I-V) and a yellow-brown pcdeosol with well-developed
structure (Strata VI-VIII) . The doninant texture of all deposits was loam
v^ch iaplies that the deposition was water flcwijig at a moderate velocity.
CmaiOMETKEC DATING
Seven radioceu±)on and two Eurchaecroagnetic samples were from 22It576
(Table 8) . The radiocarbon samples were taken front control blocks of the
midden d^sits in Block D. T\«o of the saiples were from culturally mixed
deposits, two from the Sykes-White Springs/Benton, one from the Eva-Worrow
Mountedn, and one from the Early Archaic ocnponent.
•DIBIE 8
Radiocarbon and arctiaeamaation or
beginning of the Eva/Morrow Mountain ooc\:q[)aticn. Regardless of context, the
dates do increase in age with depth below the surface, illustrating a
chronological stratified sequenoe.
TWO aurchaecmagnetic sanples &om two hearths were submitted fixm the
Poplar site. The hearths were associated with each of the two large prepared
areas identified at the site. Feature 58, was a hearth in the prepared area
oonplex in Block A (Feature 49) on the eastern edge of the site. This oonplex
feature was constructed of ccnpact clay, was well defined, and oont2dned a
flexed burial below the clay surface, clay-lined poet mold, and six hearths.
The atxhaeomagnetic date of -75 B.P. was obtained from it, and it
agrees with the documented date range of the Benton diagnostic markers
associated with the feature complex.
The second archaeomagnetic date obtained was 4,450 B.P. from Feature 50,
one of nine hearths cissociated with the complex prepared area in the center of
the site. Feature 44. This feature complex was less compact than Feature 49,
and the edges were more diffuse. The feature was mottled with stains and
charcoal, unlike the homogeneous clean surface of the other prepared area.
The date of 4,450 B.P. does not agree with the associated Benton diagnostic
markers vhich have been consistently dated at 5,000-6,000 B.P.
This feature was difficult to define, and the association of the hearth
with it was not ironclad. Two of the radiocarbon dates frcm the control block
in the midden 5 m (16.5 ft) away were taken 10 cm (3.9 in) above and 10 cm
(3.9 in) belcw the level of the feature. The date frcm 10 cm (3.9 in) above
the feature was 5, B.P. and the date frcm 10 cm (3.9 in) below the
feature is 5, B.P. The interpolated date of this heartJi was
approximately 5,700 B.P. The incorrect archaecmagnetic date probably was
related to the newness of the interpolated curve of the magnetic pole drift.
Logical inference suggests an appropriate correction.
SCMIRRy
The multi-ocmponent ®oplar site was situated on a small natured elevation
(50x50 ro: 165x165 ft) in the flooc^ledn of the Upper Tcirbi^aee VatLley. The
site was occxpied for approximately 10,000 years by all recognized prehistoric
caltural groqpe. Four large excavation units in areas of the site were
investigated during the ei^it-month fieldworic period.
The excavations ptoduoed the expected cultural oanpcnents and ocnfimed
that the Benton through Kirk Archaic deposits vwre intact. The expected
Faleo-Indian ocnpcnent was not encountered, and the oonplete Quad point
recovered in testing vies associated with the Kirk ccnponent and apparently
out-of-oontext.
The most significant results of the investigations at the Poplar site
included: 1) the isolated Early Archaic Kirk assenbleiges, 2) the plethora of
site facilities during the S^ces-White Springs/Benton occupation, and 3) the
correlation of Early Holocene (5,000-10,000 B.P.) site use changes with the
established climatic and soil dynamics.
Ten ccnponaits were identified at the site which included the prehistoric
period from the Early Archaic through the Mississippian stages. Only three of
these had both intact midden and features; i.e. Kirk, Eva Morrow Mountain, and
S^CBS-White SEsrings/Benton. Three ocnponents had intact features but lad^
midden deposits: Wheeler, Alexander, and Miller III.
Archaic Stage; The Kirk was the first period of oocv^tion of the Poplar site
and was contained within the Early Holocene peLLeosol v^ch contained three
identified zones (Strata 6, 7, and 8). The Kirk occupation was identified
only in Block D.
Features associated with the Kirk ccnponent include five chipped stone
clusters and ei^ht pits. The chipped stone clusters contained a variety of
oosplete and broken tools incliading scrapers, knives, cores, bifaoes, and
utilized flakes, as well as considerable d^itage. It is likely that an
additional xmrecognized chipped stone cluster feature was also present in the
northwestern unit of Block D in the Early Archaic zone which contained over
800 flakes. The pits contained only flakes and introduced rock.
The Kirk cultural material was concentrated in two layers vAiich were
separated by 10-15 cm (3. 9-5. 9 in) of relatively sterile sediments. These are
labeled the "lower assenrblage" and the "upper asseeblage." Both eactended
across the entire block. The "lower assenblage" contained six projectile
point/knives: Kirk (3) , Big Sandy (2) , and Greenbriar (1) . The "upper
assemblage" contained only two Kirk projectile point/knives.
The Early Archadc assenblage consisted of both ocmplete and broken tools
including projectile point/knives, scrapers, bifaoes, ^xdeeshoves, drills,
perforators, abraders, flake knives, adzes, splinter^ Tiwdgep (piece
eagudlles) , ground stone tool fragments, and utilized flakes. Hon-utilized
d^itage~was abundant (4,581) as was introduced rock (3,290 g) .
Features and midden material of the Early Archaic ccnponent suggest sene
preliminary statements about the nature of the hunan activities at the site.
Hunting and gathering food and hide, processing, tool production, and
woodworking were all likely occupations of people vbo occipied the site. The
activities inferred from the cultured material of the Early Archaic Kirk
ccnponent indicate that this locale was most likely used as a seasonal canp.
The initial Middle Archaic ccnponent was restricted to the vpper porticn
of the paleosol unit vhich had been truncated by erosion. Subsequent
occupation of the site probably caused seme mixing of the eiqxosed midden.
This inference is drawn from the presence of a few Sykes-White Springy and
Benton projectile point /knives in this ccnponent. TWo fired aggregates, one
pit, and one buried were associated with this ccnponent. Stratigraphic
position suggests that two rock clusters, three fired aggregates, three pits,
and one burial were associated with this ccnponent, but such an association
cannot be firmly established.
52
nie Eva/ttorroii ttountain midden deposits exxitained a vride range of diipped
and ground stone tools. Biotic remains, especially hickory nut^lls, and
vooA charcoal were idso associated with this conponent. It appears that
during this oocipation the site was edso used eis a seasonal camp but the
more features and materials were left bdiind. Activities oondocted included
hunting, gathering, food processing, tool production and maintenancse,
woodNorking, cooking, and inhanaticn with grave goods.
The S^ces-White l^ings/Benton ooccpation of the Poplar site was
measurably different than any prior or subsequent ocnponent. Die
distributions of Sykes-White ^>rings and Benton projectile point/knives at the
Poplar site are similar in initiation and frequency peak. Ihis similarity
causes these markers to be considered as a single cultural manifestation.
Ibis pattern was also seen at the Walnut site. Tbe primary characteristics of
this ocnponent included:
1. Activity centers or prepared areas vdiich were of fired clay loam from
2.6-4 m (8.6-13.2 ft) in diameter. Fired aggregates, pits, postholes,
burials and cremations were associated with these.
2. Ibe 30 cm (11.8 in) midden deposit had extremely liigh ooncentratiQns of
charred floral and faunal remedns, especially burned hickory nutshells.
3. Ibe midden also had dense artifacts and fired clay.
The majority of features at the site resulted from this occi^tion. The
range and type of these features indicate that the site wbis used for a
long-term, multipurpose base camp.
Ibe midden resulting from this ocnponent was distinguishable from other
portions of the dark organic zone ly the darker color deriving from the high
density of charred wood and hickory nutshells. The major feature ocnplexes
occurred vertically within 10 cm (3.9 in) . The "midden zone" of this
component was not well defined, and had many specimens from later occupations.
The designation of the 30 cm (11.8 in) as the Sjkes-White ^rings/Benton
"midden" was based only on the frequency peak of the Benton and Sykes-White
brings projectile point/knife types and the high nuntser of affiliated
features.
One of the patterns chserved in the associated midden material was an
increase in the amount of tools and debitage of Port Payne chert. The small
size of the debitage, lack of cores or preforms, presence of caches of quarry
blades, and a high proportion of projectile point/knives indicate that this
raw material was introduced to the site in the form of prepared bifaces and
that it was conserved through rejuvenation and slarpening. The increcise of
Benton projectile point/knives and the conccmitant decrease in other tool
types suggests the multiple uses of Benton projectile point/knives. This
increase in use of this foreign raw material for basic stone tools is
consistent with other Sykes-White Springs/Benton ccrponents in the Upper
Tcmbi^aee Valley.
The Popleu: site wsis occv^jied during the Gulf Formational stage, including
both the Wheeler and Alexander horizons. Unfortunately, the deposits
COTitaining the cultural naterial had been previously disturbed. Temporally
sensitive Wheeler and Alexander series ceramic types were present in
significant nurbers. Three features were possibly associated with the Gulf
Formational occvpations. These included a heeurth and two rock clusters. All
three features had associated diagnostic ceramics; hcwever, the midden weis
culturally mixed and the ceramics could have been intruded into them
accidentally.
53
nie hearth contained many mactdootanxcal specimens, including pieces of
Qircubito 3:ind. If this feature was fxon the Gulf Fomnational, it represents
the earliest docunentation of horticulture in the Upper Tcnbi^^ Valley.
Both Nheeler and Alexander series sherds were associated with the hearth.
The Poplar site was oocipied during the Woodland and Mississippian stages.
Hcwever, the integrity of the deposits destroyed, and little can be said of
the nature of the oociqpation. Thousands of sher^ front the Woodland and
Mississippian periods were recovered; however, only one ceramic cluster (a
large portion of a shell-grog vessel) (Feature 122) was encountered.
EXCAVATiaJS AT THE WAUWT SITE (22It539)
The Walnut site located in the l^:per Tcmbigbee River floodplain 16.5 km
(26.4 mi) north of Fulton, Ms (Figure 1) contained deposits front the Early
Archaic to present. This site was situated in the eastern part of the
floodplain about 750 m (2,475 ft) west of the valley wall. It was a prottinent
topographic feature elevated above the surrounding lowlying floodplain subject
to flooding during winter and spring months. The site was surrounded by
wetlands incli:iding swamps and small sloughs. The coarse texture of the sandy
loam soils in the mound indicated higher energy depositional events than those
vhich characterized the surrounding silty sediments.
The Walnut site was rou^ily oval (Figure 6) 100x70 m (328x229.7 ft) and
elevated 1.8 m (5.9 ft) above the surrounding flooc^lain. The central portion
of the site surface Weis relatively flat. The steepest slope was on the
northwest. The landform likely formed as a floo^lain bar, probably initiated
by an outlier of the nearby valley wall. Prior to testing, the Walnut site
was covered with a hardwood forest including hickory, oak, and walnut.
Clearing, construction, and maintenance operations fear the TVA
transmission line had disturbed the central portion of the site (Figure 6) .
Earlier disturbances occurred because of logging, agriculture, and looting.
Potholes of varying sizes and ages were scattered throughout the southern,
wooded part of the site. The most disturbed areas were the southwest quadrant
and euround the transmission tewer.
FTKTJ) methods
The Walnut site was recorded in the first waterway-related survey of this
cuea in (Lewis and Caldwell :44) , and was known to the loceLL
collectors for many years. Lewis and Caldwell () , as well as Blakeman
() , noted the temporally diverse materials on the site surface and
recommaided that the site be tested. Testing was performed on the site by
Bense () . The test investigations included a controlled surface
collection, i.e. two 2x2 m (6. 6x6. 6 ft) hand-excavation units and four backhoe
trenches. This information (Bense :386) indicated that there were intact,
stratified Middle and possibly Late Archaic deposits at the site.
Excavation was conducted between M2u:ch and September of . Chemical
and visiaal cores, test pits, backhoe trenches, and large excavation blocks
were all employed during site excavation. Systematic coring was the first
activity performed at the site to aid in determining excavation unit
placement. A total of 1,468 cores were examined on the Walnut site including
visual cores at 2 m (6.6 ft) intervads, chemical cores at 8 m (26.4 ft)
intervals, and two additional chemical core transects at 4 m (13.2 ft)
intervals.
54
I
i
!
Visual cores recorded charcoal concentrations fired clay ancnalies, and
stratigraphic boundaries. Using these data two large excavation units were
positioned directly over large fired clay areas. Excavations, however,
exposed two additional large (greater than 2 m or 6.6 ft diameter) fired areas
that were not identified in visued coring. Subsequent developments
dencnstrated that visual coring was not as informative for tracking subsurface
anomalies as it had been at the Poplar site.
Chemical cores identified anomalies of pH, phoqhates, and carbonates.
Sknaller excavation blocks were centejred over two chemical anomalies to
eiaoertain their origins.
Pour Icuge excavation blocks, A through D, were excavated on the Wednut
site (Figure 6) . Block A ves a 4x4 m (13.2x13.2 ft) block placed on the
wooded southeastern slope of the site in an eurea of low and high levels of
phosphate suggesting of a potential trash disposad area. No visual core
anomalies had been identified in the imnediate area.
Block B was placed over a distinct fired clay deposit approximately 60 cm
(23.6 in) below the surface observed in visual coring. No chemical anomalies
were recorded. Ihe block also served to sanple the south-central area of the
site. The block was initiated cis a 4x4 m (13.2x13.2 ft) unit but vas later
expanded to an 8x6 m (26.4x19.8 ft) block in an attempt to identify activity
areas associated with the fired clay feature.
Block C was placed over the largest and most well-defined fired clay
ancmEdy, which had been located in visual coring. Ibe fired material observed
in the cores wcis 5-20 cm (2-8 in) thick with abundant charcoal in the
surrounding azea. Numerous additional visual anomalies were observed within
15-20 cm (6-8 in) of the fired clay area, suggesting that this was possibly an
area of intense activity. Block C also served to sanple the deposits in the
northern and hipest area of the site. 11x18 began as a 12x12 m (39.6x39.6 ft)
block but was reduced to 10x10 m (33x33 ft) after the fired clay area was
better defined.
Block D was located in the center of the site over a chEuxx>al
oonoentration identified in the visual coring. Initially, a 2x2 m
(6. 6x6. 6 ft) test pit was placed over the cmcnBly (118S/103W) . Ibe results
prompted excavation of an adjacent 4x4 m (13.2x13.2 ft) block, which was later
eipanded to 6x8 m (19.8x26.4 ft) , when it became evident that the complex
feature excavations in Block C would prevent that block from being excavated
to the bottom of the site within the allotted time. Materials in the upper
meter were removea with a backhoe to the deposits below the feature zone
(Benton) which were then hand-excavated to sterile soil.
To obtain representative samples of all ccnponents of the site and to
avoid over-representation of certain components at the expense of others, all
blocks were not excavated systenatically from the surface to the base of the
site. In addition to the mechanical removed of the upper neter of mterial in
Block D, the sediments below Level 13 of Block C (the Sykes-White
Springs /Benton zone) were removed with a bex:khoe to determine if there were
pits and burials vbich extended into the lifter colored sediments below the
site.
Dturing the first three months of excavations, it became progressively
apparent that the leirge site size made it impossible to aissume that the three
major blocks under excavation (Blocks A, B, and C) were fully representative.
Four additional 2x2 m (6. 6x6. 6 ft) test units were developed to be sure that
the physical and cultural stratigraphy were better understood. These were
Icxated on the northeastern slcpe of the site (102S/87W) , on the extreme
southeastern edge of the site (146S/69W) , on the western slope of the site
56
(122S/146W) , and in the central area of the site (118S/103W) . This centred,
test pit to the placement of Block D. To gather mace data on the IcMer
levels of the site at Block D, an eu:?ed area each associated with the sene catponent. Four prepared curesis
were identified in Blcxh D, and all were associated with the earlier
Eva/Morrcw Mountain conponent. These were sanethat smaller than the
Sykes-White Springs /Benton prepared areas.
Sixty-one pits were identified at the Walnut site. While the cultural
affiliation of 28 could not be ascertained, 28 were associated with the
Archedc cxxupaticxi of the site, 22 of which were Middle Archaic. Only one pit
could be associated with the Gulf Foxmational ccnponent, three were affiliated
with the Late Wbodland, and one w^ls a Mississippian pit. Pits were the most
frequent single feature type encountered at the site (65%) , and all appeared
to be for refuse disposal.
61
other features encountered included tm ceramic clusters of Late Woodland
and Mississifpian types in Block B. One Block C concentration of chipped
stone in the Sykes-White Spring/Benton zone. Only one yellow stain was
removed as a feature, since it became clear early on that these amorpAnus
phenomena could not be properly isolated.
Seventeen burials and one cremation were discovered at the Walnut site.
Six buricds were located in Block A, seven viere in Block D, two burials and
one cremation were in Block C, and two burials were in Stratigr2^Mc Trench 2.
Most of the burial pits were identified at the base of the dark midden and
extended well into the yellow polygonal soil (Zone VII) . Hie burial pits were
over 2 m (6.6 ft) from the surface at their deepest point. Most burial pits
appeared to originate in Zone VI, the Middle Archaic occupaticxi zone. Burial
and s]celetal attributes such ais age, sex, position, orientation, type, aixi
artifact associaticxis are described in Table 14 vrfiere possible.
1SB[£ 14
Burial
Number
1
Age
Sex
Ty^
Grave
Goods
1
Block A
Indet.
Indet.
Extended
Primary, Multiple
Absent
2
Block A
Adult
Indet.
Extended
Primary, Multiple
Absent
3
Block A
Unknown
Ikiknown
Unknown
Unknown
Unknown
4
Block A
30 yrs?
Female
Extended
Primary, Single
Absent
5
Block A
40 yrs?
Male
Extended
Primary, Multiple
Present
7
Block C
Adult
Male
Flexed
Primary, Single
Absent
8
Block A
Young
Indet.
Indet.
Indeterminate
Absent
9
Block D
Indet.
Indet.
Flexed
Primary, Single
Present
10
Block D
Adult
Indet.
Exteaxied
Primary, Multiple
Absent
11
Block D
Indet.
Indet.
Indet.
Primary, Single
Present
12
Block D
Indet.
Indet.
Indet.
Primary, Single
Absent
13
Block D
Indet.
Indet.
Extended
Primary, Single
Absent
14
Block D
Adult
Indet.
Extended
Primary, Multiple
Absent
15
Trench 2
Indet.
Indet.
Extended
Primary, Single
Absent
16
TrMich 2
Indet.
Indet.
Extended
Primary, Single
Absent
17
Block D
Indet.
Indet.
Indet.
Indeterminate
Absent
18
Block C
Unknown
Unknown
Unknown
Unknown
Unknown
19
134
Adult
Indet.
Unknown
Cremation
Present
Althou^ buriads were found in most areais investigated in the site, two
areas of organized buriads were recognized, one each in Blocks A and D. The
Block A burials contained six individuals, while Block D probably contained
ei^t individuals. All burials were in the extended position except the
cremation (Burial 19) . Hie bodies had been placed in elongated, narrow pits
which were organized into rows, and the rows were oriented north-south (Figure
8) . Hie burial pits had both single and multiple interments. In Block A
multiple burials were more oonmon, and in one instance two individuals had
been laid directly on top of one aixjther (Burials 1 and 2) . Figure 9 shows
Buried 1, vrtiich was above Burial 2 in the same pit. The extent of the
cemetery eunea in the vicinity of Block A was not determined, and two burials
extended into the profiles of that unit indicating that burials extended
beyond the areas investigated. Block D burials were well defined, suggesting
a pattern of burials in rows (Figure 8) . Block C and Stratigraphic Trench 2
burials may have been in organized units, but that is uncertain because of
excavation constraiints.
62
The cremation in the Block D burial area (Figure 10) was a well-defined,
charcoal-rich area ^jpraximately 50x60 an (1. 6x2.0 ft) and 28 cm (11 in) thick
vMch contciined fragments of burned bone, apparently from a child, along with
three grooved stone effigy beads, including one of a crested bird.
Skeletal preservation was extremely poor, making age and sex determination
difficult. TBie age and sex of only four burials could be determined
(Table 14) . These included one 30-year-old female, one 40-year-old male, one
"adult" male, and one "young" male or female. Four other adults and one youth
of indeterminate sex were identified.
Artifacts were associated with four burials. The most definite
eissociations were with Burial 9 in Block D (one itiuller /pitted anvilstone, one
hanmerstone, one muller/hanmerstone) , Burial 11 in Block D (one Sy)ces-White
Springs projectile point/knife and two projectile point/lcnife fragnnents) , and
Burial 19 in Block C, a cremation, (one zoonorphic, one tubular, and one
discoidal stone bead) . Burial 5 in Block A contained a concentration of small
quartzite pebbles near the chest cavity v^ich could have been part of an
organic-cased rattle. Several other burial pits had artifacts in the fill not
associated with the body.
STRATIGRAPHY
Seven najor zones described in Figure 11 were recognized at the Walnut
site. The strata were relatively level in the central two-thirds of the site,
while on the outer one-third they were occcisionally pinched out or had a
downward dip. Ihe strata were thickest in the central two-thirds of the site.
The profile vas characterized by an upper dark midden zone 1.3-1. 5 m
(4. 3-4. 9 ft) thick corposed of smaller (^positional units. The lower portion
of the profile was a yellow-brown sanciy paleosolic loam with a single
recognizable <3epositional unit (VII) .
CHRONOMETRIC DATING
Eleven radiocarbon dates were obtained from this site (Table 15) . With
one exception, all samples produced radiocarbon (^terminations of between
8,000 and 5,000 B.P. The one sample from mixed provenienc:e yielded a (Sate of
4,594 B.P. The remaining dates correlated reasonably well for the Eva/Morrcw
Mountain, Sykes-White Springs, and Benton occupations.
T»BI£ 15
Radiocarticin dates, 22It539. _
Uncerrected
Lab #/
Field
Number
Blcxrk/
Level
C-14
(BP)
(Correcrted)
Calendar
Age (BP)
Archaeo-
magnetic
Date
Material
Cultural
Affiliation
DIC-/
539-
A/3
+95
(-)
nutshell
Late Archaic:
Little Bear
Creek
DIC-/
539-
A/7
+75
()
nutshell
Benton?
DIC-/
539-
B/6
+75
()
nutshell
Benton
66
nB£ 15
BafliocBTigp dates, 22lt539 (oaatinaed) .
Lab #/
Field
Niariber
Block/
Level
C-14
(BP)
Unoorrected
(Corrected) Arch^o-
Ccilendar magnetic
Age (BP) Date
Material
Cultural
Affiliation
DIC-/
“576
nutshell
Benton
539-
(-)
DIC-/
A/12
1
nutshell
Sykes-White
539-
()
Springs/Benton
DIC-/
A/16
nutshell
Eva/Morrcw
539-
()
Mountain
DIC-/
C/Fea
nutshell
Sykes-
539-
120
()
White Springs
DIC-/
D/17
nutshell
Eva/Morrcw
539-
()
Mountain
DIC-/
D/Fea
nutshell
Benton
539-
142
(-)
DIC-/
D/
nutshell
Eva/Morrow
539-
Bu 11
()
Mountain
DIC-/
D/
nutshell
unknown
539-
Bu 19
()
a = C3orrections calculated based an Masca curve (Ralph et al. ) .
•nie tightest date range was the Benton cxxtfxment and eissociated features.
Four dates bracket the Benton occv^jation between 5,796 B.P. and 5,335 B.P.
IVro dates were obtcdned front two separate strata within prepared area Feature
6 vdiich average 5,490 B.P. and cue within ten yecurs of overlapping at the one
sigma level. A stratified Benton pit (Feature 142) was dated at 5,532 B.P.
Feature 120, a Sykes-White Springs prepeured curea front Block C, dated
6,149 B.P. The Eva/Morrow Mounteiin zone in Block A dated 7,303 B.P. A
corresponding Eva/Morrcw Mountain level in Block D, however, dated 6,242 B.P.
This anomalous reading is probably spurious. One Eva-Morrcw Mountain burial
(11) in Block D dated 7,468+85 B.P. vrtiich agrees with the midden date in
Block A.
A nearly Ixirial (12) , which contained no grave goods, had charcoal from
the pit fill vdiich dated to 5,809 B.P. Hiis date is consistent with the
origin of the laurial pits and grave goods in other eussociated burials.
All four radiocarbon dates associated with the Benton oocvpation were
within the 6th millennium B.P. The two Sykes-White Springs dates average
about 500 years older and clvister around 6,000 B.P. TVro of the three
Eva^arrow Mountain dates, averaging 7,385 B.P. , agree with outside dates.
The one Late Archeiic date of 4,594 B.P. is eissociated with Little Bear Creek.
A total of 16 archaeomagnetic sanples from nine fired clay features from
the Walnut site were submitted for dating. The features selected had firm
diagnostic temporal markers associated, and most had a radioceu±)on date.
68
Of the 16 sanples sufcmitted fran nine features, dates \iiiere obtained from
12 saoples from six features (Table 16) . Ihese 12 date ranges viere 30-40
years, well within the standard acceptable range of variation (or di^]er8ion)
for arohaeomagnetic dates. Sanples fran four features had dispersion ranges
too greed: to be acceptable for dating purposes. 'Hie dated features consisted
both of small isolated fired areas (hearths or fired aggregates) and the
hearths within the large pr^>ared areas.
nUB 16
ArciiaeaBagnetic and oorrelatinq radiocarbon dates, 22lt539. _
Archaeo-
magnetic
Feature
Elevation
Dates (±50)
R2Kliocarix2n
95
Sy)oes-White Springs
88.16
5,9b9±115
(10 cm above
feature)
120
ST 23
Sykes-Mhite Springs/
Bentcxi
88.24-88.14
6,250
ST 19
88.21-88.27
6,200
ST 22
88.34
6,150
ST 2
88.2-88.1
6,149±95
ST 16
88.34
6,050
111
Benton
88.57
6,000
73
Sykes-White Springs/
88.34-88.27
5,950
5,706±75
Benton
(10 on above
£
feature)
o
ST 9
88.26-88.20
5,850
ST 7
88.35-88.34
5,800
ST 5
88.38-88.18
5,490±75
ST 2
Benton
88.41-88.30
5,335+75
ST 8
88.40-88.35
4,500
ST 6
88.54-88.41
4,450
ST 1
88.53-88.48
4,300
121
(very large dispersion)
108
(very large dispersion)
121
(very large dispersion)
119
In Feature 120, the largest prepared area with a 2 m (6.6 ft) wide hearth
ccrplex in the center, there were two eireeis of double- layered fired clay
separated by only 50 cm (1.6 ft) (Figure 7). Arch2«cmagnetic dates of the
northeastern hearth area reveeiled that the laist firing of the beisal strata
dated to 6,250 B.P. and the leist firing stratvm directly above it, was
6,200 B.P. The last firing of the lower stratim in the centred, hearth was
6,150 B.P. and the last firing of the stratum above it was 6,050 B.P. Hie
radiocarbon date from this feature was 6,149t95 B.P., suggesting that the
charcoal deposit likely Wcis from this heiurth.
Verticeil zonaticxi of dates in the feature were stratigrajAiically
synchronous. Hie sequence of dates indicates that this hearth ccitplex grew
higher during the 150 years of use. Although there was some overlap in
elevation of Strata 19 and 23, Stratum 19 was resting on top of Stratum 23,
and the deposits were not level.
69
The arcdiaecmagnetic and radiocarbon dates docunnent that Feature 120 was a
Sykes-Hhite Sprin^/Benton prq>ared area and that the hearths were used
sequentially, not oonteqporaneously. This si;:ports an interpretation that
this hearth area was the center of activily for an extended period of tine
(200 years) and that the average oise-life for each hearth (or fired aggregate)
was fqpproxinately 50 years. This feature vats probably the center of a
residentied. area which was used or reused continuously for several
generatixxis. Ihe hearths were nade of silt loam vdiich had been tsrxxi^t onto
the site from adjacent wetlands. The continued, rebuilding of these four
hearths in a snail area (2x4 in/6.6xl3.2 ft) resulted in a buildup of material
^iduch produced a done-shaped, chronologically ordered feature. The dense and
thin feature (15-17 cm/5. 9x6. 7 ft) grew at an average rate of 5.7 on (2.25 in)
per 50 years.
The other prepared area with multiple hearths. Feature 6, was also
archaeomagnetically dated. It was situated ^iproximately 14 m (46.2 ft)
southeast of Feature 120 on the south edge of the surface of the landforro in
Block B. The presence of many Bentcm projectile point/knives suggested that
it was younger than Feature 120. There were five fired aggregates or hearths
in an oval plan in this feature, each consisting of only one layer of fired
clay. Four archaecnagnetic and two radioc2udbon dates rauiging between 5,850
and 4,300 B.P. were obtained from this feature (Table 17) . The date range
docunents that the two prep6u:ed areais were not in use at the 82gne time and
that Feature 6 was initiated 150 years zifter the cessation of firing in the
hearths in Feature 120.
The two northernmost hearths (Strata 9 and 7) were the oldest. Stratim 9
was last fired 5,850 B.P. and Stratum 7 was last fired in 5,800 B.P. The
three other hearths (Strata 8, 6, and 1) , clustered at the south end of the
feature, were fired 1,300 years later. These hearths dated 4,500, 4,450, and
4.300 B.P., respectively. The two radiocarixxi dates of 5,335±75 and
5,490+75 B.P., respectively came from charcoal-rich Strata 2 and 5 beneath the
hearths.
Interpretation of the six dates from this feature is difficult. First, if
adl dates are accurate, this area was in vise for 1,850 years, with a hiatus of
1.300 yecurs between the use of the northern two heaurths and the southern three
hearths. If so, the hearths are not related and actually are not parts of the
same phenomenon. The older northern hearths were lower than the southern
hearths (Table 17) and could have been buried v^ien the use of the later ones
Weis initiated. However, the top of Stratim 7 (5,800 B.P.) and the bottom of
Stratim 8 (4,500 B.P.) are within 1 nm of elevation, perhaps invalidating this
reascxiing.
An alternative interpretation of the long time gap between firings is that
the archaecinagnetic dates for the southern three hearths are incorrect. Two
lines of evidence svpport this interpretaticxi. First, the ccnposition of the
feature strata documented that it was one entity. The close proximity of the
five heaurths (within a 2 m or 6.6 ft area) , as well as the presence of
contiguous underlying strata, indicate that this feature was one
activity-related phenomenon. Cleeurly, a gap of 1,300 years is too long for
related activities to take place. Second, the ra^ocarbon and the
ardiaecmagnetic dates of the northern hearths fall within a 650-340 period of
time (5,900 and 5,250 B.P.) vhich is well within the docimented range of the
Benton horizon (see Chapter II) and indicative of a shorter, more COTitinuous
use of the heeirth conplex. The younger archaeoroagnetic dates for the southern
three hecirths of Feature 6 ere out of the documented chrcnological span of the
Middle Archaic Benton horizon by at least 700 years. Therefore, it appears
that these dates are in error.
Fxon the dating of the hearths of Feature 6, it ^ipears that this edso was
an area of oonoentrated residential activities. Die date ranges indicate that
the two large residential areas Mere not ccntenporaneous. Not all of this
site was investigated, hoMeiver, and more of these features could exist.
Diree additional hearths Mere dated at the Walnut site. TVio Mere isolated
vdthin the midden matrix (Features 95 and 11) , and one (Feature 73) was a
prepared area vdth features. Of these. Features 95 and 73 Mere in Block A on
the southeastern part of the landform. Feature 95 Mas associated with
Sykes-44hite Springs tenporal roarloers and archaeomagnetically dated to
6,300 B.P. Foriy centimeters (15.7 in) above this feature in the Benton
ccnponent, another hearth associated with the prepared area (Feature 73) was
dated archaeomagnetically to 5,950 B.P. A radiocarlxn date of 5,
between these two hearths agrees with the bracketing dates. An additional
rediocarbon date of 5, was obtained &cm midden 10 cm (3.9 in) above
Feature 73. Diese separate dating methods verify each other on unrelated
hearths from two temporally separate cultural ocmponents.
Feature 111 was a hearth in Block C, the same unit 2is Feature 120.
However, Feature 111 was situated 20 cm (7.8 in) above the hio^iest part of
Feature 120. The archaecroagnetic date of the fired clay was 6,000 B.P., which
agrees well with the Benton cultured affiliation of the surrounding midden.
Except for the three young dates from Feature 6, the archaeomagnetic dates
from the five features agree with the radiocarbon dates and the associated
diagnostic cultured markers.
SlMiARY
The Walnut site was a milti-oonponent site on a large natural elevation in
the floodplain of the headwaters of the Tcmbigbee Valley. The site was
occupied for at leeist the last 10,000 years to the present. Four large
excavation units and four stratigraphic trenches produced a good senple of the
material and features left behind by previous occupemts. Following is a brief
sunmary of the site in terms of archaeological ccmponents and intra-site
patterning and activities at the site.
Archaic Stage; Die recovery of twa Daltcxi points indicates that an early
occupation could have occurred (11,000-10,000 B.P.) prior to the documented
Early Archedc oocupaticn. However, these projectile point/knives were
recovered out-of-oontext and could not be associated with a Dalton occupation
at this site.
The Early Archaic W2is the first period of occupation here and was
identified by the presence of Big Sandy and Kirk Comer-Notched hafted bifaoes
in the upper portion of the paleosol soil (Zkme VII) . Die apparent diffuse
nature of the oocuqpations, as well as intensive Middle Archedc utilization of
the site, precluded a clear separation of the Early Archaic oemponent. Die
activities probably included biface manufacture, tool maintenance, and
rejuvenation. The Early Archaic occupations were prcbably made by small
groups on an intermittent seascxial beisis, and the site was used for
specialized extraction during the time period from 10,000-8,500 B.P.
Distinctive stenmed comer-notched hafted bifaoes provite enough evidence
to define the cultuiral oemponent which follows tentatively as "Cypress Creek."
Diere is the possibility, however, that these hafted bifaoes could belong with
the Eva/Morrow Mountain occupation. Stratigraphically, however, th^ appear
to be Late Early Archaic or Early Middle Archsdc from ca. 8,500-8,000 B.P.
71
The Middle Archedc Eva/Horrow Mountain occupation of the site was much
moce substantial than previous occupaticam. The tern Bva/Morrow Mountain is
used here to reflect consistent oo-oocurrenoe of these marioeirs and blending of
the attributes of both into one projectile point/knife fom.
Pr^sared areas, representing foc£d. points of activity, first appear in the
Eva/Moonw Mountain occupation of this site. Althou^ th^ are less well
defined and perhz^ somewhat smaller in sire than those of the later
S;^oe8-4tu.te firings and Benton periods, they nonetheless indicate that these
activities were initiated. Specific tasks at these areas included the
procurement and reduction of local Camden chert cobbles into finished tools
involving heating the chert, cobble reduction via hard haniner and soft haniner
percussion, and producing flate blanks from the cobble cores. These were then
heated and further reduced by soft hammer percussion.
One of the most distinguishing aspects of the Bva/Morrow Mountain
smsenblage is the ubiquity of flake tools such 2ts hafted end scrapers and side
scTi^sers. The variety of flake tools included severed kinds of scrapers on
flakes. Flake blanks were generally expanding to amorphous in form; hewever,
seme were blade-lilce. A variety of other chipped and ground stone tools was
also found. The large amounts of fired aggregates, burned sandstone,
charcoal, fire-cracked chert, and a diversified tool kit suggest a svibstantial
occupation.
The Eva/Morrew Mountain tool assemblage suggests that a variety of
extractive and maintenance tasks were carried out cxi the site, probably in
association with the prepared areas. Such t2isks seem to have involved
hunting, fishing, and turtling as well as processing the material frem these
outings. Evidence for tool manufacture and use is present with a full
ocnplement of implements present.
Another inportant aspect of the Middle Archaic (Eva/Morrow Mountain? or
the later Sykes-White Springs/Benton) occupations was the presence of two
organized burial areas located in sq>arate sections of the site. The layout
and arrangement of the oemeteries suggest a oenmunity plan ^tihich involves the
segregation of seculeu: and ritualistic activities. The 7,468±85 B.P. date
fron Burial 11 and the 5,809±245 B.P. date fron cremation (Burial 19) provide
inconclusive evidence for developing chronological and cultured associations
in the oemeteries.
Prepared cuneeis as focal points of activity continued during the succeeding
S^ces-White Springs occupation. The large pirepared area, Feature 120,
contained four separately fired hearths suggesting repeated usage for 200
yeeurs. As in the Eva/Morrew Mountain zone, there is an indication of nultiple
tasks Ccurried out on the site; however, the incidence of eeurly stage bifaoe
manufacture is sonev^t diminished. T^ relatively large amount of bone
recovered in Block C, apparently in association with the large prepared area,
suggests that the occupation was semipermanent during the summer throu^ fall
months. This is speculative, and more sensitive seasonal indicators, edong
with better contexts, are needed to make such a determination.
TVro inhumations were recovered in Block C in the inroad late vicinity of the
prepared area. The level of origin of the pits for the interments, however,
is ixjt clear. It appears that semipermanent or permanent occupation (s)
occurred during Sykes-White Springs times. The presence of large prepared
aureas and a diversified tool kit suggest a baise camp utilization for the
Wednut site locale during this time period. It appeaurs that between 6,500 and
6,000 B.P. the Sykes-White Springs culture progressed smoothly into the
Benton.
llie Benton ccn(X)nent(s) at this site provided the best infoonation
oonoeming site patteniing, subsistenoe, technology, and overall cultural
plaoement. Diis occupation dated from 6,000-5,300 B.P. Hie data si^pport the
contention that the prepared area in Blo^ B was a focal point of the Benton
occupation. Chemical and E^siceil analysis support the contention that
introduced sediments were used to construct all prepared areas. Hie
distribution of tools on and around the prepared area indicate that activities
such as tool manufacture and rejuvenatic^ took place. Processing of animals
and plants also appeeurs to have occurred. Althou^ faunal remains were
virtually absent, this apparently is more a function of preservation than
cultural practices. Hie presence of charred hickory nutshell and wood
charcoal, along with the fired aggregates, suggests that burning was ooninon,
probably related to food processing. Hie ubiquity of the B^ton projectile
point/knife form, possibly multipurpose tools, and the numerous fracpents
manufactured from Port Payne cherl: indicate intensive rejuvenation activity.
Artifactual remains of the Middle Archaic coupled with hi^ nmbers of
pits, organized buried eueas, emd hearths indicate that the Walnut site weis a
multiple activity locus during Middle Arcdiaic times. It suggests strongly
that the site was used as a permanent or semipermanent base canp during this
time period. Hie primary season of habitation may have been during the sunmer
and fall. A flooc^lain site is more tenable during Icw^^ter periods.
Further, the abundance of hickory nutshells recovered and the types of aquatic
resources taken suggests fedl occupance.
Hie Late Arch2dc occupations of the Walnut site axe not well understood
from the recovered materials. Hiere is a distinct possibility that 2ui
occasional Ledbetter-Pickwick occupation occurred from 5,000-3,500 B.P.;
hcMever, no inferences oonoeming activities or patterning can be made.
Little may be inferred from this occupation except that there appears to have
been a najor shift in biface technology involving the use of locad heated
Camden chert in the manufacture of projectile point/knives and other tools.
It appears that the site was less intensively occupied than in preceding
Archaic occupations. It nay have been a special activity locale or a
semipermanent cenp during this time period.
Woodland and Mississippian Stages; Hie deposits of the post-Archaic
prehistoric occupations at Site 22It539 had been disturbed so that verticad
separation was not possible. Horizontal separation, with sane patterning, was
evident from the location of tenporal ceramic types. IVro possible Middle
Woodland ccnponents were detected: one with limestone-tenpered vessels on the
northwest section of the site (Block C) and one without limesbaie-tenpered
ceramics on the south porticai of the site (BlocJcs A and B) . Tuo possible Late
Woodland/Mississippian ccnponents were identified. Hie oorponent containing
shell-and-grog-tenpered ceramics was located on the north side of the site
(Block C) , and the ccnponent containing primarily grog-tenpered ceramics was
located on the southern end of the site (Blocks A and B) . Hie Gulf
Formational ccnponent appears to have utilized the entire site area relatively
uniformly.
Hie ceramic types recovered at 22It539 indicate that most of the
post-Archeiic ccnponents previously encountered in the Ufper Tombigbee Valley
were present here. Hiese include the Middle Miller I, Pharr subphase
(A.D. 100-400) and possibly Gainesville subj^ses (A.D. 900-1,100); the Late
Miller III, Catfish Bend subphase (A.D. 900-1,100); and the Early
Mississippian (A.D. 1,200-1,300) (Jenkins ).
EJflCAVanONS at the nzx site (22It590)
The Ilex site vias in northern Itawanisa Counl^ along Mackey's Cre^, a
tributary to the headwaters of the Tcnbi^bee River (Figure 1) . The site was
on the first terrace adjacent to the j\incture of the southern valley wall an
the floo^lain of the Macke's Creek valley. The site landfom was a
relatively flat terrace vhich was tangent to the hi^ Pleistocene ridge on the
south portion of the site prior to separation by a small stream. The site was
bordered by Mackey' s Creek on the north and west and a sweep on the east
(Figuire 12) . The site W2is approximately 100x60 m (330x198 ft) in size and was
vp to 2-3 m (6. 6-9. 9 ft) above the lowlying floodplain. The Pleistocene ridge
which forms the southern boundary of the site exhibited slumping, sheet
erosicxi, and mass novements of sediments.
FTKIi) METHODS
The Ilex site was initially recx>rded by Blakeman (:19) , vho cxjnsidered
the site to be signific:ant based on the presenc^e of Tr2uisiticxi^
ArthaicAkx)dland, Wocxiland and Mississippian cxxnponents identified from a
surface oollecrtion. The site was reccnmended for testing, and this was
performed in (Sense ). Testing included two 2x2 m (6. 6x6. 6 ft) units
and a conplete site profile vMch confirmed Blalcanan's results and, in
addition, identified in situ Archaic cxnponents (Bense :430).
The excavation at this site Wcis precseded by more extensive preliminary
investigations than had been enplcjyed previously to aid in the placement of
blcxh excavation iinits. These efforts iiKiltaded: 1) three stratigraphic
backhoe trenches, 2) a series of 1x1 m (3. 3x3. 3 ft) test units along the
stratigraphic trenches to correlate cultured and natural strata, 3) several
1x2 m (3. 3x6. 6 ft) test units in areas between the stratigraphic trenches, and
4) visual cures on an 8 m (26.4 ft) grid pattern.
The stratigraphic trenches exposed representative profiles of the site
without disturbing the majority of the central site area. The initial
stratigraphic trenches (1 and 2) were placed along the e^qused eastern and
northwestern borders of the site (Figure 12) . Trench 3 tremsecrbed the
southern part of the site. Trenches 4 and 5 were excavated later to
investigate the central and scuthwestem area of the site.
The test units ccnsisted of 12 1x1 m (3.3x3. 3 ft) units placed adjacent to
stratigrajhic trenches. Ei^t additional 1x2 m (3. 3x6.6 ft) test units
throughout the site area were also excavated. Excavation was in arbitrary
10 an (3.9 in) levels.
Chemical coring techniques were not enplcjyed at 22It590, because the
results at the Walnut and Poplar sites had been inconclusive. The
stratigrajdiic trenches and test units, sv^plemented by an 8 m (26.4 in) visucd
coring grid, provided much more useful informaticjn.
The strategy for subsequent excavations stressed problem-oriented
research. The pragmatic placement of excavation units to address specific
research problems was facilitated by these prelimirary investigations at the
site. The final excavation strategy stress^ recxjvery of the early cxnponents
at the Ilex site. However, evidence for all cultur^Q. cxnponents enccuntered
was scrutinized in an atbenpt to determine the relative integrity of recxvered
materials and their interpretive significanca.
The excavaticjns at Ilex included five stratigrajiiic trenches, nine 4x4 m
(13.2x13.2 ft) blcxks; cae 4x8 m (13.2x26.4 ft) and one 2x2 m (6. 6x6. 6 ft)
unit; 12 1x1 m (3. 3x3. 3 ft) test units and eight 1x2 ro (3. 3x6. 6 ft) test
74
ILEX SITE
22IT590
Itawamba County, Mississippi
^AppranniMt BoundW.. .
SuWozar Cut
a^KB
o’rr's*
TESTjfrfaa
TJIX qjTESTPITU
BLOCK H Ly BLOCK A-W
TESTPirna
TEST PIT 150
Contour Intorval; 25cm
Noik OaunmkOL
Figure 12 TOpogreqiiic map and excavation plan, 22It590.
8 8
vmits. As can be seen in Figure 12, the najor area of excavaticxi was in the
ixnrthrcentral part of the site. Here, seven major excavation blocks were
separated by a 1 m (3.3 ft) balk used to aid in stratigraphic control. HMs
area is designated as the "Master Block" and is equal to approodmately an 8x16
m (26.4x52.8 ft) excavation unit. In most units, the ipper mixed midden v«as
renoved at the site and tossed. However, in Blocks A-Z, B, and J, the entire
sequence was screened and all material recovered. Block D wais a 10x10 m
(33x33 ft) unit, however, in this unit, the ipper mixed midden was stripped
off, and the features were mapped and excavated.
CULTURAL RaiAINS
A total of 5,995 ceramics were recovered frcm the Ilex site. Based on the
temper and typology of the recovered specimens, the site was occupied
throuc^xxit the Gulf Fomnational to Mississippian stages. As at the other
midden mounds investigated, attarpts to provide a chronologicad ordering of
ceramic-bearing occipations were largely unsuccessful. Hie ceramics were
largely confined (97-98%) to the vpper 40 cm (1.3 ft) of the profile. In most
units, the shell-tempered specimens were confined to the upper 20 on (7.8 in) ,
and tte grog-tempered were confined to the upper 30 on (11.7 in) . Hcwever,
the older limestone, sand, and fiber-tempered sherds were thorou^y mixed
throuc^iout the ceramic-be^ing zone. This is likely due to the slew rate of
vertical site accretion, intense use of the site by ceramic-using occupants,
and recent extensive looting.
All known temper types are represented in the sample (Table 17) .
Scuid-tempered sherds deminate the sample (66.2%) follcwed by fiber-tempered
(20.4%) . Late Woodland (grog- and bone-tenpered) and Mississippian (shell-
tenpered) ceramics are very low in frequKicy at this site and amount to only
8.3% of the ceramic inventory. The surprisingly high amount of fiber-tenpered
ceramics (1,225; 20.4%) r^aresents the hi^iest proportion of such ceramics of
ai^ cissenblage reported in the Upper Tembi^oee Valley. Perhaps the location
of this site nearest the divide (16 km or 10 mi) with the Middle Tennessee
Valley, where Wheeler fiber-tempered ceramics developed, is significaiit. The
Late Gulf Formational ceramic assemblages (Alexander series) were
approximately hadf as frequent as the Wheeler (621 or 10.4%) . The
limestone-tempered sherds were also more frequent (5%) at this northerrmost
project site and itay be another indicator of interaction with the Middle
Tennessee Valley.
•dole 17
Tenper
Shell
117
1.9
Shell-Grog
26
0.4
Grog
329
5.5
Bone
29
0.5
Limestone
300
5.0
Sand
3,969
66.2
Fiber
1,225
20.4
Total
5,995
A total of 5,049 chipped stone inplenents were recovered frcm this site
(Table 18) . While 46.7% of the chipped stone was so fragmented that the tool
type was unidentifiable, projectile point/knives (ccrplete and fragments)
76
Gcnprised 21.1% and were nore frequent than at any other site investigated.
Other tool types included bifaoes (4.7%), prefonns (9.9%), oores (4.7%),
scrapers (6.2%), drills, perforators, and other small tools (3.8%).
18
Category
Frequency
Projectile Point/Knives
Beaver Lake
1
Benton
58
Big Sandy
6
Cotaco Creek
10
Cypress Creek
11
Dalton
5
Eva
11
Gary
9
Greenbriar
21
Hardaway
1
Kirk Comer-Notched
31
Late Woodland/Mississippi Triangular
22
Ledbetter Pickwick
16
Little Bear Creek/Flint Creek
86
Mclntire
11
Morrow Mountain
9
Residual Stemmed
94
Sykes-White Springs
20
Tcmbi^see Stenmed
4
Vauc^
2
Distal Fragment
276
Medial Fragment
170
Proximal Fra^nent
168
Unfinished Snell Triangular
3
Residual Trianguleu:
9
Plevna
4
Beachum
9
Subtotal
1,067
21.1
Bifaces
236
4.7
Cores
237
4.7
Preforms
458
9.9
Scrapers
315
6.2
Drills, Perforators, etc.
191
3.8
Other Unifaoe and Bifaoe Tools
2,545
50.4
Total
5,049
Utilized flaJces were a frequent e^qsedient tool type used at the Ilex site.
A total of 5,557 were recovered. A total of 85,091 non-utilized flakes were
recovered. Heated and unheated Camden chert dcndnated the debitage (92%) .
Within the ccnibined Camden size grades, heated materud predominates (77%) .
The dcninanoe of Camden chert during edl periods of oocipation at 22It590
conforms to a pattern recognized throughout the U^per Tcnbic^see Valley. Fort
Payne chert was the most frequent minority type ocnprising 4% of the debitage
with other types constituting less than 4% of the debitage.
nie size-^eide analysis of the d^itage indicated a pattern sunilar to the
Walnut and Poplar sites in the c^rzelaticn between analler size and hi^ier
frequency (Tidble 19) . The one-inch and larger flzdces were lowest in frequency
(0.4%) , the half-inch flakes were more abundant (15.6%) , v^le the smallest
size recorded, quarter inch, dominated the debitage (84%) . Ihis reflects both
the IcKir nxxnber of large flakes removed from Gobbles in stone tool manufacture
and the hic^ amount of manufacturing activity oonducted there. A ratio of
1:41:219 of flake size grades was characteristic of this assemblage. The
tool-to-debitage ratio was 1:17.
of defai
22It590.
Flake Size
1.0 inch
0.5 inch
0.25 inch
Ototal
325
13,289
71,477
85.091
78
Introduced rock was abundant at this site (145.6 kg) . It consisted
primaadly of femiginous sandstone and p^Dbles.
Faunal remains from 22It590 consisted of a few very small fragnents of
bone recovered and sorted during the flotation analysis. Ihe poor quality and
small nunber of firagments deviated identification. One dog burial was
encountered in a test unit during the testing (Bense ) .
LiMted nunlsers (30) of historic/modem artifacte were recovered from
22ItS90. Most of these items reflect recent land use practices, such as
fanning, tree planting, and construction. The small heterogeneous samples of
ceramics, glass, metal, and miscellaneous debris are probably less than 30
years old.
FEKHJRES
Sixty-seven features were recorded at the Ilex site and included:
7 rock cliasters
1 bone cluster
1 complex cluster
2 fired aggregates
51 pits
5 prepared areas
Pits were the most numerous feature and the vast majority contained
the same kinds of cultural materials found in the surrounding midden. Pits
generally were chcuracterized by ti» dark color of the internal fill vArLch
contrasted with the lifter color of adjoinii>g sediments. Oxidized sediments
and carbonized organics, ccrmonly found in pit fills provided evidence of
burning.
The cultural affiliation of only six of the pits was determinable.
Usually the pits contained no temporally sensitive material or had mixed
tenporal marJcers, due to digging activity. Of the six pits with either
temporal meukers or good stratigraphic eissociation, two were associated with
the Eeurly Archaic, two with the Middle Archcdc, and two were prt*ably
associated with the Late Archaic period. Pits were usually only identifiable
when they contrasted in color with the lighter and deeper strata. Therefore,
it is inferred that the v^^r portions of most pits were included in midden,
rather than with feature excavaticm.
The two fired aggregates at this site were relatively snail concentrations
of burned earth and appeeured to be part of larger prepared area features which
either decomposed or were destroyed during siibsequent occupation (s) . One
hearth was located in the Master Block area, one each in Block A-Z cuid G.
Ifcwever, they could be remnants of large fire hearths. It was difficult to
associate specific material with the fired aggregates.
All five prepared areeis were encountered in the Meister Block units A-W, X,
Y, and Z. The condition of these features was poor, however, causing both
definition and cultural affiliation problems. A greater diversity of
artifacts was recovered from the vicinity of the fired aggregate features. It
c^spears that both fired aggregates and prepared areas were peurt of larger,
pr^>ared living areas. The vertical and horizontal proximity of Features 22
and 23, (Figure 13) s\apports the proposition that these were remnants of a
single prepared area. Only two prepared areeis were well defined and could be
eissociated with a cultural component (Benton) . Even this association was
equivoccd. The size of these two features was similar: 1.95x1.66 m
(6. 4-5. 4 ft) long, 2. 0x1 .5 m (6. 6-4. 9 ft) wide, and 27 and 28 am (10.6 and
79
11.0 in) thick. These pr^iared areas had neoltiple hearths surrounded 1:^ a
nosaic of yellow and orange d^)06its (Figure 13) . The three other prq>ared
areas had been too disturbed to mec^sure or 2issociate with an occupation of the
site.
The seven rock features identified were generally well-defined clusters of
fist-sized, ferruginous fragments. Several rocks in each feature apgear to
have been thermally altered. These rode clusters were likely associated with
fire hearths, rock ovens, use areas of convenience tools or general disposal.
Cultural materials associated with the rock clxisters ccrnnonly included burned
debris, siKdi as fire-crac]ced chert, hematite, and fired clay. This implied in
situ burning in association with the rock clusters and supported the thesis
that these features were connected with fire-related activities, such as
heeu±hs or rock ovens. The absence of strong oxidation of soil matrices and
a^ in association with the clusters suggested rapid weathering and
deccnpositicmi, post-depositional disturbances, and/or short-term use of these
features.
One oonplex cluster \diich contained a variety of cultured, ddmris,
including burned sandstone cobbles, lithic inplements ard d^itage, and
pottery wsts defined. Evidence of burning and darker internal matrix color
were interpreted as indications of a snail pit, althou^ boundaries could not
be distinguished. The Wheeler sherds associated with this feature appeared to
have been from a single vessel. The feature was probably a hearth contained
in a small pit.
Three features were eissociated with recent relic digging. These were
recognized by stratigrai^c interruptions and the presence of historic/ recent
artifacts. The site surface esdiibited a series of such potholes.
Human skeletal renains from the Ilex site were limited to a single
occurrence in Block A-X and consisted of several badly decomposed bone
fra^nents. The best preserved of these fragments appeared to be a long bone
segment (f«kiur?) that was oriented in an upright position within the deposit.
A very faint color change may have marked a pit outline, although the exact
boundeuries and level of origin of this feature could not be determined. This
probably was a secondary burial in a small pit. There were no obvious
autifact associations or additional interments in the vicinity of the hunnan
bone.
STRATIGRAPHY
The site soils were loams deposited as alluvium from Mackey's Creek
through overbank deposition eind colluvium from erosiem of the adjoining
uplands. The site landform appears to have initiated as a terrace frai^nent
outlier in the floodplain, similar to both 22It539 and 22It576. Sands in the
site soils were coarser than the sands found at the Walnut and Poplar sites.
Nine strata were identified at tte site, but th^ were not uniformly
distributed throu^icut the site (Figure 14) . The upper meter (3.3 ft) (Strata
I-III) was darkly stained from organic matter and thinned appreciably to the
south. The cultural midden was under ledn by massive, undifferentiated sand
horizons of varying thicknesses. A subsurface ridge of highly weathered
sediments, oriented in a north-south direction, Weis underladd the western
margins of the site area in Trenches 4 and 5 and in Block C excavation unit.
This deposit ^jpeared to be a former terrace, possibly of Pleistocene age.
Hi^)er silt and clay fractions and gley colors characterized this terrace
remnant. Overlying this deposit was an indurated B horizon that wcis
characterized by strong reticulate mottling.
Lamellae were ocmnDn occurrences cm the north end of the terr2U3e below
ca. 50 an (1.6 ft). Ihese were generally horizcmtal, dark-colored bands of
vauriable thickness and c^peared to oorrespcnd to arecis of the site with thick
cultural midden deposits. Itiis suggested that the midden was the source of at
least some of the clays and organics present in the lamellae.
CHROMCMETRIC DftTlNG
Thirteen radicxarbon dates were captained frcm 22It590 (Table 21) . The
samples, except one, were all from midden matrix because of direct
associations between tenporally sensitive eurtifacts and charcoal in feature
context were lacking.
1MBBUB 21
Radiooartxai dates, 22It590. _
Uncxprrected
Lab #/ (Corrected) Archaeo-
Field
Number
Blade/
Level
C-14
(BP)
Calendar
Aqe (BP)
magnetic
Date
Material
Cultural
Affiliation
DIC-/
590-
5
±70
()
nutshell/
charcoal
Benton
DIC-/
590-
6
+75
(-)
nutshell/
charcoal
Benton
DIC-/
590-
9
+75
()
nutshell/
charcoal
Middle
Archaic
DIC-/
590-
10
+70
(-)
nutshell/
charcoal
Early
Archaic
DIC-/
590-
10
±100
()
nutshell/
cleurcoal
Early
Archaic
DIC-/
590-
10
±90
()
nutshell/
charcoal
Eeurly
Archcdc
DIC-/
590-
11
()
nutshell/
charcoal
Early
Archedc
DIC-/
590-
Feature
34
+85
()
nutshell/
chcurooal
Middle
Archaic
DIC-/
590-
9
+75
()
nutshell/
charcoal
Middle
Archaic
DIC-/
590-
9
(-)
nutshell/
charcoal
Early
Archaic
DIC-/
590-
10
(-)
nutshiell/
cheurooal
Early
Archaic
83
naut 21
HiUcirwTticp datcB, 22It590.
Lab #/
Field
Ntnber
Block/
Level
C-14
(BP)
Ifticorrected
(Corrected)
Calendar
Age (BP)
Archaeo-
magnetic
Date
Mater ied
Cultvural
Affiliation
IC-/
10
nutshell/
Early
590- () charcxial Archedc
DIC-/ Test Pit modem modem tree stmp
590- 18 _ _
a = Corrections calculated baised on MASCA curve (Ralph et al. ) .
Six saii{>les were from levels associated with the Early Archaic
occc^tions. Unfortunately, all of these Early Archedc dates are considered
to be too recent. The charcxsal appears to have migrated down or to liave Iseen
mixed with cluurooal from above.
TVro radiocarbon dates eissociated with the Benton ccnponent (5,777 B.P. and
5,227 B.P.) are consistent with other Benton dates frcm the Poplar and Walnut
sites. One other date was obtained from the Benton zone, but the 3,828 B.P.
date is too recent to accept.
One date of 6,200i55 B.P. was from a Middle Archaic pit originating just
below the Benton zone. The date is ocxisistent with the Sykes-White Springs
Middle Archaic carponents dated at the other sites. The modem date obtained
was not unanticipated despite the considerable depth (ca. 3 m or 9.9 ft) of
this buried tree stump. The sample vas extracted from Test Pit 18 on the
extreme northwestern edge of the site in an attanpt to date the Pleistocene
ridge remnant in Stratigr^Aiic Trench 2.
StttftPY
The Ilex site composed of fluvial and colluvial deposits was located on
the first terrace of Mackey's Creek valley. Hie terrace was surrounded by
wetlands in the floodplain of Mackey's Creek approximately 6 km (3.7 nd)
ipstream from the confluence with Big Brown Creek which forms the Tomibi^iee
River. The site apparently had an aval outline prior to being truncated on
the north and west.
Recent disturbance by heavy machinery had destroyed a considerable portion
of the site prior to testing in and excavation in . At this time, it
was ca. 60 m (196 ft) frcm Mac)cey's Creek. The site may have originally
extended to the creek.
Investigations conclucted at this site docunented that it weis occupied
continually for the last 10,000 years. Despite removal of portions of the
site, a good sample was recovered from the past occupations.
Archaic Stare; The oldest cultural materiad was contained in the Late
Pleistocene /Early Holocene terrace remnant cxi the western portion of the site
(Blocks G and H) . Most of the Early Archaic and Middle Archedc artifacts ware
ccxitained within Early Holocene fluvial sands deposited during intermittent
periods of stream aggradation.
Early Archaic occipations at the Ilex site were identified by projectile
point/knives that included Beaver Late, Big Sandy, Dedton, Greenbriar, and
Kirk stylistic markers. Cypress Creek artifacts also occurred in early
stratigraqphic contexts, but appeared to r^resent either very late Early
Archaic occupations or a transition into the Middle Archaic. Stratigraphic
and radiocarbon data from 22lt539 lUcewise suggested that Caress Creek point
styles date soroevihat more recently in time than those irtcluded in the Early
Archaic ccnplex.
Archaic projectile point/knives were recovered with greatest frequency
fron the Master Block excavation area in Levels 9-11 bet>)een 90 and 110 an
(2. 9-3. 6 ft) below the surface. Kirk, Greenhricu:, and Daltm types vgere most
frequent in the Ecurly Archedc types. General bleii^ morphology, retouching
patterns, beveling, and serration indicated that they were multiputpose tools.
Kirks, in particular, reveciled consistent eiltemate edge beveling, indicative
of resharpening vhile in the haft implying use cis a cutting iitplenent rather
than projectile.
The stylistic characteristics of the projectile point/knives ranged fron
long, lanceolate forms, such as Beaver Late and Greenbriar, to small,
comer-notched Kirk forms. The morphological diversity present in this sample
implied a significant temporal span (ca. 10,000-8,500 B.P.) for Eeurly Archaic
occipations.
Early Archedc assemblages included a wide range of iitplement forms, such
as scTc^^ers and drills, and bifacial reduction products and by-products.
Scrapers were dominated by small, unifacied vcurieties (ca. 96% of scrc^iers
recovered from 100-110 cm (3. 3-3. 6 ft) below the surface in Master Block) ,
with well-defined "thumbnail" or "keeled-end" styles being ccrinon. Other
tools, although not stylistically distinctive, included gravers, perforators,
reamers, adzes, choppers, unifacial and bifacial ]cnives, chisels, wedges, and
a small number of ground stone items dominated by ground lianatite and limcxiite
and ground flakes. A single ground atlatl weight of greenstone was associated
with Early Archaic levels in the Master Block.
Debitage conformed to the pattern of all site catponents and consisted of
heated and unheated Camden chert. Fort Payne chert vas well represented and
dominated minority raw material types in Early Archaic samples.
Ecurly Archedc features consisted exclusively of two pits probably used for
refuse disposal or storage. Site use during the Early Archedc appears to liave
consisted of relatively short-term canps, but longer term b>ase canp
ocevpations possibly occurred. The diversity of implements and eu:tif6K:ts
represented suggested beise camp settlement (s) , although the relatively low
number of items in any category combined with the proposed span of Early
Archadc occv;pations nay have accounted for this occurrence.
Projectile point/Jmife styles aissociated with Middle Archaic occipations
included Eva, Morrow Mountedn, Sykes-White Springs, Vaughn, and Beachun. Eva,
Sykes-White Springs, and Morrew Mountain specimens were most frequent. In
general, the Middle Archaic specimens did not exhibit the patterns of use-wear
or the degree of internal diversity noted in the Early Arcdiaic. The
typological variation present in the Middle Archaic, in conjunction with the
stratigraphic contexts of these materials, implied an age-reu^ge from
ca. 8,500-6,000 B.P. (ca. 6,500-4,000 B.C.). Multiple projectile point/knife
types suggested intermittent occipation of the site throughout this period.
Apeirt from the projectile point/)cnives, there were few notable differences
of the material assemblage from the preceding Eeurly Archaic. Material
densities appeared to increase sli^tly, although diversity in inpleroent
categories did not change much. Scraper forms generally laclced the
stylization noted in the Early Archadc assemblage and drills-perforators
increased in number during the Middle Archaic. Another difference was the
increase in both number and diversity of the ground stone categories. The
85
most obvious diffexenoe in Middle and Early Archaic occi^tions was the
euxumilaticn of organicedly stained cultural midden vrith diarooal flakes and
Increased organic residue, nils was consistent with 22It539 and 22It576 and
was perhaps a direct result of changes in subsistence/ settlement patterns.
Settlements of longer duration, and possibly of greater intensity, most likely
produced these cultural middens. IMs pattern may be indicative of the
establislment of base canps, rather than temporary camps, at this time
thrcu^hout the tTIV.
The Benton assemblage of the Middle Archaic components was characterized
by the diversity of forms and uses represented in the projectile point/knives.
Hie Benton projectile point/knife dominated the chipped stone assemblage and
modification of these points into secondary implement forms (e.g., drills ai>d
scrapers) Weis extensive. A wide range of uses, therefore, was represented
within the variant categories of Benton projectile point/knives (e.g. ,
projectiles, knives, scrapers, multipurpose implenients) .
Significant differences also occurred in chipped stone cind lithic raw
material procurement. The focus of these changes was an increased use of Fort
Payne chert in the manufacture of bifacial implements. Large bifaoe blades of
Fort Payne chert, not locadly available, ocmncnly were the starting points in
the manufacture of Benton projectile point/knives. This technological shift
is reflected in preform and bifaoe blade categories and by changes in bifacial
reduction strategies and an incre2ise in the occurrence of Fort Payne chezi:.
Changes in other aspects of this assemblage consisted of an increase in
numbers of items left b^ind in the midden and appeared to coiTe^xznd to
greater occipational activity and longer settlement of the site. The two
dates, averaging 5,794 B.P., obtained in the Benton ocmponent at 22It590 are
consistent with the dates and evidence of Benton ocmpcments obtained from
other sites investigated.
Mclntire and Ledbetter/Pickwick projectile point/knives recovered from the
site may represent a small Late Archaic ccmponent. Limited data and the
extent of mixing within overlying ceraraic-be2u:ing ocmpOTjents reduce the
reliability of inferences drawn.
Post-Archaic Stage; The post-Archaic deposits at the Ilex site were narked by
the ahrvpt ^ipearance of oeramics. The n\mtber of ceramics recovered increcised
in the i^per levels of the deposit, with the leugest concentration occurring
in the top 10 an (3.9 an) .
Isolation of cultural strata within the ceramic-bearing deposits at this
site was not possible. Ceramics spanning the period from middle Gulf
Focnational period (ca. 900-500 B.C.) to Mississippian stage (ca. A.D.
1,000-1,550) oocvpations were present at the Ilex site. The Gulf Fomationcd
and Middle Woodland cerandcs were far more frequent than later cultures and
may represent the major post-Archaic occupations of the Ilex site.
The intensify and/or duration of these late ccmponents, as indicated by
the high density of material, appeared to be only slightly reduced from the
Benton ooctpations. This evidence, ocmbined with the strong midden
development associated with oeramic-bearing occupatiais, suggests the
persistence of the pattern of intermittent, semdpermanent occipations.
EXCAVATIONS AT THE HICKORY SITE (22It621)
The Hickory site was located approximately 14 km (8.7 mi) north of Fulton,
Ms in the Tcrbigbee River (Figure 1) floo(%>lain 300 m (984 ft) from the valley
wall. The site was a lew ovoid knoll, 25 m (83 ft) lay 38 m (124 ft) , rising
86
60 an (2.0 ft) edx>ve the surrounding floodplain (Figure 15) . 'Ihe i;f>strearo
edge of this elevation had a steep and higher profile than the dcwnstream end.
The Hidcory site v)as located in an especially low and wet porticn of the
floodplain. Ihe area had many seep springs, flawing water, and swarnps. Hiere
were several sniall tributaries and former tributaries in the near vicinity
surrounding the site. Ihis site was an "island" of sandy soil in the swannpy
floo^lain. The landform appeared to have originated as a point or paradlel
bar deposit in the floodplain vhich grew in size through alluvial deposition.
Prior to testing, the site was covered with an oak-hickory hardlwood forest.
FIEID METHCDS
The Hickory site was discovered during the testing project with the
assistance of a local collector. The site had undergone extensive looting and
had not been recorded by previous waterway surv^s. Its proximity to similar
sites, especially 22It539, made it practicable for the site to be tested in
as part of Phase I of this project (Bense ) . Testing consisted of
two 4x4 m (13.2x13.2 ft) units placed in the center of the mound vrfiere there
was the least disturbance (Figure 15) . The high water table and rainy
oonditicns forced the abandonment of one test unit (B) after 60 on (23.4 in)
and trenching around IMit A to reach sterile soil in one half of the unit.
Althou^ the upper meter (3.3 ft) of bleu::k organic midden was disturbed,
the yellcw-brcwn zone beneath the midden appeared to be undisturbed and
cont£dned Eeurly to Middle Archaic d^x^its. Excavaticms were conducted
between November of and January of to investigate the Early Archaic
ocnpanent. Excavation focused on the highest, central portion of the site
v4iere tlie deeper deposits had occurred on oth^ similar sites. The central
sediments were least disturbed by cutting and filling qoisodes seen in Test
unit A. Between the testing and data recovery phases, this site was
accidentally included in the clear-cutting of the flocxiplain for the waterway.
Hcwever, since the intact deposits were probably preserved beneath the meter
of midden, mitigation plans were not affected.
Significant technic2d. and logistic prc^lems were encountered in
pr^)eu:ation of this site for excavation because of the lew and swanpy
floo^lain location. Further ooiplications occurred because clear-cutting had
cbliterated grid landmarks. Site preparatiCTi was initiated by iremoving the
mixed upper midden with a bacldioe in the central area of the site. This
allowed quick access to the cemented manganese stratum vhich was at the base
of the organically stained midden vhich sealed the Archaic d^)osits.
Approximately 75-85 cm (2. 5-2. 6 ft) was removed in an area approxiroately
18x25 m (59x82 ft) in size.
An area of about 250 sq m (820 sq ft) was then selected at the most
central porticxi of the site for excavation. Around this central area a trench
was excavated with the backhoe to provide a dry pedestal to be excavated
(Figure 15). The trench was 1.5-2 m (4. 9-6.6 ft) wide, 60 m (196.8 ft) in
total extent and 2-3.5 m (2.6-11.5 ft) deep and enclosed a 22.5x11 m (73.8x36
ft) area. The depth of the trench and the unstable san^ sediments made it
necessary to share, the entire length of the trench had to be shored and
cross-braced with heavy lunber. Punps were used to remove water from the
trench, and the pedestaled block remained dry.
Thiee 4x4 m (13.2x13.2 ft) blocks (Blocks C, D, and E) were placed in the
pedestal (Figure 15) , and all were excavated to sparse cultural material. One
2x2 m (6. 6x6. 6 ft) unit was excavated in each block to sterile soils.
87
HICKORY SITE
22IT621
Itawamba County, Mississippi
Figure 15 Topograj^c map and excavation plan, 22lt621.
To address the possibility of post-depositional artifact novement,
artifacts in the lower strata discovered were plotted and photographed ui situ
to record the orientation. A soil monolith or oolinii that eadiibited the total
stratigraphic sequence of the lOMer strata was removed from the west wall of
Block D.
Hie control block was eitpanded to 1x1 m (3. 3x3. 3 ft) and 2lL1 soil was
ocnpletely processed by flotation in an attenpt to recover more botaniced
remains, from the Early Archaic deposits which previously had been scarce in
all the sites investigated. Hiis and all other soil was processed in the City
of Fulton because of contaminated water at the site.
COLTORAL REMAINS
A total of 5,997 sherds were recovered from the Hickory site. All but 81
of these sherds were recovered during the testing phese in Blocks A and B.
The removal of the dark organic midden from Blocks C, D, euid E precluded the
recovery of ceramics in them. In Blocks A and B, ceramics were found most
frequently in the rpper 50-60 on (1.6-1. 9 ft) , but they were present in minor
amounts to 90 on (3.0 ft). Hie vertical distribution of temporally sensitive
ceramic types indicates that the midden had been disturbed. Consistently,
types with discrete chrcmological spans c^fiearad together in the same level.
Looting apparently disturbed the integrity these deposits. However, the
total ceramic assemblage exhibits several patterns. Late Woodland (38) and
Mississippictn (3) sherds were inconspicucus parts of the total (Table 22) .
Hie assemblage is doninated by sand-tempered ceramics (81.5%) , 70%,
including 8.4% Alexander series ceramics, reflects a strong occapa±icr\ during
the Middle Woodland and Late Gulf Pormational. Nine hundred and thirty-seven
fiber-tenpered sherds (15.6% of the total) also reflects a Middle Gulf
Formational occvpation of this site. Hiis agrees well with the Middle Gulf
Fomational sites investigated in this project in nearby sites.
miE£ 22
Oeramic frequencies by tenper, 22lt621.
Shell
3
0.1
Grog
36
0.6
Bone
2
0.1
Limestone
133
2.2
Sand
4,886
81.5
Fiber
937
15.6
Total
5,997
The 5,997 sherds recovered from Blocks A and B is s\un;n:isingly large, far
exceeding the number found at any other site. Perhaps this indicates a Gulf
Formational and Middle Woodland occupational intensity unknown at any other.
As in all other eissemblages, the most frequent category of chipped stone
implements were fragnents of unidentifiable tools (480: 32.6%) v4iich are
grouped in the "Other Uniface and Bifaoe Tools" in Table 23. Hie hi^ number
of projectile point/)aiives (32.4%) reflects significant agreement with other
assemblages. Bifaoes were the next most frequent category, followed closely
by scrapers, and drills and perforators.
89
23
Frequency
Projectile Point/Knives
Beachum
1
Benton Short Stanomed
8
Big Sam^ Side Notched
1
Bradley ^iJce
1
Cotaco Creek
3
Cumberland
1
cypress Creek
2
Eva
6
Flint Creek
22
Gary
2
Greenbricu:
2
Kirk
14
Ledbetter /Pickwick
7
Little Be^u: Creek
59
Mclntire
2
Mississippian-WOodland Triangulaur
1
Mcanxw Mountain
12
Quad
1
Residual Stenmed
29
Residual Trianguleu:
2
Sykes-White Springs
3
Tcmbi^see Stenmed
1
Vc-’jghn
1
Wade
1
^identified projectile/point/knife
1
Distal fragment
119
Medial fragnent
80
Proximal fragment
87
Lateral fragment
7
Subtotal
476
32.4%
Scr^^rs
104
7.1
Drills, Perforators, etc.
105
7.1
Bifaoes
152
10.3
Other Uniface and Bifaoe Tools
524
35.6
Cores
44
3.0
Preforms
65
4.4
Total
1,470
A total of 2,261 utilized flakes were recovered frcni the site.
Interestingly, the 0.5 inch (1.3 cm) and 0.25 inch (0.64 cm) size flakes of
utilized fladces occurred in atproxinately the sane frequencies. In ccntrast,
the non-utilized flake pattern apparently reflected a preference for l2u?ger
flakes. Of the 48,576 ncn-utilized flakes or d^itage recovered frcm the site
86.6% were 0.25 inch (0.64 cm) (Table 24), only 12.9% were 0.5 inch (1.3 an),
and 0.3% were over one inch (2.54 cm) in size. As in other midden mound
assentolages, this pattern indicates reduction of ccAbles to bifaoe tools. The
df^itage raw material was dcminated by Canden chert, both heated (78%) and
unheated (13.9%) , with minority raw materials malting up the remaining 8.1% of
the debitage.
90
■■a- T-^T—
Percentage _ J
194
0.3 ^
0.5 inch
6,272
12.9 1
0.25 inch
42,091
86.6 1
Non-iztilized flake - Prismatic
9
0.1 1
Other
10
0.2 1
Total
48,576
-J
A total of 1,470 chipped stone tools were recovered at the Hickcary site.
A total of 71 id^tifiable ground stone and 120 unidentifiable frai^nents of
ground stone tools were recovered fron the Hickory site (Table 25) .
Hannerstcxies were the most frequent tool type (21) , and they often were
canbined with millers or anvilstones. Ihe next most frequent tools were
anvilstones (17) , and these were often also used as hamnerstone or abraders.
One grooved axe was recovered along with a few ornamental artifacts. Hie
ground limonite and hematite also had been provided for ornamental pigment
extraction. The main uses of ground stone tools, however, were hammering,
pounding, and edirading.
Hie ratio of tools to debitage in the Hickory site lithic asseniblage is
1:35, much higher than at the previous sites investigated. Hie flake size
ratio is 1:32:216 and reflects production of chipped stone tools fron cobbles
at the site.
22It621.
•nOLE 25
Qcound stone tool
Ca
Abrader
Anvilstone 2 1.0
Anvilstone-Hanrnerstone 3 1.6
Pitted Anvilstone 10 5.2
Pitted Anvilstone-Abrader 2 1.0
Bead 1 0.5
Bead Preform 1 0.5
Discoidal 1 0.5
Gorget 1 0.5
Grooved Axe 1 0.5
Hammerstone 21 11.0
Mortar 1 0.5
Muller 4 2.1
Muller-Hanmerstone 1 0.5
Muller-Pitted Hamnerstone 2 1.0
Ground Hematite 4 2.1
Ground Limonite 2 1.0
Ground Stone Flakes 11 5.8
Unidentified Fragment 120 62.8
Total _ 191 _
A total of 126.0 kg of introduced rock was recovered in the
investigations. As at the other sites investigated, this was dominated (89%)
by ferruginous sandstone, locally avedlable in the nearby ijplands. Only eight
grams of faunal reroedns were recovered from the site. It was calcined and too
fragmentary for taxonomic or element identification.
91
EEMBRES
Only four cultural features were encountered at the Hickory site,
including three bcisin-shaped pits and one litliic dd^itage cluster. One Kirk
stemned projectile point Wcts ccntained in a pit, indicating an Early Archaic
association. others conteLined no diagnostic artifacts, but stratigraphic
position suggested that they were Middle Archaic (Sykes-White Springs/Benton) .
All these features were first encountered in Stratum IIB. All features
contained a relatively large amount of unmodified flaking debris and the v^per
porticos of the features produced most of the materials.
It is difficult to discern their primry function. Ihese features may
have been ctssociated with stone tool manufacturing or utilized eis refuse pits.
It is speculated that Feature 1 wzis a pile of lithic debitage. Of 141 flakes
recovered, only two were utilized (0.25 in or .64 an flakes) . These flakes
were in a pile which had a poorly defined boundary. The three beisin-shaped
pits contained a moderate amount of botanical renains in additioi to the
lithic materials. The density of botanical renains in the feature fill was
much hitler than that in the general midden. Unmodified introduced rocks were
also present in each feature. Shapes and coitents of these features indicate
use as refuse or storage pits.
STRATIGRAPHY
The Hickory site weis formed through alluvial deposition of sancfy loam
during episodic flooding. The profile was characterized generally by an upper
dark organically stained midden resting uncanformably on a truncated
yellow-brown paleosol with well-developed structure on a blue-gray gleyed clay
d^sit. Within these two general zones, eight strata were identified
(Figures 16 and 17) . The dark organically stained midden was ^proodmately
1 m (3.3 ft) thick. The lowsr half of the midden was hard and caiposed of
cemented manganese concretions. This cemented zone was probably caused by
eilmost continual water table saturation of the profile. This manganese
OOTKxetion zone was unique to this site.
The contact between the pedeosol and the overlying midden was abrvpt, and
often had an intervening lens-like deposits of coarse sand (Figure 17) . These
sandy pockets were scattered throughout the edges of the pedeosol, reflecting
a turbulent environment at one time in the site area. The sand deposits on
the top of the pedeosol likely resulted frcm the erosional episode that
removed the A and part of the B horizon at this and other sites in this part
of the flooc^lain. The pedeosol at this site was similar to other eaqxssures,
i.e. yellcw-brown color, well-developed prismatic structure with
post-develcpment polygonal cracking throughout. The sand pockets were more
frequent around the edges of the site. The central core, or epicenter, of the
landform wets still intact.
At this site, excavations continued to the bcise of the paleosol, e^gosing
the bright orange C horizon beneath the well-develqped B. The oxidation of
the C horizon was likely due to induration of iron-rich minerals in the
perched water table. Excavations continued into this zone (VIII) confirmed
tliat this was the Pleistocene valley floor.
Site profiles indicate that the initial process of site formaticai involved
the deposition of yellow-brown loam in the center directly rpon a sulanerged
clay zone. These d^x>sits built the site to the current configuration. The
paleosol developed, but portions auround the periphery were removed and
replaced subsequently lay sands. The upper midden zone then accumulated above
92
BLOCK 0 WEST PROFILE
t03Sral marlcers. Feature 11 at the Beech site
nnn^-jt-iwpri a Gary and a Little Bear Creek projectile point/lcnife vhich dated to
4, B.P. Feature 1 at the Oak site oont£UJied a Little Bear Credc, a
Mclntire, and a Benton projectile point/knife, vhich were dated at 3,
B.P. Both dates are fairly consistent vrith others in neaihy areeis. Severed
Middle Archaic pits vdth diagnostic projectile point/knives were also dated.
However, only two eire consistent vdth others in this project. These are Xmo
dates &rxn pits containing Benton projectile point/knives. At 22It623, a date
of 5, B.P. was obtained from a pit containing only Benton projectile
point/knives. One segment of Feature 7, a oontpound pit/ landfill ocnplex at
the Oak site, contained a Benton, projectile point/knife dated to 5,S B.P.
However, two other se^nents of this sane feature dated 4, B.P. 6md
4, B.P. These segments did not contain projectile point/knives and
likely are related to a later use of this area during the Late Archaic period.
Feature 7 did contain cme Mclntire. This projectile point/knife type was
associated with a 3, B.P. date in Feature 1 at this same site.
ME 30
lf«<ied front the
Early Archaic through the Late Woodland stages. Although data recovery
focused on the Late Archaic, informatif»i on other occ\5)ations was obtained.
Archaic Stage; Perhaps the most surprising aispect of these sites was the
discovery of the Early Archaic paleosol in an atypical landform to other
occurrences.
Only a small saitple of the peileosol was screened, and it produced chipped
stone tools and debitage. No diagnostic projectile point/knives were
recovered in the sample. However, there were three Kirk's recovered ficcm the
site, and these probably came from the paleosol. The initial Middle Archaic
Eva/Morrcw Mountain usually found in the top of the paleosol also was present.
Although only one Eva was recovered, 11 Morrow Mountains were found. This
sv^ports the Middle Archaic occipation at the site, and the ccaponent is
likely conteiined in the top of Strata VI (paleosol) and V and IV. This
deposit appecirs light in artifact density, but has high integrity.
The late Middle Archaic Benton occupation signals a more intense use of
these sites. The main archaeological deposits from this and the following
Late Archaic occipation were pits. Unfortunately, the intensity of pit
digging and mixing of both pit and midden material caused problems in
interpretation of the activities. The presence of so many pits, however,
indicates storage and/or refuse disposal during this and later periods.
The ceremonial blade cache, artifact clusters, broken siltstone atlatl,
and many of the ornamental items probably were associated with either the
Benton or Little Bear Creek occi;pations.
While the focus of the investigations at this site was the Late Archciic,
Little Bear Creek period - corroboration of stratigraphic sequence vreis only
partly successful, because it was difficult to separate the coiponents.
Mixing of materials mitigated against adequate interpretaticn. In spite of
the mixing and poor association of pit features, the Benton oocipation at this
site was differentiated by fired aggregates from hearths and prepared areais
with multiple hearths. These features suggest that the Benton period
occupations was more transient than at Walnut, Poplar, and Ilex sites. The
Benton occi^jation there also appears to have been light at the Hickory site as
well.
Sane settlements investigated had hearths and large prepared cu:e2is and
burials, and others did not. The dates ctotained on the Benton occupation are
later than nost other, yet still cotpatible, averaging 5,300+72 B.P.
106
I
Four pit features associated with the Little Bear Cre^ ocnponent; three
were dated and averaged 3, B.P. These are the (xily ones thus
obtained associated with the late Archaic Little Bear Creek horizon in the
Upper Tcnbi^see Valley. The material in these pits included a wide assortnient
of chipped stone tools, ground stone tools, d^itage, and charred plant
reenins, eepecially hickory nutshells. These materials indicate a wide range
of acti^ties, including subsistenoe-related, hide-working, and tool
manufacturing activities. The presence of 53 Little Bear Creek and Flint
Credc projectile point/knives svpport a substantial Late Archaic occupation of
these sites. It is likely that at least seme of not ed.1 of the 14 poshnolds
were from this occupation.
The reneins of the later Gulf Formationed. and Middle Woodland ooctfiations
were mixed, but two Middle Gulf Formationzd and three Woodland pits weire
encountered. The Gulf Fozmtian£d pits at the Oak site contadned sparse
material, but they provide evidence of stone tool manufacturing activities.
The Woodland pits contzdned more material and revealed subsistence-related,
tool manufatcturlng, hide and woodworking activities.
The investigations demonstrated that these sites were essentially
abandoned after the Middle Woodland. Only intermittent visiting during the
Late Woodland in inferred, and virtually no occupation during the
Mississippian is documented.
EXCAVATIC»«S AT THE ARALIA SITE (22It563)
The Aralia site was located approximately 13.4 km (8.4 mi) northeast of
Fulton, Ms (Figure 1) . It was situated at the edge of the floodplain of the
Tonbigbee River (Figure 21) . The Hickory site, a floodplain midden mound
investigated in this project, was located only 750 m (246 ft) from the Aralia
site in the floodplain. The site surface followed the steep slope (10-15%) of
the wall in the eastern two-thirds, but was semeshat flatter (8-12%) in the
western third and possibly extended on the floodplain edge. However, the
western edge of the site had been disturbed by a dirt road and floodplain
clear-cutting of the vegetation. The sloped surface of the site also had
severed erosion£d gullies and tree throws at the time of testing and
excavation.
The site appears to have been a narrow sandy terrace lying against the
valley wall which subsequently has been covered Isy oollxivial and alluvial
materials. Colluvial deposits resulted from downslope movement of the sanc^
sediments cdong the valley wall. Several seep springs occur at the base of
the vcdley Wcdl near the site at a simileu: elevaticxi. Active small slurps euid
mass movement cuecis are oonmon in this area.
FTETJ) METmODS
The site was initially located during a surv^ of the Caned Section of the
Tennessee-Tcmbigbee Waterway by Blakeman (:19) , who noted transitional
Archaic Afoodlland, Miller I, Miller II, and Miller III cultural material and
recormended it for further testing. The site was tested in as part of
the large-scale Canal Section testing program (Bense ) . Testing included
excavation of two 2x2 m (6. 6x6. 6 ft) units and a surface collection in the
clirt road along the base of the slope (Figure 21) . Testing indicated that
vdiile traces of Late Archadc through the Late Wbodland period use were
present, the site deposits were deminated by a middle Gulf Formational stage
(Henson Springs phase) ocevpation. Little was known of this culture in the
107
waterway at that time as the deposits in other sites had been mixed and at
best limited to a few intact features. Because of the limited knowledge and
the abrupt appearance and the unusual oeramic series associated with this
culture, the isolation and integrity of this oonponent at Aralia, it was
reoamnended for data recovery.
Excavaticxis were ocnducted at the Aralia site between Noventer of and
March of . Ite objective of the investigations was the recovery of a good
sanple of the Henson ^;>rings oonponent. As at the Ilex site, which was under
investigation at the same time, badchoe trenches (five) were the initial means
of site investigation to aid in the placement of hand-excavation units. These
provided Informaticxi on the extent of the site, since the midden zone was
buried 20-30 on (7.8-11.8 in) below the surface. The midden appeared to be
discontinuous, approximately 50 m (165 ft) north-south by 40 m (132 ft)
east-west. Based on the infonnation from the trenches and test units, the two
test units were placed in areas of the richest midden deposits, and these were
expanded accordingly (Figure 21) . No chemical or visual coring was performed,
since the necessary infonnation had been obtained throu^ stratigraphic
trenching.
The main excavation units were initicdly 4x4 m (13.2x13.2 ft) in size
(A and B; Figure 20) . These units were later e:q>anded to follow features and
esqnsed midden. Block A eventually incorporated 40 sq m of excavated area,
vhile Block B contained 36 sq m. Tb explore the horizontal extent of the site
and to identify possible activity areas, an additional seven 1x2 m (3. 3x6. 6
ft) test units and two 1x1 m (3. 3x3. 3 ft) units were placed to the east and
south of Blocks A and B. Four shovel trenches were excavated by hand to
investigate the extent of the midden or features e)qx>sed in other units.
These were 50 cm (1.6 ft) de^ and varied from 2 m (6.6 ft) to 8 m (26.4 ft)
in surface curea.
CULTURAL REMAINS
Artifacts recovered from the site included ceramics, stone tools, historic
eutifacts, and floral and faunal remains. The following sections summarize
each material class recovered from the site.
A total of 3,041 sherds were recovered from Aralia (Table 31) . While four
tenper types were present, grog, limestone, and fiber made vqp only 2.1% of the
sanple. Sand-tempered ceramics dominated the assemblage (97.9%) and of the
specimens identified, 69% were Henson Springs types (Alexander series,
Columbus Punctate or Shuthsonia Zone Stamped) . The combination of Henson
Springs with the large amount of Plain (28.2%) , suggests strongly that this
was the Henson Springs culture of the late Gulf Formational stage. The few
fiber-, limestone-, and grog-tenpered sherds (65) likely reflect short visits
during these cultural episodes.
While sand-tempered ceramics were usually the most frequent types
encountered in most floodpleiin sites investigated in this project. The Aralia
site yielded them at the rate of 60-70% greater than other sites. Other
ceramic assemblages contain a much lower percentage of diagnostic Henson
Springs ceramics (ca. 10-20%) than the 69% from Aralia.
Grog
47
1.5
Limestone
9
0.3
Sand
2,976
97.9
Fiber
9
0.3
Total
3,041
A total of 983 diipped stone artifacts were recovered from the site
(Table 32) . As in most cisseirblages in this project, there was a high
percentage (40.9%) of fragnents of unidentifiable chipped stone tools. Ihis
category was included in the "Other Uniface and Bifaoe Tcx>l" group, and caused
it to be the most frequently (43.8%) encountered of the chipped stone tools.
The second most frequent tool type recovered v«s projectile point/knives
(30.7%) . The relative amount was even higher (52%) if only identifiable tools
were considered.
Findings here agreed with the other flooe^lain assemblages, except for the
high range of projectile point/knives. The Hickory site, only 750 m (2,461
ft) away, had a similar high percent of projectile point/knives.
TSffiU 32
stone tool frequencies by type, 22It563. _
Tool Clatss/Type Frequency Percentage
Projectile Point/Knives
Cotaco Creek
1
Flint Creek
98
Gary
2
Little Bear Creek
36
Mud Creek
1
Residual Stenmed
10
Wade
1
Distal fragment
66
Medial fragnent
27
Proximal fretgment
51
Subtotal
293
30.7
Bifaces
64
6.7
Preforms
91
9.5
Cores
15
1.6
Scrapers
27
2.8
Drills, Perforators, etc.
46
4.8
Other Uniface and Biface Tools
417
43.8
Toted
983
Preforms (9.5%) and bifaces
(6.7%) were the next in
frequency among
chipped stene materials. The drills and perforator group follcwed with 4.8%,
and cores were fcwnd at the lowest rate.
A total of 1,881 utilized flakes were also recovered from the site. This
large nuntoer reflects a consistent use of these expedient and disposable
tools. The 0.5 inch (1.3 cm) flakes were almost eis abundant (43.8%) as the
0.25 inch (.64 cm) flakes (49.9%), indicatir>g selec:tion of larger flakes,
rather than a proportional use of aveiilable flakes.
A total of 43,475 flakes of debitage vnere present in the Aralia
eisaenblage. The dGminaoit raw material was Camden chert (97.3%) , 85.7% vets
heated. The remaining 2.7% of the d^itage was made up of several local and
exotic raw materials.
Snell, 0.25 inch (.64 cm) flaJoes (87.2%) dominated, 0.5 inch (1.3 cm)
(12.7%) were next in abundance, and few flakes greater than one inch (2.54 cm)
in size (0.1%) were recovered (Table 33). The debitage was quite localized,
and the 2x2 m (6. 6x6. 6 ft) test unit 76.4S/87W that was the furthest v^lope
(Figure 21) contedned the most d^itage and the hi^^iest proportion of
projectile point/knives. The 2x4 m (6.6x13.2 ft) unit 4 ra (13.2 ft) to the
west (78S/98W) oontadned the largest amount of debitage.
of dehi
22It563
ze Grade
0.5 inch
0.25 inch
5,501
37,935
43.475
The tool-to-dd3it2^ ratio at this site is 1:46, one of the lowest ratios
of all sites investigated in this project. The proportion of flake sizes is
1:141:973, also the lowest of all sites. The likely reflects that tool
manufacturing was a primary activity at this site which results in a lot of
debitage.
A total of 145 ground stone inpleroarts viere recovered fcon the Aralia
site. Most (71.0%) were fragments of unidentifiable tools (Table 34) and
flakes of tools (13.8%) . Of the 22 identifiable ground stone tools, the most
abundant were hammeretones (5.5%) . Pounding stones sudi as mailers, mortars
2uid anvilstones were also present (7.0%) . Ground hematite, limonite, and a
stone bead inplied personal adorrxnent.
T«£ 34
GkDund stone tool
22It563.
Hamnerstone
Pitted Anvilstone
Muller
Mortar
Bead
Ground Limcnite
Ground Hematite
Unidentified Ground/Polished
Stone Fragment
Other (Ground Flake)
A tot£d of 94,868 grams of introduced rock was recovered from 22It563.
Sandstone wzis the most abundant material found, followed by oobble/pebbles.
Fire-cracked chert chunks and petrified wood were next in abundance. One
cautionary note should be made, however, concerning the amount of introduced
rock. Sandstone and oobble/petoles (91.2% of the grorqp) occur naturally in
the sediments of the site, and the amount brought on the site by the farmer
occupants could not be determined.
w
Ihe historic remains were recent and consisted largely of dtnped household
d^ris. Shell casings were also a frequ^t surface artifact reflecting local
hunting activities.
Faunal remains frcm 22It563 weze represented by a few very small and y
usually calcined fragments. Since these sanples were too fragonentary to
permit positive indentification/ no further analysis of these sanples were
conducted.
FEAITORES
A total of 12 features were identified at the Areilia site and all were
apparently associated with the Henson brings occupation. The features
included:
8 ceramic clusters
2 pits
2 dark stained midden features
All the ceramic clusters were concentrations of fragments from single
broken vessels. The clusters were all lying on or in the buried midden zone
(Strata IV) and were not contained in any visible pit. i^jparently th^ were
piles of sherds (and occasionally other materials) vAiich were at or near vri>ere
the vessel had broken. Of the eight vessels, the ceramic types included five
Alexander Pinched, two Alexander Incised, and one Residual Plain. At least
two of the vessels were wide-mouth utilitarian bowls.
Ihe broken vessel concentrations in Block A (four) and two in Block B were
outside the dark organically stained midden of Features 10 and 11,
reflectively. In addition, one ceramic cluster was found in test unit 78N/98W
and 84N/100W. These were found within the Henson Springs midden zone, the |
fiparent occvjpation land surface of the time.
The two pits encountered, both in Block B originated in the midden zone
containing the Henson Springs conponent and were outside the dark organically
stciined midden of Features 10 and 11. One contained diagnostic Alexander
Pinched ceramics. These features were, apparently, small V-shaped gullies
filled with refuse, rather than purposefully dug pits. They were similar in I
size; 43 can (1.4 ft) long, 40 can (1.3 ft) wide, and 45 can (1.5 ft) deep.
The largest and most ccttplex features encountered at this site were the
two large cjrganically stained areeis designated as Feature 10 in Blocdc A and
Feature 11 in Block B, In both Ccises, only the eastern edge of the features
was defined, ard the blcjck units were extended to the west in efforts to find
the western perimeter. The exigencies of time, hcwever, precluded the I
cxanpletion of that task. The features appeared to have been organic^ly
stained, somewhat more cxaipact and circular than areas within the midden zone
vdiLch were present in most of these sites.
Convoluted sediments were found along the top of the features. The
sediments were not hcmogenecxis, and small lenses c:a. 50 cm (1.8 ft) in
diameter and 5-10 on thick (2-4 in) thick were identified. Seme mixing with |
the subsoil was noted in the field. These features were probably either dunps
or areas of residential activity. The features both contained a hic^r
density of cultural material than other midden areas.
STRATIGRAPHY ^
The site stratigraphy was characterized by sandy, alluvial and colluvial
sediments. Lamellae were canton belo? 60 an (1.97 ft) and were probably
2
associated with a perched water table levels. The depositior\al record at
22It563 documented periods of extensive erosion, oolluvial deposition,
suqper-saturation and flooding produced by Tcmbi^bee River flooding. These
processes continue to affect the site envirotment to the present, airl no doubt
they had a significant influence on prehistoric settlement of the eastern
valley margins.
One of the more peculiar aspects of the Aralia site was the steep slope of
the site surface. The present slope was e^jproodmately 10% and examination of
stratigrajMc profiles revealed that the slope of subsurface units toward the
floo(^lain reached 15%.
The surface at the site and adjoining eu?eas at the base of the Vcdley wall
were alternatively eiqieriencing episodes of erosion and deposition at the time
of excavation. Riviilet, channel cutting, and downslope movement were all
active. The base of the slope at the viestem edge of the site revealed
cut-and-fill deposits prctebly due to the meandering of the Tcnfei^see or
tributary streams. A channel could have been active in this position during
some of the prehistoric occupations.
In all likelihood, the base of the slope did contain evidence of
prehistoric settlannent that was destroyed by erosion and recent disturbance by
improvannents to the access road. The recovery of 21 Wheeler sherds in this
road during the (Bense ) testing investigations provided the only
indication of the temporal placement of occupations in this area. It is
noteworthy also that the position of the river was likely much closer to the
eastern vedley wall them currently (1 km: 0.6 mi) , as indicated the
cut-and-fill deposit along the toe of the slope. If so, this may have been an
important factor in the selection of the site area for settlement.
Nine strata were identified at the Aralia site (Figure 22) . Ml were
loams or sandy loams with high sand contents. One stratum (II) 20-40 cm
(7.8-19.7 in) below the surface had been visibly altered by the introduction
of organics, and it contained the most cultural material. Althou^ this zone
was darker than non-culturally modified zoies, it wets not as dark as the
floodplain midden zones.
CHRONOMETRIC DATING
One radioccirbon date was obtained frcm Zone 2, the cultured midden of the
Henson Springs. The date was 2,379+50 B.P. , vAiich is consistent with the few
other dates of this culture in the Tcmbigbee Valley. The date, along with a
paucity of ceramic types such as Smithsonia Zone Stanped and Columbus
Punctate, cenfirm the ocev^tion of the site early in this phase.
SOMiARY
The investigations conducted at the Aralia site confirmed that the primary
occupation was during the Henson Springs jtose. The materials ar»d features
within Zone 2 exhibited no mixing and can be considered as an isolated
component of this piiase. However, the distributiOTi of artifacts within the
upper three stratigraphic zones revealed the predominance of Alexander series
ceramics and Flint Creek/Little Bear Creek projectile point/knife forms,
regardless of stratigraphic position. Secondary ocevpations were present,
although they were limited. These coiponents were identified Wheeler,
Baytown, and Miller ceramics. The majority of these were recovered from the
surface of the site along the base of the slope during testing (Bense ) or
fror. disturbed contexts within the Block B excavation unit.
113
Ths HensOTi Springs assarblage was chcuracterized by its honogeneit^ across
the site. Although eroded sand-tenpered sherds ocnprised the majority of the
overall oerandc sanple, the Alexander series, including pinched and incised
varieties, ytddti constituted the majority of the decorated sherds. A detailed
study of this ijnix:)rtant assemblage is presented in Giapter IX of this report.
Lithic implements were dominated by ocnplete cmd fragmentary Flint
Creek/Little Bear Creek projectile point/knives, i.e. 13.8% of all chipped
stone tools and 90.9% of all identifiable projectile point/knives. Other
implements included relatively limited numbers of bifaoes and cores and a
variety of preforms, scrapers, and drills-perforators-reaners. Although
scrapers occur in a variety of styles, they are outncnbered by drills-
perforators-reeDners by a ratio of apprcodmately 2:1. Overall, chipped stone
lithic debitage and implement samples reveal that stone tool manufacture was
at least a consistent activity and ah overwhelming preference for heated
Camden chert (97.3%) . Finally, ground stone implements were poorly
rq>resented in this assemblage; however, several mailers, mortars, pitted
anvilstones, emd a single, fragnentary bead were recover^. Ihe diversity of
lithic inplements suggest a wide range of activities within the Henson Springs
phase ccnponent.
Oocvpation at the site was eppar^tly semipermanent, althcu^ seme
movement of population to other sites during portions of the year probably
occurred. The relatively low number of items in certain iirplement categories
(e.g., ground stone, scrapers) suggests a low occupaticxi intensity, a limited
number of intermittent, serniipermanent occupations, eind/or use of multipurpose
tools. Ihe projectile point/knife forms were used for a Veuriety of purposes,
e.g., projectile, knives, drills/perforators, which perhaps obviated the need
for other speci£j.i2ed tools.
The lithic sample from 22It563 provided further clarification of the
stylistic elements and the technological patterns in an cissemblage from the
Henson %>rings phcise in the l^per Tcmbigbee Vedley. As indicated earlier,
both the Flint Creek and Little Bear Creek projectile point/knives in this
sample vere probably a part the Henscxi brings assemblages. Althou^
separated for descriptive purposes, these two forms overly in terms of
stylistic attributes, size, and technology of manufacture.
Eeurly stages in the hafted biface reduction were represented by Stage 1
and 2 preforms. Primary and secondary decortication flakes produced by the
reduction of ocbbles to the preform stage were not well represented in the
debitage suggesting that at le6ist seme initiad reduction tock place off-site
at the sources of cobbles. Ihe initial manufacturing sequence at the site
apparently involved both bifacial reduction of cc±bles and large flakes
derived from cobbles. Ihe reduction of relatively large flakes, or possibly
split cobbles, appears to have been the favored steurting point in the
production of projectile point/knives in this sample.
The more extensive shaped and thinned stage is reflected in the biface
bletdes. The retention of flat, unthinred bases on finished bifaces is
evidence of the initial striking platform produced by a proximal flake-blank
orientation. Flat or faceted bases, often ctxisisting of ooUale cortex,
ccmmcanly occurred in samples of Flint Creek and Little Bear Creek projectile
point/knives (Jolly :18; Cambron and Hulse :51,82). llus is
particularly true of this assemblage and others from the Ipper Tcmbi^aee
Valley. Nearly half of the 98 Flint Creek projectile point /knives in this
sample have faceted or "imfinished” bases.
In the distribution of materials and features at the Arcdia site, three
analytical units were formulated: Zone 1 included all strata and materials in
115
strata I-III, levels above the midden horizQn, Zone 2 was the defijied middoi,
and Zone 3 included 2dl strata below IV. This isolated the oonoentration of
cultural debris in the dark-oolored midden. Strata IV, including the two large
stained midden features (10 and 11) vAiich appear to have been coeval segments
of a discontinuous cultural midden. Outside of these areets, the cultur2d zone
was nuch lifter in organic staining with a gradual horizontal change frcm the
features. In the esaavation units placed furthest t^lope no staining was
encountered; however, cultural ddaris was still confined to Strata IV.
The eutifacts reoovered from Block A and B indicated possible differences
in the kinds of activities conducted in each area. Block B had mc»:e cultural
material, althou^ the ceramic frequency recovery was nearly identical. Block
B h^td some mixing of diagnostic ceramic t^pes and no such mixing was apparent
in Block A. Lithiics in the two blocks revealed l2u:ger nixrbers and a greater
diversity of lithic implements in Block B, with an emphasis on maintenance/
rejuvenation, and some seccxidary manufacturing. Perhaps this was the location
of a lithic workshcp).
In contrast, the tools in Block A assemblage had a wide range of types
suggesting a gradual accretion of ddsris. The occurre3x:e of the dark-oolored
midden areas in both blocks also implied ^trcurtive or residence activities
that resulted in the deposition of organic residue.
As noted previously, the 2x6 m (6.6x19.7 ft) (78S/98W) test unit produced
a disproportionately hi^ nvmber of projectile point/knives. Of these, 28
(62%) were clMsified as Flint Creek. Prelimineuy examination indicated
multiple uses such cls drilling, cutting, and projectiles. Perha^ some
maintenance and/or rejuvenation of these Implements was adso represented. The
fact that the midden in Blocks A and B was not present in this upslope area
suggested that it was not an intensive processing area for biotic resources.
The Henson Springs oonponent v®s interpreted as a base camp settlement.
This interpretation was based on the diversity of implements, as well cus
activities. The time span of the occvpations represented within madden Strata
IV was most likely no more than several hundred years. Gcnponents bracketing
the Henson Springs occvjpation were represented, but such occupations were
apparently of shorter duration and intensity.
EXCAVATIONS AT SITE 22It606
Site 22It606 was located approximately 13 Jon (20.8 mi) north of Fulton, in
Itawamba Co. , Ms (Figure 1) . The site was situated on an isolated segment
outlier of the valley wall which was part of the Pleistocene terrace
overlooking the floodplain of the Tcmbigbee River. Mud Creek flowed along the
southern and eastern borders of the landform and separated it from the
adjacent uplands by dawncutting in the soft sandy sediments. The surface area
of the terrace remnant was cpprcximately 60x140 m (196x460 ft) . The southern
and western edges of the landform were very steep (Figure 23) , however, the
east emd north slopes v^re much more gradual. The site area was perch^ 4.5 m
(14.8 ft) above the flood^ledn and adjacent to it. The site had been cleared
and a hcmesite, garden, and store were just being demolished at the time of
testing in , and heavy machinery associated with waterway ccxistruction had
damaged the northeastern third of the site area.
FTKTD METHCnS
The site was tested in Phase I of this project between September and
October and included a controlled surface collection and two 2x2 m
Contour Intorvat: 50cm
Figure 23. Topographic map and excavation plan, 22lt606. Note: Arbitrary Datum = 100.0
Figure 23 Topographic nap and excavation plan, 22It606
(6. 6x6. 6 ft) units in areas vihich produced materials diagnostic of a Late
Woodland/Mississippian occvj^tion (Figure 23) . liie test vmits were excavated
to an average depth of 100 on (3.25 ft) . Five additional stratigraphic cuts
averaging 50 cm (1.8 ft) deep were made to obtain data on cultural, and naturad.
stratigraphy and geonorphological processes at the site. Testing adso
included machine-*stripping of the plow zone in two 2.5 m (8.2 ft) wide
trenches axxoss the site. IMs exposed 14 prehistoric features below the plow
zone, all of which were either conpletely removed or sampled. Ihe informaticxi
fron testing revealed that this site contained intact featires of the Late
HOodland/Mississippian period. In all other sites investigated in the Upper
Tcnbic^bee Vadl^ in this project, this coiponent had consistently been very
mixed through looting and/or cultivation. Features at this site of the late
aboriginal occupation of the area, itede it a logicad choice for excavation to
recover a good sample of the Late Woodland/Mississsippian features.
Intensive data recovery wais conducted in Ihaise II of this project during
December of . * focus of these investigations was to expose the Late
Woodland/Mississippian features and e»::avate a r^}resentative sample.
After olearing of the vegetative cover, mostly tadl weeds, with a tractor
and bush hog, the plow zone south of the eaist-west testing trendi was
machine-stripped with a bulldozer. A snaller dozer then removed the final few
centimeters over the features. The daurk stains contraisted with the lifter,
yellcwer, subsoil were flagged and initially investigated by shovel skimming.
The features were then excavated using the standard procedures.
CULTURAL REMAINS
Cultural materials retrieved from 22It606 investigations included
ceramics, lithic materials, historic artifacts, and biotic remains.
A total of 1,730 sherds were recovered from the site, and they consisted
of five major temper grotps: shell, grog, bone, limestone, and sand.
Sand-tempered sherds were the most frequent (45.1%) usually followed by grog
(37.8%) (Table 35) . Shell-tempered ceramics were the next most frequent
(11.7%), vdiile limestone (3.3%) and bone (2.1%) were the least Sequent. The
Late Woodland diagnostic tenper types (grog, bone, and shell) vAien ccmbined,
daninated the assemblage (51.6%) . The sand-tempered ceramiics included a few
late Gulf Formational (1.3%) and Middle Woodland types (10.6%), however, the
bulk was composed of either eroded or plain specimens (88.1%) .
•OBIE 35
Oeacamic frequencies by temper, 22It606.
Temper
Frequency
Percentage
Shell
Grog
Bone
Limestone
Sand
Total
202
654
37
57
780
11.7
37.8
2.1
3.3
45.1
1,730
A total of 457 chipped stone tools were recovered from the site
(Table 36) . The most frequent category vas Other Unifaoe and Biface Tools
(69.6%). However, most were (97.5%) fragments of unidentifiable tools. The
next most frequent category was projectile point/knives ‘ ,.6.4%) . Screpers
vdiich was (59.4%) of the identifiable chipped stone tools. Bifaces were the
next roost frequent tool type (8.5%) , followed by scrapers (3.9%) , and drilling
118
Percjen
Projectile Point/Knives
Big San ip in CO c
r " S
5 C ^ i > C
> 5 2 g <=> 5
^ o> O) c S
:. allll
■ '5 O (O CO n u
II li
(0 9 (0 0)
E ig I s
1 -5 i l
2 0) > O
c5 c * — c
(0 o >. '? .
* w-o « C u
^ £ c cj 'q. (q
®=il -t
SS5 “
cc c« 8 2? ® >
> ® oicc o -
2 « £ > ^ 5"
i^£i«5E
c »-■ -S a -
■£ O 3 ^ ^3 ^
o»S-'§ « 2
i c a J ® O)
= '^ o 5 3 —
® '5 § ® g O)
>■ 10 £ > c i£.
® S = t
c s ®
o E o ®
g “
o E S ^
a a “ $
. o ^ o>
as ® ®
S 5 “ c
iS9-5 2
n V 2 nr ■“
to g 52 -2*
V T- o
DC c £ i o>
> > « s s
2 6 ■o 5 « . «
® ?. a 8 i
c^sT® g-2
2 SE sr i 2
to O c O O 8 sc
^ 0) o»
•ts « c
» 0) s
^ C ts
||i
1 is
^ S 3
<0 U
iii
>«£ ®
(D (0 ?
o o I
;* >. >s c
I 2 (0 ts E
2 to to •“
C -O ® -O
i I- <» I-
S “ c a
■5 5 ® 5
tN ^ o g> O
Hi 8 £ ® ® ® •
E^
Of
£l^= =
l*'wi
■lalliS^I
tai
n
iw!
tim.
iSSsS
m.
WM
/^§si
is*^sA
Figure 24 Stratigr^iiic profiles, 22It606
BadiocaaitMn dafaes, 22lt606 ((
Material
Cultural
Affiliation
oontJnuRd) .
Feature/
Lab
C-14
Uncorrected
(Corrected)
Provenience
«
Date
Date *
20
Segment B,
(middle
stratixn)
DIC-
730±55
A.D.
(A.D. )
wood charcoal
28
Segment B
(middle
stratvm)
DIC-
580±60
B.C.
( B.C.)
Middle
Archaic
45
North 1/2
DIC-
860±60
A.D.
(A.D. -)
wood charcoed
* Corrections based on fomula in Redpdi et al. .
SlMIRRy
Site 22It606 was located hi^ above the flooc^lcdn on a flat-topped
Pleistocene terrace vdiich had been cut off from the i:plands by a small
tributary stream (Mud Creek) . The site had been occupied intermittently since
the E2u:ly Archaic period to the present; hcMever, the major occupation was
during the Late Woodland/Mississippian period (A.D. 1,100-1,440). The
investigations focused on the features from this occupation, 12 of vdiich could
be definitely documented (all pits) . Other features likely associated with
the late Woodland/Mississippian ocnponent were a large fire hearth and
sandstone cluster. During this period, the site was apparently intermittently
occupied.
Some trends in the change of tempering was docunented 1:^ radioceubon
dates. During the Late Woodland/Mississippian period grog tempering was
preckxninant, ccnprising 50-60% of the ceramic assemblage, except from about
A.D. 1,250 to possibly A.D. 1,400, vdien it was 70-90%. Hie leurgest tempering
minorities were shell, shell-and-grog, and sand. Bone- ai>d limestone-tenpered
sherds were rarely encountered.
The shell-tenpered ware weis pnbably the most sensitive time indicator.
Although infrequently found, it displays meaningful tempor2d. variations.
Shell tempering was apparently introduced early, perhc^ by A.D. 1,000, and it
comprised at least 30% of the ceramic assemblage. By A.D. 1,250, it accounted
for less than 5% of the sherds, and in very late prehistoric increased to its
original frequency. A detailed modal analysis of this ceramic assemblage was
performed and is presented in Chapter IX of this repoxrt.
Other materials provided informaticxi about the Late Woodland/
Mississippian. Most features oontcdned a proportionately large quantity of
lithic tools and debitage. Hie majority were relatively small pits, possibly
suggestive of short-term, recurring occupation or use. Several of the pits
were stratified, suggesting reuse, and some had large amounts of secondarily
deposited charcoal. The lack of prepared areas, ceremonial features, burials,
or structural renains reflects a utilitarian and short-term use of the site.
Severed Kirk points, both in the g«ieral midden and in the features eue
indicative of an Early Archaic occvpation. In some cases, Feature 2, it
123
appeared that these a3:±i£acts were collected and reutilized by later groups.
Feature 28, with diagnostic artifact (a Sykes-White Springs point) and a
5, B.P. radiocartxsn age, identifies a small Middle Archaic ccnponent.
Late Archaic projectile point/knives fron the general midden and a small
pit. Feature 27, suggest a possible low intaisil^ Late Archaic ocnponent. The
large-shaped fire basin, Feature 30, and many of the steamed points associated
with the assemblage indicate an Alexander ocnponent. The fire basin which
became a trash pit begun in Alexander times, suggests sli^tly less transient
use of the site during this oocipation.
Sand-, limestone-, and bone-tempered ceramics in varying but small amounts
in the midden may have been frcni Early and/or Middle Woodland activity, but
could edso have been merely minority wares during the major period of
habitation, i.e. the Late Wocsdland/MLssissippian.
During the 500 years of the late prehistoric period the site was Weis still
tindergoing li^t, short-term, intermittent use. The site was an ipland bluff
canpsite overlocddng the Tcnbigbee floodpledn near a small stream v^iere wild
foods were harvested. It may have been a hunting station or wild plant
collecting locale visited sporadically. This site represents but a small
piec:e of the Late Woodland and Mississippian settlement pattern moscdc,
ooiplementing the arrangement of larger agricnoltural and ceremonial
settlement.
EXCAVATIONS AT THE MUD CREEK SITE (22lt622)
The Mud Creek site was located in the Tcmbigbee River floodplain in
northern Itawamba Co., Ms, approximately 11 km (6.9 mi) north of Fulton
(Figure 1) . The site was approximately 400 m (1,320 ft) south of a tributary
stream (Mud Creek) and the Tombi^pee River. Only C3a. 175 m (574 ft) north of
the site was Site 22It606.
The site was approximately 1 m (3.3 ft) above the general elevation of the
surrounding flocxSplain and roughly circular in outline (Figure 25) . The
surface scatter of artifacts covered an area roughly 85x75 m (23.4x20.6 ft) .
Prior to the investigations, the site had be^ cleared and farmed for
approximately 50 years, and the v^jper 15-20 cm (5. 9x7. 9 in) has been disturbed
by plowing. No potholes were observed; however, the surface probably had been
collected. Recent waterway oonstructicmi road for equipment had been placed on
the center of the site. The site was located during Phase I of this project.
After exploratory information doctmenbed it as a multi-ocnponent mound, it was
scheduled for testing.
FIELD METHXS
The initial reconnaissance fieldwork included a general surface
collecticn. Subsequently, two 1x1 m (3. 3x3. 3 ft) test units in the center of
the mound (Figure 25) were selected. Block A was excavated to 48 cm (1.6 ft)
below surface, but sterile sediment was not reached because heavy machinery
had adtered the stratigrai^y before excavation was catpleted; Block B wzis
excavated to 58 cm (1.9 ft) . Because of the presence of multi-components and
the possibility of a stratified site, the site was reccmmended for testing.
In pr^)aration for testing, the site vas mewed and disced in an area
approximately 88x84 m (290.4x277.2 ft). Testing was conducted in September
and October of and began with a controlled surface collection of 17% of
the site. There were 78 4x4 m (13.2x13.2 ft) \^ch were randomily selected.
One 4x4 m (13.2x13.2 ft) unit excavated was located in the center of the site
between and tangent to the originail 1x1 m (3. 3x3. 3 ft) units.
Figure 25 Topographic map and excavation pl^, 22It622.
CULTURAL REMAINS
Cultxiral renains recxTvered fzon the Mud Credc site ocmsist^ of oeraniic
aiid lithic material. Only 41 ceramics were recovered frcnt the site, which is
a particularly lew figure coqpared to other sites in the floodplain. 11:ie
temper types consist of fiber, sand, and grog. Fiber tenpering doninated the
ceramic assemblage (78.0%), folloMod by grog (14.6%), and sand (2.4%) (Table
39) . The snail nixiber of ceramics suggests that the Mud Creek site was only
briefly occi;pied during the post-Archaic. This situation contrasts sharply
with other tnv sites.
TABE£ 39
Oeramic fcequencies by temper, 22It622.
_ ra -
Temper
Grog
6
14.6
Sand
3
2.4
Fiber
32
78.0
Total
41
A total of 301 chipped stone tools were recovered
frem the Mud Creek site.
The most frequent tool type was "Other Ibiiface aixi Bifaoes" (48.8%) (Table
40). However, 97.9% of these
specimens vexB unidentifiable fragments of
chipped stone tools. Projectile point/knives iiade tp
cotprising 21.9% of all
chipped stone stools and 40.5%
of all identifiable chipped stone tools.
Preforms (11%), bifaces (8.6%)
, and scrapers (7.3%) were proportiOTately low.
but relatively even in amount.
Drilling tools (2.7%)
and cores (1.7%) were
the least abundant of the chipped stone tools.
TABUS 40
Chipped stone tool frequencies by type, 22It622.
Type
Projectile Point /Knives
Little Bear Creek
10
Ledbetter/Pickwick
3
Flint Creek
1
Wade
X
Baker's Creek
X.
Crawford Creek
i
Tcxnbi^3ee Stenmed
]
Residual Stemmed
3
Kirk
4
Greenbriar
2
Unidentifiable fragmentt^
39
Subtotal
66
21.9
Bi faces
26
8.6
cores
r
1.7
Scrapers
22
7.3
Preforms
33
11.0
Drills, Perforators, etc.
8
2.7
Other Uniface and Bifaoe Tools
141
48.8
Total
301
126
A total of 242 utilized flakes were in this assemblage. Ihe most frequent
was 0.5 in (1.3 on) (57.4%) follcMed by 0.25 in (.64 cm) size flalces (38.0%).
One inch (2.54 an) flakes (2.1%), chunks (2.1%), and prismatic blades (0.4%)
were low in frequency.
Debitage wcis abundant in this assemblage, totaling 9,030 flakes
(Table 41) . As in other assemblages, 0.25 in flakes were the most dominant
(82.4%) .
•aOE 41
Size
Frequency
1.0 inch
30
0.3
0.5 inch
1,556
17.2
0.25 inch
Total
7,444
9,030
82.4
Half-inch (1.3 an) flakes o^rised 17.2% of the debitage and only 0.3%
were one inch or larger. The distribution of the flakes sizes reflects stone
tool maintenance and production at this site.
The tool-to-debitage ratio is 1:30 which is low and likely reflects the
manufacture of chipped stone tools at this site during the Archaic stage. The
proportion of flaJce sizes is 1:52:248 vhich further sv:?:ports the stone tool
production activities.
Only 12 ground stone tools were recovered. While most (12) were fra^nents
of lanidentifiable tools, one hanmerstone, one steatite sherd, and one
sandstone concreticai were found.
ETEftlURES
Only two features were encountered in the testing investigation, i.e. two
pits located in Block C. Neither pit contained any diagnostic material,
although one did have several chipped stone tools.
STRflTIGRAPHY
The stratigraphy indicated that the site was formed by the fluvial
deposits. The landform was likely a parallel bar similar to the Poplar,
Walnut, and Hickory sites.
Six identifiable strata were recognized in Block C (Figure 26) . Five of
the six strata were continuous within Block C. Stratum 6 was an extremely .
compact sediment vhich was horizontally discontinuous. The upper three strata
(50 cm: 1.6 ft) contained the most cultural d^xjsits and were slightly
organically enriched, although much less so than in other flocx^lain mounds.
The loany cultural zone was riddled with krobovinas and root stains, and it
erfiibited much pedoturbation.
SUfABY
The Mud Creek site was located in the floot^lain of the Torbic^Dee River
Valley and was a small parallel beir formed during flooding episodes. The
cultural material recovered indicated that the site had been occi;pied
intermittently from the Early Archciic to Woodland stages, and the materials
had been mixed, probably from the intense bioturbation.
127
Evidence recovered stiggest that this site was never oocvjpied intensely.
The site probably served as a location for short-term activities. The nunber
of diagnostic projectile point/knives and ceramics recovered suggests that
this site was used most frequently during the Late Archcdc and middle Gulf
F0oaation2d..
Although this mound was smaller and lower than the others investigated, it
is unusual that it appeared to have had negligible Woodland occvjqpation. Only
nine sherds of the Woodland stage were recovered. Ccnpared to the thousands
present at other sites, this is remarkable. The absence of shell tempering is
also different.
The shallow and mixed midder and low frequency of features mitigated
agadnst further investigation.
EXCAVKTKXIS AT THE SMILAX SITE (22M)
The Smilax site was located in Monroe Co., Ms, approximately 3.2 km
(2.0 mi) northeast of Aberdeen and 15.6 km (9.7 mi) south of Fulton (Figure
1) . This was one of two sites investigated in this project in Monroe County.
The site was cai a low rise in a plowed field on the edge of the first terrace
400 m (1,320 ft) east of the Tantoi^3ee River channel. The edge of the terrace
on which the site wais located had been cut off from the remainder by a
tributary stream, and the site landform was a low outlier surrounded by
wetlands. Vine Creek entered the floodplain and joined the Tcmbigbee ca.
1,000 m (32.8 ft) to the north of the site. The site area was approximately
90x50 m (297x165 ft) on a small rise of approximately 60 cm (2 ft) on the edge
of the terrace outlier overloolcing the floodplain. An embankment and ditch of
an abandoned 19th-century railroad were present cxn the eastern edge of the
site (Figure 27) . The western edge of the site was the terrace edge of the
floo^lain which dropped 1-1.5 m (3.3-5 ft). The southern boundary was
defined by the decrease in surface artifact density. T\iro other sites, 22Md676
and 22Mo677, occvpjy low knolls on the edge of the landform adjacent to the
flooc^lain about 100 m (330 ft) and 200 m (660 ft) to the west-northwest and
west-southwest, respectively, of 22Mo675, and cultural material was present in
low quantities on the surface between the concentrations of the three sites.
The soil survey indicated that this site was cleared less then 20 years
ago (uses ) . Prior to this the site area had been covered by forest.
FIELD METHODS
The Smilax site was originally located in Blakeman's (:76) survey of
the Canal Section of the waterway. It was recatmended for additional
investigation beised on the suspected presence of a Late Archaic cenponent
(Blakesnan :74) . The paucity of information from this cultural period was
poor, and the site was tested in Pliase I of this project to determine its
information potenticil.
Testing was ccnducted in Phase I of this project in January and February
of and included a controlled surface collection, mechanical stripping,
and test units. The controlled surface collection was a 22% stratified randem
sanple of 4x4 m (13.2x13.2 ft) collection units. Six 2x2 m (3. 3x3. 3 ft) test
pits were excavated (Figure 27) . Five were placed randomly, and one test pit
(Test Pit 6) was judgmentally placed. Only Test Pits 4 and 6 were excavated
into the sterile subsoil. The others were excavated only through the plow
zone to cbtcdn a sample and check for features.
Itiree judgmentally placed 24x2 m (79.2x6.6 ft) mechanically stripped
tTcmsects were excavated with a small tractor and bcx-scraper (Figure 27) .
All potential features were examined and investigated.
OJLTURftL REMAINS
Only 78 sherds were recovered from the Staiilax site. The assemblage was
dominated by sand-tempered sherds (73.1%), followed by 23.1% grog, with only
one specimen each of shell/grog-, limestone-, and fiber-tempered sherds (Table
42) . All sherds from Units 1-5 were recovered from the plow zone (0-30 cm or
0-11.8 in) below the surface. In Test Unit 6 ceramics ed.so were recovered
beneath the plow zone.
TABUS 42
Oeranic frequencies by temper, 22Kd67S.
Temper
Frequency
Percentage
Shell-Gtog
1
1.3
Grog
18
23.1
Limestone
1
1.3
Sand
57
73.1
Fiber
1
1.3
Total
78
A total of 24 chij^jed stcxie bools were recovered (Table 43) . Hie most
frequent category was projectile point/knives (37.5%) , including one Late
Woodland/Mississippian Triangular, one Geury, two Little Bear Credc, and one
Flint Creelc. The next most frequent tool type encountered was "Other Iftuface
and Biface" tools (25%) . However, as in all other assenblages, this was made
up primarily of unidentifiable fragnents (83.3%) . The remaining chipped stone
tools incliKie drilling tools (12.5%), scrapers (8.4%), preforms (8.4%),
bifaces (4.2%), amd cores (4.2%).
TMBU 43
Chipped stone tool frequencies by type, 22Wd675. _
Percentage
Projectile Point/Knives
Flint Creek
1
Gary
1
Late Woodland/Mississippian Triangular
1
Little Beair Creek
2
Distal fragnent
2
Medial fragment
1
Proximal fragnent
1
Subtotal
9
37.5
Bifaces
1
4.2
Preforms
2
8.4
Scrapers
O
4m
8.4
Drills, Perforators, etc.
3
12.5
Cores
1
4.2
Other Uniface and Biface Tools
6
25.0
Total
24
A total of 31 utilized flakes were recovered. Of these 15 (48.4%) were
0.5 inch (1.3 cm) flakes and 18 (58.1%) were 0.25 inch (.64 cm) in size. The
najority (93.5%) were made frcm Camden chert.
A total of 510 flakes of debitage were recovered (Table 44) . Quarter inch
flakes were most frequent (88.8%), v4iile 0.5 inch (1.3 cm) flakes corposed
only 11.2% of the debitage. No flakes greater than 0.5 inch (1.3 cm) were
recovered, viiich likely reflects a lack of initial cotble reduction at the
site. Camden cherts made up 93.3% of the d^itage, and the remaining 7% of
the flalces were cotposed of heated emd unheated Tuscaloosa gravel, blue-green
Bangor, Fort Payne, fossiliferous Fort Payne and Pidodck cherts, quartzite,
ferruginous sandstone, and unidentified raw materials. Ground stone tools
were few and limited to one unidentifiable ground stone tool and one ground
stone flake.
•nOE 44
Size
Frequency
0.5 inch
57
11.2
0.25 inch
Total
453
510
88.8
The tool-to-debitage ratio is 1:22, vAiich is relatively high. However,
the flake size grades contained nor^ greater than 0.5 inch (1 .3 cm) and a
ratio of 1:8 for the 0.5-0.25 inch (1.3-0.64 cm) sizes.
No ground stone artifacts were recovered frcm this site, ltds is the only
site in the project which produced no such tools.
Historic material incltided two brass shotgun shell bases, one fragment of
slag/cinder, one fragment of xmidentifiable clear glass, and one aluminum can.
All were contained in the plow zone.
FEATURES
No cultural features were documented in the testing of the Stnilax site.
All suspected features identified in Figure 27 were natxiral in origin.
STRATIGRAPHY
Three major strata were documented at 22Mo675 (Figure 28) . The strata
consisted primarily of a plcw zone 10-20 cm (3. 9-7. 8 in) thick which was part
of a moderately dark organically enriched stratum 5-15 cm (2-5.9 in) thick.
This had mostly been disturbed by plowing or cultivation. The basal stratum a
ccmpact yellow-brown silty loam of irKleterminate depth. It was culturally
sterile .
SIMMARY
The Smilax site was located on a terrace remnant, surrounded by the wet
floodplain of the Tombigbee River, containing many other sites. Testing
investigations documented that the site had been occupied during the Historic
(Euro-American) , Late Woodland, Middle Wbodland/Late Gulf Formational, and
Late Gulf Formational periods. The position of all diagnostic artifacts
recovered frcm test imits located hcwever were confined to the plcw zone. It
appears that cultivation appeirently had destroyed the vertical and horizontal
integrity of the cultural ccmponents.
132
One pattern was seen in the lithic dehitage vdiich was ocsnoentrated in Test
Pits 5 and 1. Diagnostic artifacts recovered fron these imits were confined
to the plow zone and range from Late Archaic to Late Wcxxiland. Chipped stone
tools were concentrated in Test Pit 6. All diagnostics, however, were
confined to the plow zcxie indicating that this uncultivated wood^ area.
Of the diagnostic material recovered, more was from the Late Archaic
through Middle Woodland periods, suggesting more use during this time.
However, this site was never intensely used during any period and can be
considered to have been the scene of very limited activities.
EXCAVmONS AT THE DOGWOCO MOUND SITE (22Jto531)
The Dogwood Mound was located in Mcauroe Co., Ms, approximately 8.3 km {5.4
mi) northeast of Aberdeen (Figure 1) . The site lies approximately 520 m
(1,706 ft) east of the present channel of the Tcmbigbee River and on the edge
of the first terrace. 'Hie Dogwood Mound was a conical e2urthwork mezisuring
approximately 17 m (55.8 ft) in diameter and 1.85 m (6.1 ft) in hei^t
(Figure 29) . A pothole was in the center of the mound approximately 3.75x4 m
(12x13.1 ft) in size. The mound was situated in an old field, which had been
planted in pines. Second growth oak and hiclcory grew on the earthwork and
within about 5 m (16.5 ft) of its perimeter. The understory consisted of
dogwxd, briars, and cliirbing vines. i
The terrace edge was approximately 40 m (132 ft) west of the mound, and it
marked the boundary of the floodplain vMch svpports a rich hardwcxxl forest.
A relic channel of the Tanbi^aee lay near the site in the floo^lain which
ocjntained a small stream. The terrace on vdiich the Dogwood Mound was situated
was broad and level. ^
FIELD METHCOS
The site was first located and recorded during the first survey of the
waterway (Lewis and Caldwell :18,111) and was investigated further by
Adkinson (:114) and Blalceroan (:75) vdx) identified it as a probable
Miller I period burial mound. I
Blakeroan (:75) hypothesized that the earthwork was a "Miller I Period
burial mound" and noted that "as such this site joins a relatively elite group
of sites in northeast Mississippi including the Pharr Mounds and Bynum
Mounds." Blakeman's reocritendations were sipported Atkinson (:114)
v^n he expressed the concern due to the rsurity of such mounds in the Upper
Tcmbigbee Valley and the threat of vandalism. I
The site was subsequently tested in Phase I of this project from February
to March of to assess the cultural affiliation of the mcjund and to
evaluate its integrity. A large recent pothole, dug by relic collectors,
intruded into tlie top of the mound (Figure 29) .
Testing investigations included a hand-excavated trench into the mound and
several deep auger borings. Two sides of the earthwork were cross-sectioned I
with a series of 1x2 m (3. 3x6. 6 ft) units separated by 10 on (3.9 in) balks
(Figure 29) . All units excavated in the mound, except one, were dug to the
subsoil contact. Soil texture, color, and structure were used to separate the
construction fill from the sub-mcund soil. Soil characteristics and cultural
phencmena were noted or mapped during level excavation.
134
Figure 29 Topographic map and excavation plan, 22Mo531.
CULTORAL RHftlNS
Prehistoric cultural debris was found scattered throughout the fill of the
earthwork. Cer^Bnics and lithics indicate that the earth used to construct the
mound was removed from an area which contained cultural material. Ihe silt
loam character of the fill suggests that the dirt was quarried neeu±y. No
borrow area was observed, however. Historic land vise and periodic flooding of
the locale apparently obliterated any traces of such a borrow pit.
Ceramics provided the most definitive clue to the cultural affiliation of
the mound. Ihe entire assemblage consisted of Middle Woodland limestone-
(3.3%) or sand-teqpered (96.7%) specimens (Table 45). The presence of Furr's
Cord Marked, Saltillo Fabric Marked, Pleiin sand-tenpered sherds, cuid Mulberry
Cre^ Plain indicated a Middle to Late Miller I context (ca. A.D. 1 - 300)
(Jenkins :257-259). Further, the number of Furr's Cord-Marked sherds (9)
in comparison to those of Saltillo Fabric Marked (33) hints that the fanner
constitutes a major type in spite of being outnvmbered by the latter. If that
is the case, the ceramics prc±»ably represent a Late Miller I occi:qpation, and
the mound was probably constructed at ca. A.D. 200 - 300 (Jenkins
:258-259) .
TAHUB 45
Ceramic frequencies by taiper, 22ito531.
lype
Limestone
10
3.3
Sand
296
96.7
Total
306
A total of 37 chipped stone implements were irecovered (Table 46) .
Identifiable projectile point /knives included only one Little Bear Creek and
two Residual Stemmed types. The most frequent category was "Other Uniface and
Biface" tools, but 88.9% were fragments of unidentified tools. Atypioally,
cores followed projectile point/knives in frequaicy (13.5%) followed 1^
scrapers (10.8%), preforms (5.4%), and bifaoes (2.7%). Utilized flakes were
the most frequent tool type (rp=119) , and 51.3% were 0.5 inch (1.3 an) in size.
Six ground stone items, including two hanmerstones , a muller, one piece of
ground hemnatite, and two unidentified fragtnents, were found.
19IBLE 46
Qi-inpwH stone tool frequencies by type, 22Mo531. _
Type _ Frequency _ Percentage
Projectile Point/Knives
Little Bear Creek
1
Residual Stemmed
2
Distal fragment
2
Medial fragment
2
Subtotal
7
18.9
Cores
5
13.5
Preforms
2
5.4
Bifaces
1
2.7
Scrapers
4
10.8
Other Uniface and Bifaoe Tools
18
48.6
Total
37
136
Debitage flakes were quite minerous (n=3,086) . As seen in Table 47, 83.7%
were 0.25 inch (.64 cm) in size. Hadf inch (1.3 can) cxnprised only 16.2%, and
flakes greater than one inch (2.54 can) acxxunted for only (0.1%) of the total.
The primary raw material was Canden cdiert (94.1%) with Fort Payne chert inaicing
up most of the remainder of the raw materials.
TMUB 47
Size-grade frequencies of debitage, .
Type^
Frecjuency
1.0 inch
4
0.1
0.5 inch
499
16.2
0.25 inch
2,583
83.7
Total
3,086
The tool-to-debitage ratio in the fill of this mound is 1:87 and is
relatively low. The proportion of the flake sizes is 1:124:645, the lowest of
edl sites investigated.
Historic artifacts included 15 6d cxnmc^ cut nails and 19 seed beads. All
but three of these items were recovered in asscxriation with a burial in the
mound. The remaining specimens were recovered from the pothole in the mound.
Testing demonstrated that the mound had been utilized as an historic
cemetery, probably during the 19th c:entury. Ex<3avaticr\ dcxnxnented at least
two interments. The scxtth edge of the grave shaft of the southern historic
burial was clearly discernible in the profile and extended from the root mat
within a few centimeters of the surface to a cSepth of aJxwt 6 ft (1.83 m) .
The recent pothole destrcjyed a major portion of the center of the earthwork.
It was originally 3.5x4 m (11.4x13.1 ft) wide at the top and tapered to 2x1.5
m (6.6x5 ft) at the botton and was 1.6 ro (5.2 ft) deep. This pothole
destroyed at least cue historic burial as well as the northern margin of the
grave of Burial 1. The impact of the vandal's activities on any prehistoric
interments that may have lain at the base center of the earthwork could not be
determined.
FEATURES
Three features were docnmented during the Dogwood Mound testing project.
Feature 1 appeared to be a dark brcwn Icsad of midden used in construction. It
was irregular in plan and profile, 90 cm (2.9 ft) belcw the surface, 51x43 cm
(1.7x1. 4 ft) wide, and 25 cm (9.8 in) thick. Artifacts in it included
aboriginal pottery and a few grams of introciuced roct.
Feature 2 was 150 cm (4.9 ft) below the surfeKO and was a deposit of red
cx:her. It was oval in shape with slightly tapering sicSes with a beisin-shaped
cross-section measuring 14.5x17.5x12 cm (5. 7x6. 9x4. 7 in) thicdc. This feature
was probably an intentional deposit in the mound fill.
One burial was also Icxated in the mcund. It was in the coenter of the
mound 179 cm (5.9 ft) belcw the surface. Only the western end of a pcxjrly
preserved coffin and the upper remains of a Caucasian female, whose age at
death wets estimated as 50± years, were exposed. The major porticsn of the
interment extended eastward into an unexc^vated section of the site. The
burial cx>ffin was manufacrtured with exit cxxtmon nedls. The use of these nails
suggests that the cxiffin was manufactured from one-inch (2.54 an) planks. The
vertical orientation of nails alcang the southern and western sec^u-cans of the
exaffin suggests that the sieJes were set on the botton and nailed from beneath.
The width of the exaffin probably ranged from 30.5-40.6 an (12.16 m) ; the
lieight and depth of the exantainer was prcbably more than 30.5 cm (1 ft) .
137
A series of eight green seed beads, ca. 2 nm in diameter, and 11 black
seed beads, ca. 3 inn diameter, v«ere found in the thoracic area of the burial.
Diese beads lay in a position v^ch suggested a single strand arrangement
vAiich V4as perhaps appliqued to a garment. No other grave goods were
recovered.
The cranium and mandible of the interment were examined in the field by
the physical anthropological consultant. His observations documented that the
individual was of Caucasian extraction ani a female. The age estimation was
based on the thinness of the parietals, maturity of the cranial sutures
(closed on both the interior and exterior surfaces) , and teeth wear. The
condition of the teeth would argue for a lower age, but balanced with the
suture closings and parietal thinning suggested the older age.
All teeth present were in relatively good condition with no developmental
defects or evidence of jAiysical traima. One cavil^ was found in the i^per
ri^t second molar and no developmental defects, or evidence of p4^sic2d
trauma. The first molars on both the left and right sides of the mandible and
maxilla were absent. All first molars were lost pre-mortem as indicated by
the alveolar resorption and remodeling vbich had occurred.
A circular penetration of approximately 5.6 nm of the left peurietal
immediately posterior bo the coronal suture was interesting emd possibly
significant. The penetration was quite close to the bregma and resenbled a
wound, possibly caused a .22 caliber bullet. It was inpossible to
determine vbether this wound v^s the prcscimate cause of death. Without dcubt
this particulcu: wound did not result in the iimediate death of the individual.
Ihe inner table of the paurietal was not shattered. Examination of the
interior of the cranium showed that repair and rebuilding of the inner table
had proceeded for several months prior to death. Ihe surfaces of the inner
cranium did not give an indication of markings left by a bullet either as an
exit wound or as ricochet trails. Although the area surrounding the cranium
was searched, no bullet was found. Fran the apparent sngle of the wound (if
indeed it was a bullet) an exit through the foramen magnum is not
inconceivable .
Other pathologies included considerable infectious disturbance occurring
in the mastoids. Even with some remodeling the mcistoids displayed mcuked
indications of mastoiditis. The infection of the mastoids did not ^pear to
have been active at the time of death. The stperior surface of both orbits
displayed sane porosities of an appearance similar to cribra orbitalia. The
left orbit was the more severely affected. Althou^ the causes of cribra
orbitalia are not cleeurly established there is seme evidence suggesting that
depletion and/or insufficiency of iron avadlable to the organism may produce
lesions of this type. Such iron deficient associatiois with cribra orbitalia
have been noted in tropical eureas where parasitic infection is quite corantiOTi.
NO hyperostosis spongiosa orbitae was noted in either facial or cranied. bones.
Both the atlas and axis were cotpletely normal in appearance, shewing no
pathologies or degenerative changes.
To sunmarize, Burial 1 was a Caucasian female approximately 50 yesurs old
vbo was interred, presumably, in an exterxied position in a burial container
assumed to be a simple rectangular coffin. A mid-nineteenth century date was
estinated for the burial. Further, the association of the historic Dogweod
Mound grave plot and an historic farmstead site within 300 m (990 ft) conforms
with an expected settlement pattern of rural residence units.
Although Burial 1 was the only interment formally defined, crani2d remeiins
of a second individual and 6d nails were recovered from the fill of the
vandal's pit just to the north of Feature 3. This indicates that the Dogwexsd
Mound contained at least two historic interments.
All artifacts except the beads and nails associated with Feature 3,
Burial 1 were reoovezed frcxn the mound fill and yihat may be a buried A horizon
at the base an the mound on the original land surface. No distribution
patterns within the mound were detected. Material reoovezed frcm the
off-iiDund units indicate that prehistoric occvpation debris was confined to
the plow zone.
STRATICS^APHY
The silt-loam soil in the inroediate area was documented in the two test
pits excavated approximately 5 m (16.4 ft) fran the south and west edges of
the mcund. This soil was encountered beneath the mound and served to identify
the base of the earthwor)c.
Five stratigraphic zones were defined in the Dogwood Mound (Figure 30) .
■Hie stratigr6phy of the mound fill indicated that the earthwork probably
represented a single constructicxi episode. Stratum I and II were e^parently
the result of weathering of the mound's vpper surface since construction.
Stratum III accounted for the major volume of the earthwork and probably
represented a single depositional {base. Stratum IV appeared to have been the
buried A liorizon, or a redeposited A horizon. It is postulated that the
absence of this stratum in the western section of the esurthwork can be
accounted for by its lack of distinguishing characteristics in this section of
the site or by its prehistoric renovaUL.
StMIARY
Test excavations at Dogwood Mound revealed no aboriginal interments or
cremations. Although inhumations were expec±ed at the base center of the
mound, testing revealed only two historic intrusions in this location. The
U.S. Anty Corps of Engineers ordered excavaticxi to cease after t)ie ^ situ
Euro-American burial that lay at the approximate center of the earthwork was
discovered. No evidence of prehistoric ljurials was encountered 1 m (3.3 m)
west or 2 m (6.6 m) to the south.
Despite the lack of aboriginal interments, the material contents of the
fill indicate that the earthwork was constructed during the Middle Wbodland
period. A ceramic complex, containing Saltillo Fabric Marked, BHirr's Cord
Marked, Plain sand-tenpered sherds, and Mulberry Creek Plain, argues that the
mound was probably constructed during the latter part of the Miller I pbcise or
ca. A.D. 1-300 and quite possibly during the Late Miller I subphase (ca.
A.D. 200-300).
The stratigrajby of the earthwork sv»ggested that the mound constsnjcted
as a single unit. Differentiation of stratigrajbic zones was considered the
result of post depositional weathering emd possible burial of an A horizOTi. A
historic grave, inferred to date to the 19th century, crosscut or intruded
into all but the base strata of the site. Although it might be suggested that
the earthwork wzis an artifact of the historic period for use cis a cemetery
plot, the crosscutting and truncation of stratigraphic zones within the
earthwork negate this possibility. Based c»n the aveiilable evidence, eObeit
circunstantial, there is little doubt that the mound was a prehistoric
feature, most probably associated with mortuary practices.
The use of aboriginal earthworlcs far Euro-flmerican cemeteries was in
Iceeping with the historic settlement pattern practice of selecting prominent
topographic features for the location of graveyards. The site of the Dogwcxxl
Mound historic grave plot and the location of a farmstead site approximately
139
SOUTH TRENCH
Easi Prollle
;MOdefn (ntrusion
230 m (759 ft) to the east, however, reflec± a locational pattern ocnmon to
rural households of the Southern Piecknont during the 18th and 19th centuries.
The Euro-American interment at the Dogvrood Mound requires one of tMO
actions. The site can either be pi^served and protected in perpetuity, or the
historic interment (s) can be moved and rdxoried vhich vpould necessitate
mitigating the impact of this action on the prdiistoric resources of the site.
SCMIRRy OF EXQAVRXICXE
Of the 11 sites investigated in this project, four were excavated only,
four were tested and excavated, and three were tested only. The sites were
located on and adjacent to the flooc^lain of the Upper Tcnbi^see Valley. In
situ stratified deposits of the Archaic stage were encountered and excavated
at six of the sites. In addition, one Late Gulf Pormational and one Late
Moodland/Mississippian component was excavated. Four of the sites (Poplar,
Walnut, Ilex, and Hickory) were midden mounds whicdi were islands of
well-dradned loam in the wet flooc^lain. Of these the walnut and Poplair sites
were the most intensely occtpied and oantained the most material and features.
The Beech and Oak sites (22It623 and 22It624) were in the flooc^ladn, but not
as intensely ooopied as the others.
The information recovered from these sites will be more integrated in
Chapter X of this report.
141
CHAPTER VI SOII£ AND GBGMORPHQlJOGY STUDIES
INTRCDUCnCN
Although archaeological studies were an ijnportant part of the
investigations, other disciplines, especially soils science, produced a great
deal of useful information. In this chapter, the soil studies conducted in
this project, especially soil chemistry and mor{4iology, are r^xirted and
interpreted. These soil studies ha\^ contributed markklly to the
understanding of midden mound morphogenesis, past landscepe stability,
intensify of site use by hunan groups, and consequent effects on the soil
profile. Data acquired were largely original, and they supported more
realistic interpretations about both the nature of past htnan behavior at
these sites, and they contributed to an understanding of extant landforms.
The soil studies conducted as part of these investigatixxis Identified and
described paleosols, anthrosols, and soil development processes at five
flooc^lain and one tpland site. Since much of the soil science data and
discussion in this chcpter are probably outside the ken of neny
archaeologists, the editor has included a section at the end of this chapter
vhich describes and interprets the soil studies especially for them.
The goals and objectives of the soil and gecnorphological studies
performed in this project were as follows:
1. To investigate and describe the pedogenic/geanorphic relationship of
r^resentative sites in their natural setting.
2. To characterize the physical, morphological, chemical, and mineralogical
soil parameters of selected representative sites.
3. To characterize the organic-stained epipedons of typical sites.
4. To describe and quantify the paleosols at selected sites.
MBIH3DOLC3GY
Preliminary soil examinations of sites and adjacent areas were made using
positioned transects and soil auger observatiOTS to depths of 2 m (6.6 ft) .
The morpjxjlogicsd examinations included gecmorphic position, soil color,
texture, structure, consistency, horizonation, boundaries, coarse fragnent
contents, and hydrological characteristics.
Detciiled soil examinations of the sites were made in excavation units and
stratigraphic backhoe trenches. The soils were described and sanpled using
the standcurd method (USDA Soil Survey Staff ) . Sanples were sealed in
plastic bags for subsequent analyses.
Soil samples were odr dried, crushed with a wooden rolling pin under
gentle pressure, and sieved throuc^ a 2 nm sieve for laboratory analyses.
Particle size distribution was determined by the hydrometer method and sieving
(Day ) . Soil organic carbon was determined by wet cembustion (Allison
) . Extractable acidity was determined by the barium
chloride-triethanolamine method (Peech ) . Exchangeable eduminum was
determined by potassium chloride extraction following the procedure of Yuan
() . Free iron oxides were determined by sodium-dithianite extraction and
potassium dichromate titration (USDA Soil ConservatiCMi Service ) .
Exchangeable cations were extracted with neutral IN NH4QAC and determined by
143
atonic adsorption spectxxjphotonetiy. Soil was me2isured in water using a
1:1 soil-to-liquid ratio.
Nitrogen was determined by the Kjeldahl method (Brenmer ) using soil
ground to pass throu^ a 60-niesh sieve. Citric acid-soluble phosphorus was
extracted with 1% citric acid after shaking thirty minutes and determined by
the molybdovanate method (AOAC ) using a Baiisch and Lotb Spectronic 21
spectrophotoneter. Phosphorus fractionation on selected sanples was
determined by modification of the Chang and Jackson procedure (Peterson and
Corey ; Meixner and Singer ) . Determination of hutnic/fulvic acid
ocnpounds of selected samples was determined by color ratio at 400 and 600 nm
using a Spectronic 21 spectrophotometer (Tan and Giddens ) .
Clay fractions were s^>arated by centrifugal sedimentation . They were
anedyzed by X-ray diffraction (Jackson ) with a Norelco Geiger counter
spectrophotoneter using Cu Ka radiation and a nickel filter. Mineral type and
content were estimated from the basad spacings and X-ray peak intensity.
Differential thermal analysis wais conducted on selected samples using
magnesium-saturated sanples equilibrated to 56% relative hvmidity and a
Deltatherm DTA instrument. Microscopic examinations were made of selected
soil peds using reflected light microscxpy to 150 power.
Soil samples were described and collected from profiles considered
representative of the site after chservation of the site and the adjacent
off -site soilscape. GeomrEhic surfaces were studied by field transects
between sites using a soil auger for subsurface investigations.
The informatiCTi recovered from individual site investigaticxis will be
presented here from five selected-as-representative sites. The detailed
information will be presented for each site with an integrative summary
following.
THE VJALNOT SITE (22It539)
The site was in the eastern part of the Taibi^aee River floo^lain about
750 m (2,475 ft) west of the eastern valley wall, and was a prominent
topograptiic feature elevated approxinately 1-1.5 m (3.3-5 ft) above the
surrounding flooct>lain (Figures 1 and 6) . The site had slopes of 2-5% in
contrast to slopes of 0-2% in the adjacent floodplain. Lower-laying parts of
the site were subject to winter and spring flooding, vhich reguleurly scours
and fills the floodpledn creating microrelief.
The elevation appeared to be a natural topographic feature resulting from
fluvial deposition. Small slou^s pairtially surrounded the site, and their
silty bottom sediments indicate an aggrading status. The coarse texture of
the Seindy loam soils in the site suggest hi^er energy depositionail events.
The steep valley walls bounding the flooe^lain are composed of mature,
well-developed soils with thin ochric epipedons and well-develcped illuviated
argillic horizons (Bt) and eluviated E horizons. Sbiithdade soils (Figure 31)
dominate the eastern valley wall. These soils are deep, well drained, and
permeable, with red subsoils. They formed in thick beds of loamy materials cn
side slopes ranging to 40%. The argillic horizons have subangular blocky
structure and oriented clay skins on ped faces. Soils of the western vedley
wall are less steep, and the Ora and Savannah soils contain dense, firm
fragipan horizons in the argillic horizons. The Mathison soils of the western
valley wall have relatively high silt contents. The upland soils are very
strongly acid, highly weathered siliceous Ultisols (Table 48) with lew base
saturation levels.
144
SOIL LEGEND
SYMBOL NAME
1 Kirkvilie^antachie association
2 Mathiston silt loam
3 Ora fine sandy loam. 2 to 5 percent slopes, eroded
4 Ora fine sandy loam, S to 12 percent slopes, eroded
5 Savannan loam. 2 lo 5 percent slopes
6 Smithdale line sandy loam. 5 to 8 percent slopes, eroded
7 Smithdale fine sandy loam. 8 to 1 7 percent slopes
6 Smithdale association, hilly
9 Savannah loam. 0 lo 2 percent slopes
Figure 31 Soils in the vicinity of 22It539.
I
Soil Series
Kirkville
Mantachie
Mathiston
Ora
Savannah
Staiithdale
Cleussification _
coarse-loany, siliceous, thermic Fluvaquentic Dystrochr^xts
fine-loany, siliceous, acid, thermic Aerie Fluvaquents
fine-silty, siliceous, acid, thermic Flavaquents
fine-loany, siliceous, thermic Typic Fraguidults
fine-loany, siliceous, thermic Typic Fraguidults
fine-loany, siliceous, thermic Typic Pedeudults _
Kirkville and Mantachie soils catiprise the floodpleiin bounding the site
(Figure 31) . These soils are Dystrochrepts and Flvtvaquents vdth minimal soil
development (Table 48) . They typically had brown and yellowish brown surfaces
and gray or li^t gray subsxarface horizons (Table 49) . Textures ranged from
sandy loam to silty clay loam with occasional loamy sand, vAiich reflects the
textural stratification. The floodpleiin soils esdiibited little profile
development and had seme cambic Bw horizcxis. They were strongly acid.
Ihe culturally altered soils of the site developed in loany, fluvial,
siliceous sediments. Ihese soils were readily distinguislied by very thick,
humus-rich, dark reddish brown epipedons (surface horizons) , vdiich were due to
prolonged cultural activity and habitation. The past occi^xaticHi of the site
has drasticadly altered normal pedogenic features of color, structure,
consistency, horizonation, organic matter content, and certain chemiced.
parameters. The soil comprising the site differed greatly from adjacent
floodpladn soils and was readily distinguished.
Profuse populations of earthworms, crawfish, rodents, and other diverse
microfauna and microflora thrived in the organic-rich mound, vdiich was
elevated above the adjacent flooi^lain cind seasonal wetness. Pedoturbation
has tended to mix the v:pper meter (3.3 ft) of soil and affected normal
pedogenic exqxression. Horizonation tended to be marked ty intense
dark-oolored humic staining of the skeletal matrix.
TKBUE 49
Mansell color of selected horizons of representative soils in the
floodplain adjaoent to Site 22It539.
Sample
Depth (cm)
Munsell Color (moist)
Terrace east of the site above
15
-
30
Yellowish brewn (10YR5/4)
the floodplain
30
—
83
Brownish yellow (10YR8/8)
Middle of the floodplain
0
—
30
Brown (10YR5/3)
east of the site
50
-
75
Gray (10YR8/1)
75
—
105
Gray (10YR8/1)
Floodplain 75 m
30
—
60
Li^t brownish gray (10YR8/2)
east of the site
75
—
100
Light gray (10YR7/2)
Floodplain 20 m
25
—
50
Brown (10YR5/3)
east of the site
85
—
125
Gray (10YR6/1)
Floodplain 75 m
5
-
30
Dark gray (10YR4/1)
west of the site
62
-
88
Grayish brown (10YR5/2)
146
TMU 49
Mansell cxilar of selected horisons of representative soils in liie
floodplain adjacent to Site 221t539 (contiBiMd) . _
Sanple
Depth (on)
Mansell Color (moist)
Floot^lain 120 m
15 -
37
Dark bixwn (10YR4/3)
west of the site
40 -
50
Dark yellowish brown (10YR4/4)
Floodplain 40 m
25 -
50
Gray (10yR5/l)
south of the site
55 -
85
Dark gray (10YR4/1)
90 -
125
Gray (10YR5/1)
Flooc^lain 100 m
5 -
37
Grayish brown (10yR5/2)
south of the site
100 -
125
Gray (10YR5/1)
Floo(^lain 75 m
15 -
37
Gray (10YR8/1)
north of the site
87 -
112
Light gray (10YR7/1)
raysicaL desgription
Hie nound soil was dark reddish bzown and reddish brcwn with Mansell hues
of SYR in the upper 1.8 m (5.9 ft) (Table 50), idiich differs markedly firm the
adjacent floodplain soil vMch have hues of lOYR (Table 51) . Hie site
epipedon had a Mansell color value that shifted one unit with wetting and
dorying. The dcurk reddish brown epipedon had a distinct "greasy" feel when
rubbed between the fin^rs. Individual quartz grains had a continuous coating
of hunic stain. The thick, dark-colored epipedon graded into bric^ter colored
subsoil materials at depths below 1.3-1. 8 ro (4. 3-5. 9 ft). Hie subsoil had
dominant colors in the lOYR hue. Humic staining conucnly extended into the
upper part of the bri^ter colored subsoil and coated vertical surfaces of ped
faces. Topically » below the dark-colored ^ipedon and ininediately above the
bri^ter colored paleosol, the humic materiads formed distinct horizontal
bands or lamellae, vhere vertical water flow had been i?etarded by the less
permeable underlying paleosol.
TMU 50
Padon deacription of repgeaeptative pcofile, 22It539. _
Depth Itescription _
(on) (moist colors)
0-15 Dark reddish brown (5YR3/3) sandy loam; moderate fine and medium
granular structure; slightly firm in pleu», veiy friable vhen
disturbed; many fine and median roots; few entail black (10YR2/0)
charcoal fragments; greasy vhen rthbed; medium acid; gradual wavy
boundajY-
15-37 Dark reddish brown (5YR3/3) sandy loam; moderate fine granular
stonicture; friable; many fine and oonitpn medium oxaots; few small
charcoal fra^nents; nuneinus krotoviovi etnd worm casts; ccnmon small
and median gray (10YR3/1) and dark gray {10YR4/1) potaherd; greetsy
when nafched; raedioim acid; gradual wavy boundary.
37-60 Dark reddish brown (5YEt3/3) loam; weak fine granxolar structure;
friable vhen disturbed; ccmmon fine aixi few medium roots; few small
charcoal fragments; numerous krotovina and vrorm casts; greetsy vhen
rubbed; strongly acid; clear anooth boundary.
file, 22It539 (oontinaed)
!an) (moist colors)
60-100 Dark reddish brcwn (5YR3/2) sanc^ loam with few medium fednt very
dark brown (10YR2/2) mottles; weak fine granular structure; friable
vdien disturbed; ooccisional latninae of strong brown (7.5YR5/8) loam
in lower part of horizon; occasioneil mottled dudcy red (2.5YR3/2) ,
reddish brown (2.5YR4/4), and yellowish red (5YR5/8) "fired
aggregates"; ccmtion dicunroal fragments; few fine roots; sand stripping
evident on ped faces; numerous krotovina and worm cas^ts; medium cicid;
gradual wavy boundary.
100-150 Dark reddidi brown (5YR2.5/2) sarxfy loam with few pockets of strong
brown (7.5YR5/8) loatry sand; weak fine granular stnacture; friable
vrfien disturbed; many charcoal fragnents and few "fired aggregates";
few black concretions in lower part of horizon; few fine roots;
strongly acid; clear smooth boundary.
150-180 Dark reddish brown (5YR3/2) and strong brown (7.5YR5/6) san^ loam;
weak fine granular structure; friable disturbed; few charcoal
fragments; sand stripping on verticad ped faces; few "fired
aggregates"; strongly acid; gradual wavy boundary.
180-195 Dark yellowish brown (10YR4/4) sancfy loam with contnon medium strong
brown (7.5YR4/6) ard yellowish red (5YR5/8) mottles; weak coarse
prismatic parting to weak fir^ subangular blocky structure; firm;
vertical seams filled with very pale brown (10YR7/4) fine sand and silt
form polygonal structure, sard stripping has occurred in seams; ocmmon
fine rounded black concretions; purple stains exterd vertically along
ped faces in upper part of horizon; strongly acid; gradual sncoth
bourdary.
195-250 Brownish yellow (10YR6/6) sandy loam with common medium dark yellcwish
brown (10YR4/4) , strong brown {7.5YR4/6) ard yellowish red (5YR5/8)
mottles; massive parting to weak coarse prismatic structure; slightly
firm in place; polygonal seams filled with very pale brown fine sand
ard silt stripped of clay; catmon black round concretions; mediisn acid;
gradual diffuse boundary.
250-275 Mottled yellowish brown (10YR5/8) , dark yellowish brown (10YR4/4) ,
strong brown (7.5YR5/8) , and pale brown (10YR6/3) sandy loam; massive;
slightly firm in place, friable when disturbed; few black
_ ferrcroanganese concretions; strongly acid. _
Close examination of the soil profile revealed subtle differences within
the dark-colored epipedon. Differences in text\ire and consistency were most
apparent with gradual changes in the class or grade of granular structure.
Because the site had elevations above the adjacent floodplcdn it had been
subjected to intense pedoturbation by insects, crawfish, and burrowing rodents
vbich formed krotovinas and retarded pedogenic development.
The dark-colored epipedon graded into a well-developed paleosol at depths
of 130-150 cm (51.2-59.1 in). The paleosol exhibited a pronounced color
t
I
148
I
i
diange with hues of lOYR and increcised clay content accotpanied by
well-€aq)ressed prismatic structxare vAiich parted to sxibanguleir blocky
structure. Ped faces in the paleosol had oriented clay slcLns and sand
bridging ly clay and iron oxides. Micro-norj^ological analyses of the
peileosol revealed peileo-argillans throughout the matrix. The morphological
and physical features are diagnostic of argillic horizons. There was no
evidence of the ancestreil surface (A) horizon of tlie buried paleosol. Coarser
textured, humic-coated sediments of different morphological characteristics
rested upon the p>aleosol suggesting truncation of the original surface by
fluvial erosion.
The p>cLleosol had a distinctive polygonal morphology ccnprised of prisms
(0.5-1. 5 m or 1.7-5 in diameter) sepsarated by leached, silty and very fine
sand seams. The seam materials were stripped of clay. The seams ranged frtxi
1-3 an (.4-1.2 in) wide at the top of the pialeoargillic horizcai, and they
became thinner with increasing depth. Seams ccrmonly extended to depths of
1 m (3.3 ft) and bisected preexisting structural features and pods with
horizontal cutans.
The thidcness of the pjaleoargillic horizon and the well-developod
structure and oriented clay s)cins on pjed faces indicate the p)cdeosol developed
over a long period of landscape stability prior to the buried. The soil
development and pedogenic expression in the peleosol was ccmparable to that of
well-developed, upland mature soils of adjacent Pleistocene upland surfaces.
Particle size data (Table 52 and 53) indicate discrete fluvial
depositions. The highest silt content occurred in the surface l^er
(0-15 an/0-5.9 in) and decreased with depth. Sand contents generally
increased with deprth. Highest clay ccffitents were associated with the
psaleoargillic horizon of the pjaleosol. The sand fraction was doninated by the
fine (0.25-0.10 mm) and very fine (0.10-0.05 mm) classes, Althcu^ textural
variations occurred within the site as reflected by data in Tables 51 and 52
from different locations on the site, a higher degree of textural uniformity
existed than was detected in the adjacent floodplain as shewn in Table 53.
Higher clay contents and greater variations in silt and clay contents occurred
off-site. Textural bedding planes were also evident in the adjcicent off-site
soils that were not readily evident in the soils cerprising the site vhich
suggests greater stability. Howsver, different depositional energy gradients
are shown by the constant sand fabric (Figure 32) .
T»B[£ 51
Particle size distribution of selected soil
22It539.
Depth
(2 - 0.05
Texture
(on)
0-15
53.00
percentage -
42.08
4.92
sandy loam
15-37
37-60
53.64
50.10
36.88
39.69
9.48
10.21
sandy loam
loam
60-100
56.17
31.10
12.73
sandy loam
100-150
55.92
30.29
13.79
sandy loam
150-180
55.65
25.84
14.51
scindy loam
180-195
71.34
21.11
7.55
sandy loam
195-250
69.88
20.45
9.67
sandy loam
250-275
64.76
23.67
11.57
sandy loam
Level
Texture
(an)
(2-0.05 ran)
(0.05-0.002 ran)
(< 0.002 ran)
1
0-10
66.94
27.40
5.66
sandy loam
2
10-20
62.14
31.00
6.86
sandy loam
3
20-30
62.36
30.79
6.85
sant^ loam
4
30-40
62.33
29.04
8.63
sand^ loam
5
40-50
63.86
29.04
7.10
san^ loam
6
50-60
65.43
27.72
6.85
sand^ loam
7
60-70
64.58
28.33
7.09
santfy loam
8
70-80
63.77
26.84
9.39
sandy loam
9
89-90
63.51
26.31
10.18
sancty loam
0
90-100
64.11
26.47
9.42
san^ loam
1
100-110
64.27
28.61
7.12
sand^ loam
2
110-120
64.33
28.56
7.11
sandy loam
3
120-130
62.94
28.94
8.13
sanc^ loam
4
130-140
58.95
29.08
11.97
sarxty loam
5
140-150
56.03
31.15
12.82
sandy loam
6
150-160
53.60
29.77
16.63
sandy loam
7
160-170
56.92
28.07
15.01
sand^ loam
8
170-180
66.76
23.91
9.33
sandy loam
9
180-190
67.52
23.15
9.33
saiK^ loam
1
200-210
70.81
17.81
11.38
sandy loam
2
210-220
69.98
16.82
13.32
sandy loam
Sand Fraction
vel
Very Coarse Coeirse
Medium
Fine
Very Fine
(2-1 ran)
(1-C .5 ran)
(0.5-0.25 ram) (
- Dprcraitaop
0.25-0.10 ram)
masssm
\
Earticle size distribution of
fijos Blodc A, 22It539 (oontinnad)
Semd Fraction
Level Very Coarse Coarse Meditin Fine Very Fine
- - JlHi'.W - 76'.lPff'.l?anJ (O.lVO.ftitmJ
- percentage -
2
0.13
3.70
41.04
17.21
3
0.07
0.10
3.65
41.33
17.21
4
0.06
3.27
41.12
17.85
5
0.03
0.07
3.37
43.02
17.37
6
0.18
4.22
44.70
16.30
7
0.02
0.09
3.78
43.06
17.63
8
0.11
3.94
43.46
16.24
9
0.01
0.08
3.64
43.32
17.46
10
0.06
3.68
41.48
18.89
11
0.04
0.11
3.64
42.65
17.83
12
0.00
0.08
3.93
42.77
17.55
13
0.01
0.04
3.60
42.07
17.18
14
0.12
3.05
38.35
17.38
15
0.08
3.32
36.07
16.52
16
0.04
2.57
34.72
16.26
17
0.02
0.04
2.66
37.55
16.65
18
0.03
2.81
44.27
19.64
19
0.01
0.01
2.57
44.35
20.58
21
0.01
3.14
48.80
18.86
22
0.02
3.37
49.49
mu 53
Itaticle size analyses and pH of soils adiaoent to 22It539.
Sanple
Sand
Slit
Texture
Terrace east of site
15-30
68.0
14.7
SL
1.9
Bordering floo^lain
Micldle of floodplain
45-63
62.8
12.6
24.6
SCL
4.8
east of
site
0-30
25.8
43.2
31.0
CL
4.8
Middle of flooc^lain
east of
site
50-75
30.7
39.7
29.6
CL
4.4
Middle of floodplain
ecist of
site
75-105
36.4
33.3
30.3
CL
4.8
Floodpleiin 75 m
ecist of
site
30-60
28.2
42.9
28.9
CL
4.8
Flooi^lain 75 m
east of site
75-100
35.2
40.0
24.8
L
4.7
Floodpleiin 20 m
east of
site
25-50
43.0
32.5
24.5
L
4.7
Flooc^ledn 20 m
east of
site
85-125
47.1
23.2
29.7
SCL
A.l
Flooi^lain 75 m
west of
site
5-30
40.7
36.2
23.1
L
5.0
Figure 32 Constant sand fabric, 22It539.
15-37
40-50
25-50
55-85
90-125
100-125
15-37
Silt
Cla:;
36.3
24. (
29.0
20. (
13.0
8,;
20.3
14.;
Article size analyaes and pH of soils adiaoent to 22It539 (ocfitinaed) .
le Depth Sand Silt
lam 75 in
west of site 62-88
Floodplain 120 m
west of site 15-37
Floodplain 120 m
west of site 40-50
Floodplain 40 m
south of site 25-50
Floodplain 40 m
south of site 55-85
Floodplain 40 m
south of site 90-125
Floodplain 40 m
south of site 5-37
Floodpl2dn 40 m
south of site 100-125
Floodplain 75 m
north of site 15-37
Floodplain 75 m
north of site 87-11
* Texture; L = Toot; SL = sandy loam; SCL = sand^ clay loam; LS = loany sand;
CL = clay loam; SiCL « silty clay loam
The presence of illuviation, greater soil structure developmmt, sand
bridping and oriented cl2y skins ext ped faces in the svibsoil indicates greater
pedogenic developnent in the deeper strata in ccxnparison to the upper 1.25 m
(4.1 ft) . Uie soil morphalogical expressions appear to be relic features of
previous soilscapes that were subsequently buried by fluvial sediments. Tlie
pedogenic development in the subsoil of the site contrasted sharply with the
undifferentiated, gleyed, stratified soils of the adjacent floo^lain
(Table 49) .
CHEMICAL DESCRIPnOW
Calcium and hydrogen (acidity) were the dominant exchangeable cations of
the pedons analyzed (Table 54) . Calcium contents varied in different levels
and generally decreased with increasing dep>ths. Cedcium/magnesivin ratios
exceeded 7:1 throu^iout the pedon. Potassium levels were highest in the
surface leyer and in the i?3per horizons of the paleosol. Trace levels of
sodium were present. EHxchangeable cdixninum levels increased with d^^th
reaching maximun values in the paleoargillic horizons. Ebcchangeable alumimxn
levels are generally associated with intaise weathering and age. Cation
exchange capacities were greater in the upper l^ers and decreased in the
buried p6deosol reflecting differences in organic matter content and
mineralogy. Exchangeable cation levels and caticxi exchange capacity values of
the pcdeosol were simileu: to upland Pleistocene soils of the adjacent valley
L53
MU 54
r**— chancacteristicB of rqpteBen^ye 22It539^ -
_ ExcAianqeable Cations _
Base
Depth pH Ca Mg K Na m- Al* Total** _ Saturation
"ot - - - Qipl (pf) kg-1 -
0-15
5.9
6.73
0.90
0.18
0.02
9.85
0.04
17.68
44.29
15-37
5.8
3.81
0.46
0.09
0.02
6.79
0.06
11.17
39.21
37-60
5.5
5.03
0.38
0.08
0.02
8.12
0.34
13.63
40.43
60-100
5.6
6.76
0.65
0.08
0.02
10.02
0.22
17.53
42.83
100-150
5.5
6.24
0.65
0.13
0.03
10.16
0.84
17.21
40.96
150-180
5.5
5.32
0.44
0.11
0.02
7.98
0.81
13.87
42.47
180-195
5.5
2.73
0.22
0.06
0.02
3.15
0.57
6.18
49.03
195-250
5.6
3.07
0.27
0.06
0.02
3.01
0.62
6.43
53.19
250-275
5.2
3.59
0.27
0.08
0.02
4.91
1.46
8.87
44.64
not ine'lxided in tot£tl
** cation exchange capacity by summation
+ acidity
Organic natter content was greatest in the surface horizcai and had ahrvpt
changes with depth (Table 55) . Organic natter and C levels varied across the
site as shown by differences in Tables 55 and 56 for different pedons. Total
N levels showed similar distribution differences. Large variations occurred
in the C/N ratios for different layers (Table 56) The C/N ratio of soil
surface horizons normally ranges from 10-12/1 and decrease with depth
(Fitzpatrick ) with some middle horizons having ratios as low as four.
Ihe lower C/N values of some subsoils has been attributed to high contents of
anmonium ions fixed by clay as a possible mechanism (Fitzpatrick ) . In
contrast to natural undisturbed soils, the pedons analyzed (Table 53) had
dramatic increases in the C/N ratio in the layers frcm 50-150 an
(19.7-59.1 in) with ratios exceeding 20/1. Ratios abnptly decreased at
depths below 160 on (63 in) , with lowest values occurring in the deepest
layers analyzed.
TABU 55
Ocgaiuc natter, free iron oxides, total P and 1% citric acid extractable
P205 levels of selected pedon, 22It539. _ _
IIHSilllil
Organic Matter
Fe203
Total P
1% Citric Acid
ExtraK:±able P205
(on)
- percentage -
—ppm -
0-15
2 75
1.3
509
183.9
15-37
0.73
1.8
422
223.8
37-60
0.96
1.3
574
305.8
60-100
1.38
1.2
563
368.5
100-150
1.03
1.4
702
532.2
150-180
0.46
1.2
667
591.8
180-195
0.09
0.7
327
181.8
195-200
0.06
0.8
336
142.5
250-275
0.08
1.3
535
182.9
154
Vtax 56
Ckganic
PH» cazbon, nitrogen cxxxtents, cartxWnitrogen ratio, and 1%
Level
Depth
Qrgamic Matter
-EH
C
N
C/N
1% Citric Acid
Soluble P2C6
1
(am)
0-10
- percentage —
4.20 5.1 2.43
0.390
6.2/1
ppm
414.78
2
10-20
1.15
6.3
0.66
0.066
10.0/1
395.83
3
20-30
0.61
6.3
0.35
0.039
8.9/1
469.39
4
30-40
0.67
6.1
0.39
0.031
12.5/1
468.91
5
40-50
0.77
5.9
0.45
0.028
16.0/1
425.96
6
50-60
1.25
5.9
0.72
0.025
29.0/1
344.13
7
60-70
5.9
0.53
0.023
23.0/1
343.78
8
70-80
6.1
0.60
0.026
23.0/1
526.85
9
80-90
1.14
6.1
0.66
0.026
25.0/1
528.99
90-100
1.25
6.1
0.72
0.024
30.0/1
507.10
11
100-110
1.55
6.0
0.90
0.025
36.0/1
548.78
12
110-120
1.58
5.9
0.92
0.023
40.0/1
549.34
13
120-130
1.58
5.8
0.92
0.024
38.0/1
752.03
14
130-140
1.40
5.8
0.81
0.029
28.0/1
.25
15
140-150
1.40
5.6
0.81
0.035
23.0/1
.85
16
150-160
1.28
5.8
0.74
0.035
21.0/1
.34
17
160-170
0.79
5.8
0.46
0.033
14.0/1
.31
18
170-180
6.0
0.11
0.017
6.5/1
626.90
19
180-190
0.15
5.9
0.08
0.017
4.7/1
506.59
21
200-210
0.13
6.0
0.07
0.016
4.3/1
284.26
22
210-220
0.11
5.9
0.06
0.016
3.7/1
284.55
Soil pH levels varied sli^tly with depth and location across tlie site
(Tables 52 and 56) . The low j* value of the surface (0-10 an/0-3.9 in) layer
of the pedon of Bloc)c A is associated with the higher organic matter content.
The soil levels in the site were COTsiderdsly hi^ier them adjacent
floodplain soils, vdiich had average levels below five. The hi^ier pH values
in the site are associated with hi^%r calcium levels.
In addition to specific requirements for organic matter content, color and
other properties; anthropic epipedons are required to have at least 250 ppn
PjO- retractable in 1% citric acid (Soil Taxonony ) . All the layers
analyzed in the Block A pedon had levels exceeding 250 ppn (Table 56) , with
nexisum values occurring at depths of 130-170 an (51.2-67 in). The citric
acid (1%) extractable P-O- levels could be associated with population, time of
occupation, or types of ^activities. Some variation occurs in the citric acid
soluble PoOc across the site and with d^5th as shewn in Tables 56 and 57, as
might be rpected. Citric acid extractable P^O- levels of adjacent off-site
soils were less than 150 ppm for selected saitples analyzed.
MINEa^ALOGICAL DESCRIFTICXf
Kaolinite was the dominant mineral in the clay fraction with lesser
amounts of illite, vermiculite-chlorite intergrade, anectite, and quartz.
Vertical layers did not eadubit discrete clay mineral suites, but reflected
the depositional fluvial environment. However, vermiculite-chlorite
intergrade content decreased with depth, and it was not detected in the
underlying p>aleosol horizons. The absence of chlorite-vermiculite in the
155
I
I
I:;
r-
( ■
j; •
l :
I
i
i
I
i
I
i
paleosol relative to the siperjacent horizons suggests greater weathering in
ihe paleosols. Ttiis finding agrees with data of Suhe et al. () vdx>
studied pedeosols in Indiana. Sand and silt fractions were doninated by
quartz with lesser amounts of mica, glauccaiite, and feldspar.
MICRO-MM«OILOGICAL DESCRIPTION
Examination of natural soil peds and separated sand fractions were done
for seleoted sanples via conventional li<^t microscopy. Representative
saiqples of the dark-colored, hiitiic-stcdned upper layers and subjacent paleosol
layers were examined in detadl.
IPIPEDON
The skeletal matrix (sand and silt) was cornprised doninantly of sand
grains which vere uniformly coated with reddiidi brown and black organic
matter. The individual coated sand grains were bound or cemented by hunus
into clusters or domains vdiich fanned moderate to strong granular structures.
The structuraul units are resistant to disruption by water emd they tend to wet
very slowly, suggesting a high cetpillary contact angle. Shaidng the soil in
0.05 N NaOH removes or "strips" the huius coating revecding angular and
rounded, clecu: quartz grains. The sand grcdns beccme loose and single grained
after removal of the coating and tend to re-wet readily. The epipedon of the
midden mound presents a striking contrast in ccnpariscxi to adjacent flooc^lain
soils. The adjacent soils have a very thin (0-6 cm/ 0-2. 4 in) surface layer
that is not as dark colored with Munsell hues of lOYR and which has a "scdt
and pepper" ^pearance with mostly unooated quartz grains and unbound organic
matter. The ncxi-site soils have much less defined structural units and tend
to wet more readily.
Additional tests were ccxiducted to characterize further the nature of the
hmus coating. Immersion of the himic-stained soil in water for 12 hours with
shaking resulted in no unbound or "free" organic matter, and the structural
domains persisted. Treatment of the epipedcsi with 0.5 NaCW and gentle shaJcing
removed the humus coating and resulted in a reddish brown extract. Addition
of 0.5 HCl to the colored extract resulted in ocmplete flocculation of the
organic material leaving a clear supernatant indicating dominantly humic acid
ccnpounds (McKeague ) . The color ratio of 0.5 N NeiCXI extract solutions
was determined to characterize further the coating materials. The color ratio
of: E4/E6 = extinction (absorbance) at 400 rm
extinction (absorbance) at 600 nm
has been used for differentiation of humic substances (Tan ) . Pulvic
acids yield spectra with a steep slcpe in contrast to humic acids. According
to research by Tan () , a lew color ratio less than seven corresponds to
himic acids and related cenpounds with hi^ moleculeu: wei^ts.
Color ratio values for the representative pedon of Block A (Table 57)
indicate dominantly hvatiic acid coipounds. Values sli^tly exceed seven at
depths of 210 can (82.7 in) and greater.
TABIZ 57
Color ratio values of MaOH extracts of pedon from Block A, 221t539. _
absorfaeuToe 400 rm
Level Depth (cm) Color Ratio Vedue = absorbance 600 nm
1 0-10 6.13
2 10-20 5 72
156
I
•OBtX 57
Onlcr ratio values of NaOH extracts of pedcn fm Block A, 22It539
(ocntifiedl . _
Level
Depth (on)
absodsanoe 400 rm
Color Ratio Vadue - absorbance 600 nm
3
20-30
5.52
4
30-40
5.37
5
40-50
4.55
6
50-60
3.84
7
60-70
3.82
8
70-80
4.42
9
80-90
4 34
10
90-100
3.96
11
100-110
3.78
12
110-120
3.52
13
120-130
3.72
14
130-140
4.21
15
140-150
4.08
16
150-160
4.54
17
160-170
5.26
18
170-180
6.15
19
180-190
6.50
21
200-210
7.33
22
210-220
7.50
PALElOSaL
Microscopic examinations revealed that paleosols ccmnonly have a
fine-grained S-matrix with void argillans ai^d embedded grain argillans
dispersed throu^iout the matrix. Voids were typically smaller in the
pedeoargillic horizons than exist in the overlying epipedons. Ihe paleosol
had a high content of vesicular pores. Hie pedeosol tended to have ccnpound
structure with prismatic parting to well-developed subangular blocky
structure. The polygcml seams between prisms were highly leached and
stripped of fines.
SLMMRRy
Site 22It539 was located in the eastern part of the TOnbigbee River
floodplain about 700 m (2,310 ft) west of the eastern Pleistocene valley wall.
Lower-lying parts of the site were subject to flooding during winter and
spring months. The site soil was distinguished by very thick, hcrais-ridi,
dark reddish brown sandy loam upper layers ranging one meter (3.3 ft) and
greater in thickness. The dark-oolored ^ipedon had Munsell hues of SYR vivich
ccaitrasted sharply to adjacent off-site soils. Organic matter contents ranged
fron 2.75% in the surface layer to 0.08% at depths below 250 on (98.5 in) .
The dark-oolored epipedon graded into a well-develcped paleosol at depths of
130-150 cm (51.2-59.1 in). The paleosol had color hues of lOYR and increased
clay cxMitent eiccempanied by well-developed structured units. The paleosol had
a distixKtive polygonal morphology cenprised of prisms separated by liighly
leached silty and sandy seams. Soil development and pedogenic ejqsression in
the paleosol was cotiparable to well-developed mature soils of adjacent
Pleistocene upland surfaces. Sand contents increcised with depth, and the sand
fraction was deminated by fine and very fine sand. The paresence of
157
llluviaticn, greater soil stnictural development, sand-bridging, and oriented
cle^ skins ped faces in the paleosol indicate greater pedogenic development
in ocnparison to the i^iper 1.25 m (4.1 ft).
Calcium and hydrogen (acidity) vere the dominant exchangeable cations of
the site soil. Calciixn contents varied in different levels and generally
decreased with increcising depths. Calcium levels were several fold greater
than levels of adjacent non-site soils. Potassium levels were hi^iest in the
surface l^er and the upper horizons of the paleosol. Trace levels of sodium
were detected with no accunulations. Exdiangeable aluminum levels increcised
with depth with maximum values occurring in the paleoargillic horizons. Soil
pH varied with depth and location across the site, but values were
considerably hi^ier than off-site soils. Base saturation levels were also
higher in the site soil than adjacent off-site reflecting cultural addition of
cations to the site.
Total nitrogen contents decreaused with depth and did not correspond to
increaised organic carbon levels at different depths. Hie site had erratic C/N
ratios vMch differed markedly from non-site soils. All the layers of the
site soil contained greater than 250 ppm citric acid (1%) solvible PjOe/ vMch
is one criteria for anthropic epipedons. Maximum P205 vadues occurfea at
depths of 100-180 an (39.4-70.9 in) reflecting cultural additions.
Kaiolinite wais the dominant mineral in the clay fraction with lesser
amounts of illite, vermiculite-chlorite integrade, smectite, and quartz. Sand
and silt fractions were doxdnated by quartz with lesser anounts of mica,
glaixxxiite, and feldspar. Vertical layers did not eadiibit discrete clay
minerad suites.
The skeletal matrix (sand and silt) of the daurk-oolored epipedon was
oonprised of sand grains which were uniformly coated with reddish brown and
black humus. The coated sand grains were cemented by hutius into clusters
forming granular structure. Ihe humus coating was not removed by repeated
wetting and drying. Analyses indicated the amorpSteus coating was dominantly
conprised of humic acid ocnpounds.
THE POPLAR SITE (22It576)
The site occvpied a topographic hi^ in the eastern peurt of the Tcirbi^aee
River floodplcdn about 300 m (990 ft) west of the valley wall. The floodplain
had slopes of 0-2% in contrast to slopes of 2-5% for the occupation mound
(Figure 33) . Loced microrelief existed in the flooc^ledn due to scouring and
filling by flood waters. The level floodpledn merged abruptly with the steep
vcdley wcdls.
Hie site appeared to lie a topographic feature caused natural fluvial
deposition. It was surrounded l>y a narrow slough (Figure 33) which appeared
to be aggrading. The slough was much wetter thrcughcut the year than adjacent
areas in the floodplain.
Uie bottom sediments of the slough were dominantly blue-gray and
olive-colored silt loam and silty clay loam overlying loarty materials. Ilie
blue-gray color reflected the gleyed conditions resulting from wetness and
lack of aeration. The floodpladn sediments were siliceous but contained
considerable glauconite and mica.
Mature, well-developed soils with distincrt eluviated E horizcxis and
illuviated argillic horizons (Bt) oonprised the tpland areas adjoining the
floo(^leiin. amithdale soils (Figure 31) of the ipland areas directly east of
tlie site had red subsoils that contciined tp to 35% clay in the argillic
horizons. The upland soils were very strongly acid, highly weathered, and
158
siliceous with low base saturation levels (Ultisols) . The Ora soils located
in the uplands northeast of the site had dense, firm fragipan horizons at
depths of 50-75 an (19.7-29.5 in) .
The floodplain surrounding the site was caiprised of Kirlcville and
Mantachie soils vdiich had ndnimal pedogenic developnent (Table 58) . These
soils had brown and yellowish brown surface horizons and gray, li^t gray, or
pale brown (Table 59) subsoils with loany textures. The floo(^lain soils had
canhic Bw tiorizons (color B) with little eluviation and illuviation. They
were strongly acid.
T»BI£ 58
Clafwl f1 cation of soils
Soil Series Classification
Site 22It576 and vicini
1KB[£ 59
Mmsell cnior of selected berizons of representative soils in tihe flno^lain
adjacent to Site 22lt576.
Sample
Depth (an)
Munsell Color (moist)
Hole 1 North
75-90
Gray (10YR5/1)
Hole 1 North
100
Gray (10YR5/1)
Hole 2 South
30-50
Li^t gray (10YR7/2)
Hole 2 South
50-70
Light gray (10YR7/2)
Hole 2 South
70-92
lii^t brownish gray (10YR6/2)
Hole 2 South
92-112
Gray (10YR5/1)
Hole 3 South
50-75
Light gray (10YR7/2)
Hole 4 East
25-50
Gray (10YR5/1)
Hole 5 East
10-30
Dark brewn (10YR4/3)
Hole 5 East
50-70
Li^t brownish gray (10YR6/2)
Hole 6 West
80-100
Gray (10YR8/1) with (10YR5/6) mottles
Hole 7 Southwest
65-75
Gray (10YR8/1)
Hole 8 Southwest
0-30
Peile brown (10YR8/3) and gray (10YR5/1)
Hole 8 Southwest
30-40
Light gray (10YR7/2)
The culturally altered soils of the site developed in loany, flxivial,
siliceous sediments. Tliese soils were readily distinguished 1;^ very thick,
humus rich, dark reddi^ brown epipedons (surfaces) vAiich were due to
prolonged cultural activity and habitaticai. Past occupation of the site had
drastically altered normal pedogenic features of color, structure,
consistency, horizonation, organic matter content, and certain chemical
parameters. The soil ccrprising the occupation locale wcis distinct, and it
differed greatly from adjacent soils of tl^ region.
Profuse populations of earthworms, crawfish, rodents, and other diverse
microfauma and microflora thrived in the organic-rich site vdiich was elevated
above the adjacent flooc^lciin and seasoned wetness. Faunal and floral
pedoturbation in addition to the human activities had tended to mix the upper
meter of soil cind affected normal pedogenic develcpment. The dark-colored
humic staining of the upper meter (3.3 ft) also tended to mask the natural
horizonation.
PHYSICAL DESCRIPTION
The upper meter (3.3 ft) of the mound soil was dark reddish brown with a
moist hue of SYR (Table 57 and Figure 34) v^iich differed markedly from the
adjacent floodplain soils which had hues of lOYR (Table 60) .
•DfflUS 60
Pedon description of repaceaentative profile, 22It576. _
Depth Description _
(cm) (moist colors) ^ '
Dark reddish brown (5YR3/2) loam; moderate-fine and medium granular
structure; very friable; many fine and medium roots; few small black
(10YR2/0) chcurcoal fragments; greasy vhen rubbed; strongly acid;
clear smooth bcundary.
0-10
A 0-l0ctn
Abi 10-S0
Abj 50-70
Aba 75-115
Ab« 115-130
2B lb 130-200
PALEOSOL
DARK REDDISH BROWN LOAM
GRANULAR STRUCTURE
DARK REDDISH BROWN LOAM
GRANULAR STRUCTURE
DARK REDDISH BROWN LOAM
GRANULAR STRUCTURE
DARK REDDISH BROWN AND
BLACK LOAM
GRANULAR STRUCTURE
VERY FRIABLE AND LOOSE
NUMEROUS CHARCOAL
DARK REDDISH BROWN
SANDY LOAM
GRANULAR STRUCTURE
DARK YELLOWISH BROWN
LOAM
ANGULAR BLOCKY STRUCTURE
FIRM
POLYGONAL SEAMS
Figure 34 Soil profile, 22It576
i
4
raeuB 60
Bedon deacription of reppeaentative profile, 22It576 (coptinaed) . _
Dep^ PesCTiption
(an) (moist colors)
10-51 Dark reddish brcwn (5YR2.5/2) loam; moderate- fine granular structure;
sli^tly firm in place, friable vAien disturbed; many fine and madiun
roots; few small charcoal fra^nents; canton anall and medium very
dark gray (10YR3/1) and dark gray (10YR4/1) potsherd in v?>per part of
horizon; medium acid; greasy when rubbed; gradued wavy boundary.
51-75 Dark reddish brcwn (5YR3/3) loam; weak-fine granular structure;
slightly firm in place, friable when disturbed; cannon snail roots;
few mottled dusky red (2.5yR3/2), reddish brown (2.5YR4/4), yellowish
red (5YR5/8) "fired aggregates"; common small black charcoal frag¬
ments; medium acid; greasy when rubbed; gradual wavy boundary.
75-87 Dark reddish brown (5YR3/2) and strong brcwn (7.5YR5/6) loam with
streaks of light brownish gray (10YR6/2) and black (10YR2/0) ;
weak-fine granular structure; very friable; abundant black and
very dark gray nulti-sized chcuccoal fragments; few fine roots;
medium acid; gradual wavy boundary.
87-97 Dark reddish brcwn (5YR3/2) and strong brown (7.5YR5/6) sandy loam
with ccrmon medium light brownish gray (10YR6/2) mottles; weak-fine
granular structure; friable; coanon black charcoal fragments; medium
acid; gradual wavy boundary.
97-113 Reddish brown (5yR4/4) and strong brown (7.5YR5/6) sandy loam;
weak-fine granular structure; slightly firm; few fine charcoal
fragments; few fine round black concretions; slightly acid;
gradual wavy boundcury.
113-120 Strong brown (7.5YR5/6) and dark brcwn (7.5YR4/4) loam; weak-fine
grcinular structure; slightly firm; few fine black concretions;
purple stains extend vertically; slightly acid; gradual wavy bcundary.
120-130 Strong brown (7.5YR4/6) loam with cannon medium brcwn (10YR5/3)
mottles; weak fine granular structure; slightly firm; cannon fine
and medium round black concretions; purple stains extend vertically
along ped faces; medium acid; smooth wavy boundary.
130-146 Dark yellowish brown (10YR4/4) with cannon medium strcaig brown
(7.5YR4/6) , dark brown (7.5YR3/4) and yellcwish red (5YR5/8) mottles;
weak fine platy structure that parts to weak- fine subangular blocky
structure; firm; thin seams filled with very peile brown (10YR4/4)
silt and very fine sand form an intermittent polygonal network; firm;
patchy clay skins and intergranular bridging; cannon black ferro¬
manganese concretions; strongly acid; gradual irregular boundary.
4
i
A
A
J
J
162
of
Description
(moist colors)
five pirofile, 22lt576 (ocntinued) ,
146-184
Mottled pale brown (10yR6/3) , brcwnish yellcw (10YR6/8) , strong brown
(7.5YR5/8), yellowish red (5YR5/8) , and red (2.5yR4/8) loam;
massive parting to weak fine subangulau: structure; firm; seams
ranging to 0.5 an (.2 in) width and filled with very pale brcwn
(10YR4/4) silt and very fine sand form a continuous polygonal
structure, sand stripping has occurred in the seams; common ferro¬
manganese concretions; clay skins on ped faces and within larger
pores; strongly acid; gradual irregulcu: boundary.
184-200+ Mottled yellowish brown (10YR5/6) , pale brown (10YR6/3) , olive yellow
(2.5Y6/8), brownish yellow (10YR6/8) , and reddish yellow (7.5YR6/6)
loam; massive; slightly firm; few black ferromanganese concretions;
strongly acid.
The mound soil epipe’on had a Munsell value vMch changed at least one
unit from wet to dry in o' ntrast to adjacent soils which did not exhibit this
color change. The dark reddish brown epipedon iiad a distinct "greasy" or
slick feel v^en rubbed between the fingers. The dark epipedon graded into a
paleosol at depths ranging from 130-150 an (51.2-59.1 in). The paleosol
exhibited a pronounced change in color with hues of lOYR and increases in clay
content accompanied by a structviral change to sulaangular blocky. Ped faces in
the paleosol had oriented clay skins and sand bridging lay clay vdiich is
characteristic of argillic horizons. A very pronounced polygonal network
permeated the paleosol, with the polygons separated by seams filled with silt
and very fine sand v^ch had been stripped of clay. Clay content decreased in
the lower horizon of the paleosol and structure changed to massive. The
horizons from 130-184 cm (51.2-72.5 in) had well-developed ped faces
characteristic of argillic horizons viiich occur on upland mature soils of the
region, l^piccilly, below the dark-colored epipedon and above the brighter
colored paleosol, the humic materials formed horizontal bands or lamellae
vdiere vertical water flow had been retarded by the less permeable underlying
paleosol.
The thickness of the paleoargillic horizon and the well-developed
structure and oriented clay skins on ped faces suggest the paleosol formed
over a period of landscape stability prior to burial. The pedogenic
developnent in the paleosol was comparable to well-developed, upland soils of
adjacent Pleistocene geomorphic surfaces. Micro-morphological analyses of the
pcdeosol revealed the paleo-argillans extended throu^out the matrix. Ihere
Wcis no morEhological evidence of the ancestral surface horizon of the buried
paleosol. Other than the missing surface horizon, the paleosol ajpeared
intact and well preserved under the burial sediments.
Particle size distributions (Table 61) suggest discrete fluvial
depositions in the UR^er meter (3.3 ft) and reflect argillation in the
underlying paleosol. Si]t contents were generally higher in the siurface
layers, and clay increased with d^th reaching maximum levels in the
p>aleoargillic horizon. Fine (0.25-0.10 nm) and very fine (0.10-0.05 nm) sand
fractions were donincint with only trace anounts of coarse (0.25-0.10 nm) and
very coarse (2-1 nm) sand. Although textural variations existed in the site,
a higher degree of textural uniformity e:cisted than was detected in the
163
adjacent flocx^lain soils (Table 62) . Praninent textural bedding representing
fluvial depositions was very evident in the adjacent off-site soils in the
active floo<^lain, which indicates the site had greater stability for
pedogenesis.
!I3VB[£ 61
Barticle size distribiition of representative pedon, 22It576.
Sand
Silt
Clay
(an)
0 -
10
47.2
42.8
10.0
Loam
10 -
51
48.4
37.2
14.4
Loam
51 -
75
46.6
33.3
20.1
Loam
75 -
87
48.8
30.9
20.3
Loam
87 -
97
57.7
25.5
16.8
Sandy Loam
97 -
113
56.5
29.5
14.0
Sandy Loam
113 -
120
50.4
35.3
14.3
Loam
120 -
130
42.7
39.6
17.7
Loam
130 -
146
39.0
38.1
22.9
Loam
146 -
184
38.3
36.5
5.2
Loam
184 -
200
48.0
30.2
21.8
Loam
Sand Fraction
Very Coarse
Coeurse
Medium
Fine
Very Fine
Depth
(2-
-1 nm)
(1-.5 nm)
(.5-. 25 itm) (.25-
.10 nm)
(.10-. 05 mm)
(can)
*^peircen ■■
0-10
0.42
0.42
3.90
27.33
15.15
10 -
51
0.10
0.30
3.84
27.96
16.22
51 -
75
0.02
0.10
1.63
24.72
20.14
75 -
87
0.06
0.25
3.44
29.63
15.39
87 -
97
0.01
0.12
7.24
34.95
15.33
97 -
113
0.07
0.14
5.52
34.03
16.77
113 -
120
0.02
0.23
2.65
28.95
18.50
120 -
130
0.01
0.25
1.53
23.19
17.67
130 -
146
0.01
0.04
1.20
20.52
17.24
146 -
184
0.01
0.05
1.06
18.94
18.28
184 -
200
0.00
0.04
1.54
27.12
19.32
TABEf 62
Particle size distributicn, pH, and organic matter ocsitents of representative
soils ^ri[jaoent. to Site 22It576. _
Hole Qrgeinic
Depth
Sand
Silt
Clay
Texture
■ PH
Matter
cm (2-
.5nm)
(.05-.002nm)
(<.002nm)
1 North
75-90
52.6
26.7
20.7
SCL
4.4
1.02
1 Itorth
100
56.9
23.4
19.7
SL
5.0
0.43
2 South
30-50
45.2
29.9
24.9
L
4.9
0.57
2 South
50-70
76.6
13.5
9.9
SL
4.6
0.21
2 South
70-92
84.3
10.6
5.1
LS
5.1
0.15
2 South
92-112
79.6
9.8
10.6
LS
5.3
0.17
TABUB 62
Bsorticle sixe distrlbuticn, pB, and organic matter oontents of representative
Hole
Sample
HSU
Sand
Silt
IHH
Texture
pH
Organic
Matter
cm (2-
.5nm)
(.05-.002nm)
(<.002tnn)
3 South
50-75
47.6
28.0
24.4
SCL
4.3
1.54
4 East
25-50
24.4
36.1
39.5
CL
4.3
2.69
5 West
10-30
42.7
27.4
29.9
SCL
4.8
1.42
5 Wfest
50-70
41.4
27.7
30.9
CL
4.9
0.60
6 West
80-100
55.9
22.9
21.2
SCL
4.8
0.33
7 Southwest
65-75
40.6
30.5
28.9
CL
4.6
0.65
8 Southwest
0-30
36.8
38.5
24.7
L
4.9
1.58
8 Southwest
30-40
41.9
34.3
23.8
L
5.2
1.08
L=loam; SLr=sandy loam; SCL=sandy clay loam; LS=loany sand; CL=clay loam
Hie constant sand fabric illustrates the d^x^sitional gradients of the
Poplar site (Figure 35) . Hiis methcxi has been used to analyze the
d^xisiticoial environtnent characterizing a soil profile, cussuming the sand
fabric to be the skeletal matrix through vAiich clay would move. Clay
distrifaution suggests a downward translocation of particles from the t^per
horizons with a major discontinuity existing between the tpper clay maxima
about 60 on (23.6 in) and the subjacent paleoargillic horizon.
Greater soil structural development, argillation accoipanied by sand
bridging, and oriented clay skins on ped faces in the paleosol indicate
greater pedogenic development in the deeper strata relative to the upper
130 an (51.2 in) . The profile position, lack of organic staining, and
pronounced pedogenic expression in the paleosol are indicative of its greater
chronological age. The soil morphological expressions appear to be relic
features of previous soilscapes that were subsequently buried by fluvial
sediments.
CHEMICAL DESCRIPTION
Ceilcium and hydrogen (acidity) were the dominant exchangeable cations
(Table 63) . Calcium oontents varied in different levels with maximum values
occurring at depths of 75-87 cm (29.5-34.3 in) in the pedcai ancdyzed and
decreasing with depth. Calcium levels were three to four times greater than
levels off-site in the adjacent floodplain. Ca/Mg ratios exceeded 8:1
throu^out the pedon, with vedues generally decreasing with increasing depth
except in the surface layer. Potassivan levels were much higher in the surface
layer, with a second naxima occurring at depths of 51-97 cm (20.1-38.2 in) .
Trace levels of sodium were present as might be eaqected in the highly leached
environment. Exchangeable aluminum levels increased dranaticedly in the
underlying p)aleosol, with maximum values occurring at depths of 180-200 cm
(70.9-78.8 in). Higher eiorhangeable aluminum values are gemrally cissociated
with intense weathering and reflect age. Cation exchange capacities were
greater in the upper layers reflecting the higher organic matter contents.
Base saturation levels increased with depth, and they were much greater than
levels of adjacent floodplain soils.
165
D*pth
HOE 63
ciiaracteristics of uepteaentative Pedon# 2^. _
Exchangeable Cations
Base
BSSuSHI
pH
Ca
Mq
K
Na
H
AL
Total Saturatican
an
Qnol (
prj
0 -
10
5.3
11.32
1.28
0.42
0.03
15.54
0.09
28.59
45.65
10 -
51
6.0
12.48
0.84
0.09
0.03
11.45
0.03
24.89
54.00
51 -
75
5.7
12.90
0.83
0.10
0.04
12.46
0.15
26.33
52.68
75 -
87
5.8
15.00
0.69
0.14
0.03
12.18
0.05
28.04
56.56
87 -
97
6.0
9.34
0.48
0.11
0.04
7.28
0.04
17.25
57.80
113 -
120
6.1
8.62
0.48
0.08
0.03
6.49
0.00
15.70
58.66
120 -
130
6.1
8.49
0.48
0.09
0.03
6.72
0.04
15.81
57.50
130 -
146
5.7
8.61
0.60
0.07
0.04
6.64
0.45
15.96
58.40
146 -
184
5.4
8.99
0.70
0.09
0.04
8.52
1.53
18.34
53.54
184 -
200
5.3
10.00
0.85
0.10
0.07
8.94
1.91
19.96
55.21
200 -
230
5.3
8.86
0.82
0.09
0.06
6.34
0.88
16.17
60.79
Soil p« levels varied with depth (Table 63) . Hi^iest pH values ocscurred
at depths of 87-130 an (34.3-51.2 in) , and they were associated with Icwer
exchangeable acidity. The soil pH levels in the site were considerably higher
than adjacent floodplain soils, v^iich had average vadues of 4.8. The higher
pH values of the site are associated with higher exchangeable calcium and
lower acidity levels.
Organic matter levels generally tended to be greatest in the surface
horizons (Tables 64 and 65). However, levels exceeding 3.5% were detected in
the 50-60 on (19.7-23.6 in) layer of BlocJc D, vhich indicates the variation
across the site. Organic matter levels exceeded 1% in the t5)per 80 an
(31.5 in) and coincided with the darJc reddi^ brcwn 5YR hues. Total nitrogen
levels also decreased with d^xth and did not correspond to increased organic
matter levels that occurred at various depths. This suggests a depletion of N
or concentration of C. This trend is also reflected in the C/N ratio which
increeises below the surface layer and does not substantially decrease until
depths of 80 an (31.5 in) and greater. The large C/N ratios are due to the
relatively low levels of N cotpared to C, vdiich suggest concentration of C
and/or depletion of N. The C/N ratio of natural, non-disturbed soils of the
region ranges fron 10 or 12/1 eind decreases with depth. The hicb C/N ratios
and presence of eppreciable amounts of C in deeper layers is a ^arp contrast
to undisturbed soils of the region. It is interesting to note that the
Munsell color hue does not change with these variations, and little change was
noted in chrana and value. The paleosol had organic matter levels less than
0.5%. Free iron oxide contents (Fe20,) , shewn in Table 65 exceeded 3% in the
upper 50 on (19.7 in) and decreased^abrvptly before increasing again in the
paleosol. The Fe-0- levels in the paleosol are similar to levels exxurring in
argillic horizons'^or adjacent vpland Hapl\idults. Higgler ^620, levels are
generally cisscxriated with hi^er clay contents in soils of we eurea. However,
clay contents of the surface horizons were less than deeper horizons, posing
an enigma for the higher levels detected in the surface layers.
167
64
Cfcganic aatter, caztxn, nitrogen contents, caztcn/nitzogen ratio, and 1%
Level
Orrrnic
Depth (cm) Matter (%) C (%)
N {%)
C/N
1% citric Acid
Soluble P205 (pim)
1
0-10 2.75
1.59
0.109
14.5/1
466.6
2
10 - 20 1.83
1.06
22/1
3
20 - 30 2.29
1.33
26/1
934.6
4
30 - 40 2.27
1.31
0.041
31.9/1
627.7
5
40 - 50 2.43
1.41
0.043
32.8/1
771.7
6
50 - 60 3.53
2.05
0.046
44/1
1,230.4
7
60 - 70 2.61
1.51
0.052
29/1
2,116.2
8
70 - 80 1.59
0.92
0.042
22/1
1,603.3
9
80 - 90 0.98
0.57
0.031
18.4/1
838.4
10
90 - 100 0.81
0.47
0.030
15.7/1
920.2
11
100 - 110 0.47
0.27
0.030
9/1
774.7
12
110 - 120 0.27
0.16
0.028
5.7/1
634.6
•mOE, 65
organic matter and Fe203 contents of typical pedon, 22It576.
Depth (cm)
Organic Matter (%)
Fe203 (%)
0-10
3.16
3.7
10 - 51
1.77
3.0
51 - 75
1.53
2.0
75 - 87
1.90
2.0
87 - 97
0.74
1.4
97 - 113
0,77
0.7
113 - 120
0.50
1.3
120 - 130
0.23
1.5
130 - 146
0.23
1.8
146 - 184
0.21
2.0
184 - 200
0.12
1.4
Anthropic epipedons are required to have at least 250 ppn P-0_ extractable
in 1% citric acid in additicxi to definitive color and organic matter contents
(Soil Taxonany ) . All the layers analyzed in the Block D pedon had PpOc
levels well above 250 ppn (Table 66) . Maximum vadues occurred at d^Jths of^
50-80 cm (19.7-31.5 in) , and the levels were much hi^ier than l^ers above and
below these depths. The hi^er levels at these cJepths nay be indicative of
greater population, habitation time, or other activities vrfien the layers at
50-80 cm (19.7-31.5 in) were the antec:edent surface horizcxi. In cxmparison,
citric acid extracrtable ^2^5 levels of adjac^ent off-site soils were less them
120 ppm for selected soil'^samples analyzed.
•DVBLf 66
Soil phoephorus
repreaentative profile from Block D, 22lt576.
Riosphorus Freictions _
fractions of
rus Fractions
804
768
838
6
3
5
14
04
50
201
32
31
34
28
29
1,652
1,132
1,448
809
749
208
29
1,796
.-M
926
851
338
35
2,149
1,007
314
173
30
1,523
714
323
9
23
1,069
0 944
225
9
26
1,204
0 814
295
4
21
1,133
- ' -
0 831
349
4
21
1,205
. i
Phosphoms is one of the least nobile soil constituents. However, under
intense ^feathering conditions with warm tenperatures and excess precipitation,
or over long periods of time, soil phosjitorus undergoes chaises in chendcad
form and locaticxi in the profile (Walker and Syers ) . Bie levels of
various soil phosphorus fractions have been used to quantify soil developnent
and chronological age relationships (Walker ) . Research in New Zealand
(Walker and Syers ) shewed the soil phosphorus transformations consisted
of the dissolution of calcium phosphates and the formation of organic and
aluminum and iron-oxide phosphorus species. Ihey reported that with
pedogenesis all forms of soil phosphorus were transformed almost ocnpletely to
the organic and oocliided forms. The phosphorus fractions extracted by
sequential laboratory extractions in these studies were grouped into
functional P fractions. These groupings exjnsisted of non-occluded P (Pnoc)
which has been associated with aluminum and iron oxide surfaces and is readily
avadlable for plant use; occluded P (Poc) which has been associated with
aluminum and iron-oxide lattices and is not readily avedlable for plant use
except under a reduced envirorment; calcixm-bound P v^ch is considered not to
be readily avedlable for plant use. Meixner and Singer () studied
Edxjsphorus fractions fran soil profiles formed in mixed alluvium in California
of a chronosequence ranging in age from 300-250,000 years. reported that
occlxxled phosphorus (Poc) generally increased, and the content of non-occluded
phosphorus (Pnoc) decreased in B horizons. They found that calcium-bound
phosphorus (Pea) decreased over time in the surfeu:e horizons, but changed
little in B horizons. In this study, the organicedly bound phosphorus (Po)
did ixjt change with time.
The discrete soil phosphorus fractions determined by sequential laboratory
extractions for a representative pedon froii 22It576 are paresented in Table 66.
Organic P levels have been used widely in archaeologiced studies (Griffith
) and related to hunen oocup)ation. It is interesting to note that maximum
values of non-occluded, occluded ceilcivin, and toted phosphorus occur at depths
of 50-70 on (19.7-27.6 in) , and maxinum values for occluded phosphorus occur
at depths of 90-100 on (35.5-39.4 in). These data seem to indicate greater
additions of phosphorus materials from occupation that occurred when this
layer was the exposed surface of oocupiation. It may be depictive of a long
period of occupsation or larger pxjpulation on this occupational surface. The
organic phosphorus levels remain relatively low to depths of 120 cm (47.3 in) .
169
The ncxi-occluded jAiosiAionis fraction, which has been associated with aluminun
and iron oxide surfaces, was the doninant fraction (Table 67) . Occluded
phosphorus ccn(3rised less than 42% of the total. The organic p^iosphorus
feaction contents ranged from 4.3% in the surface layer to 1.9% at a depth of
110-120 can (43.3-47.3 in). The literature suggests PCa contents might be
expected to decrease with time as weathering relecises P from this fractiai for
conversion to other forms. The PCa levels did not exhibit a steady decreaise
until depths below 80 cm (31.5 in).
•EMBUS 67
SodLl pho^jwruB factions percentage of total phosphorus oonlenL of
representative profile froa Block D, 22It576. _ _
Level
Depth (cm)
. . . ■
Phosphorus Fractions
Non-occluded
(Noc)
Occluded
(Oc)
Calcium
(Ca)
Organic
(0)
- percentage-
1
0-10
49.0
40.7
5.9
4.3
2
10 - 20
54.3
39.6
3.1
3.0
3
20 - 30
48.6
37.1
12.2
2.1
4
30 - 40
67.8
26.8
2.8
2.6
5
40 - 50
57.8
37.9
2.2
2.1
6
50 - 60
45.0
41.7
11.6
1.7
7
60 - 70
43.1
39.6
15.7
1.6
8
70 - 80
66.1
20.6
11.3
2.0
9
80 - 90
66.7
30.2
0.9
2.2
10
90 - 100
78.4
18.7
0.7
2.2
11
100 - 110
71.8
26.0
0.3
1.9
12
110 - 120
68.9
28.9
0.3
1.9
I
MINERALCGICAL DESCRIPTION
Kaolinite was the dcminant clay mineral with lesser amounts of illite,
vermiculite-chlorite integrade, smectite, and quartz. The different layers
did not ejdiibit dissimilar minercil suites, but had a hi^ degree of similarity
as might be expected, since they had a ccrmon fluvial parent material.
Hcwever, vermiculite-chlorite integrade contents decreased with depth, and it
was not detected in the underlying paleosol liorizons. The absence of
chlorite-vermiculite in the paleosol relative to the superjacent horizons
suggests greater weathering in the paleosols. This finding agrees with data
of Rute et al. () , vho studied paleosols in Indiana.
The sand and silt fractions were dcminated by qucu:t.z with lesser amounts
of mica, glauconite, and feldspar.
micro-morphological DESCRIPTION
Examinations of natural soil peds and separated sand fractions were dcxie
for selected samples via conventional li<^t microsctpy. Representative
samples of the dark-colored, humic-stained upper layers and subjacent paleosol
layers were examined in detail.
170
EPIDEDON
Hiie skeletal matrix (sand and silt) was ccnprised dcminantly of sand
grains v^ch were uniformly coated with reddish brown and black organic
matter. Hie individual coated sand grains were bound or cemented by hunus
into clusters or dcmains \diich formed moderate to strraig granular structure.
Hie structural aggregates tended to wet very slowly, initially repelling the
introduced water, suggesting a hig^ c^illary contact angle. Hie structural
units viere resistant to disruption by water and persisted vAien r^ieatedly
ininersed in water and dried. H^ organic coating was not removed by prolonged
shaking in water. Shaking the soil in 0.05 N NaOH removed the humus coating
reveeiling angular and rounded, clear queurtz grains. Hie sand grains became
loose and single grained after removal of the organic coating, and they tended
to re-wet readily. Hie epipedon of the midden mound presented a striking
contrast in ccmparison to adjacent non-site floodpledn soils. Hie adjacent
soils had very thin (0-6 cm/0-2-4 in) surface layers that were not as dark
colored with Munsell hues of lOYR, and they liad a "salt and pepper" appearance
due to uncoated vhite queu±z grains and unbound black organic detrius. Hie
non-site soils had much less defined structural units and tended to wet more
readily. Hie organic-stained epipedon materials had a distinctive smooth or
"greasy" feel vrfien rubbed between the fingers in contrast to a coarse feel for
the non-site soils.
Mditional tests were conducted to characterize further the nature of the
humus coating. Immersion of the humic-stained soil in H_0 for 12 hours with
sheJdng resulted in no unbound or "free" organic natter, and the structural
dcmains persisted. Treatment of the epipedon with 0.5 N NaCXi and gentle
shaJdng readily removed the humus coating and resulted in a reddish brown
extract. Addition of 0.5 N HCl to the colored extract resulted in conplete
flocculation of the organic material leaving a clear sv:q?ematant indicating
dominantly humic acid compounds (McKeague ). The color ratio of the 0.5 N
NaCM extract solutions was determined to characterize further the coating
materials. Hie color ratio has been expressed as:
E4/E6 = extinction (absorbance) at 400 nm
extinction (absorbance) at 600 nm”
and it has been used for differentiation of humic substances (Tan ) .
Fulvic acids yield spectra with a steep slope in contrast to humic acids.
According to research by Tan () , a low color ratio less than seven
corresponds to hunic acids and related cotpounds with high molecular weii^ts.
A color ratio value greater than seven generally corresponds to fulvic acid
groups.
PALEXDSOL
Microscopic examinations revealed the peileosols ccrmonly have a
finf -grained S-matrix with void argillans and embedded grain argillans
dispersed thrcwghout the matrix. Voids tended to be smaller in the
pcdeoargillic horizons than existed in the overlying epipedons. The paleosol
horizons had a high content of vesiculaur pores that were not connected. Hie
pedeosol had compound structure with prismatic parting to well-developed
subangular bloclcy structure. The polygonal seams between prisms were hi^ly
leached and stripped of fines.
Color ratio veQ.ues for the representative pedon of Block D (Table 68)
indicate dcminantly humic acid compounds. Values gradually increased with
depth but were less than seven.
171
•raBEf 68
OoiLar ratio values of HaCB extracts of tcpcsscntatiuo psion from Block D,
22It576.
Level
Depth (cm)
Color Ratio Valuer absorbance 400 mt
absorbance 600 mt
1
0
-
10
3.73
2
10
-
20
3.83
3
20
-
30
4.04
4
30
-
40
3.89
5
40
-
50
3.83
6
50
-
60
3.68
7
60
-
70
3.95
8
70
-
80
4.34
9
80
-
90
4.47
10
90
-
4.64
11
100
-
110
5.44
12
110
-
6.36
SUM«Rir
site 22It576 occupied a topograjiiic high in the eastern part of the
'harbi^Dee River floodplain about 300 m (990 ft) west of the Pleistocene valley
waill. Part of the site floods during winter and spring months. The v^per
meter (3.3 ft) of the site is dark reddish brown with Munsell hues of SYR, and
it has loany textures. Organic matter contents exceed 1% in the ipper meter
(3.3 ft) and decrease with depth. Hie dark upper strata are underlain by a
lighter colored paleosol vbich contains a paleoargillic horizon. Maximum clay
contents occur in the paleosol, which has well-developed structural aggregates
and macro-structure ejpressed by a distinctive polygonal net^rk, with
polygons separated by seams of leached silt and sand. The thickness and
pedogenic development of the peileosol suggest landscape stability and soil
formation over a substantial time interval prior to burial. There was no
evidence of the ancestral surface horizon of the buried p>cdeosol, vrtiich
suggests fluvial truncation may have occurred prior to subsequent burial.
Cedcium and hydrogen (acidity) were the daninant exchemgeeble cations.
Calcium levels were much hi^er than adjacent off-site soils, aivd contents
varied in the different horizons. Potassium levels were highest in the
surface horizon ./ith a second maximum occurring at depths of 51-97 cm
(20.1-38.2 in). Trace levels of sodium were detected. Exchangeable aluminum
increased in the UTiderJying paleosol suggesting greater weathering and age.
Soil ptfl values ranged from 5.8 in the surface to 6.0 and greater at depths of
87-130 cm (34.3-51.2 in) before decreasing with depth. Soil pH levels were
higher than off-site soils.
Total nitrogen levels decreased with depth and did not correspond to
increased organic carbon levels at different depths. The site had very high
C/N ratios ranging as high as 44/1 at depths of 50-60 cm (19.7-23.6 in). A
very large increase in citric acid soluble P2O,- occurred at depths of 50-80 cm
(19.7-21.5 in). Citric acid (1%) soluble P_Dc^levels exceeded 460 ppm
throughout the sola. These levels cure several hundred fold greater than
values of adjacent off-site soils and reflect cultural additions. Maximum
organic phosphorus levels occurred at depxths of 50-70 cm (19.7-27.6 in). The
site soil contained calcium phosphorus fractions ranging to 338 ppm at 60-70
cm, which is unusual in the highly leached environment and also reflects
cultural additions.
172
Kaolinite was the dcminant cdxsninvxn phyllosilicate clay with lesser
anaounts of illite, vermculite-chlorite integrade, smectite, and quartz. Hie
sand and silt fractions were doninated by quartz with lesser amounts of mica,
glauconite, and feldspar. Hie different layers did not esdiihit dissimilar
mineral suites.
Hhe u{]per meter (3.3 ft) skeletal matrix (sand and silt) was oonprised of
sand grcdns uziifonnly coated with reddish brown and black liumis vAiich imparted
structural, stability to the soil characterized lay granular structure that
persisted against repeated wetting and drying. Analyses indicated the
amorphous hvimis coating material wais caiprised dominantly of humic acid
ccnpounds.
THE ILEX SITE (22It590)
Hie Ilex site was located on a lew fluvial terrace at the juncture of the
southern vadley wall and flooc^lain of Maclcey's Credo. Hie nearly level
terrace appeared to have been joined to the Pleistocene ridge to the south
prior to lateral dissection by fluvial actions. Hie site was 2-3 m (6. 6-9.9
ft) above the active flooc^laiin, and it had a slope of 0-3%.
Hie terrace deposits appeared to be a combination of fluvial sediments and
colluvium/alluvium moved downslope from the adjacent Pleistocene sideslopes.
Althou^ covered with forest vegetation, the steep sideslopes exliibited many
slippage areas.
Steep valley walls with slopes of 25-50% adjoin the site to the south.
Mature, well-developed soils with distinct E Jiorizons and illuviated argillic
horizons comprised the upland areas. Dominant soils were mapped as the
Smithdale association, hilly (Figure 35) . Shiithdale soils tiad red subsoils
that contained up to 35% clay in the argillic horizosns. The upland soils were
highly weathered, strongly acid, cmd siliceous with low bcise saturation levels
(Ultisols) .
Kirlcville and Mantachie soils caiprised the Mackey's Creek flooi^lain
bounding the site. Hiese soils had minimal pedogenic development and had
cantoic (Bw) horizons with little eluviation and illuviation. They were
strongly acid, and surface horizons exhibited scouring and deposition from
recent fluvial action. H«se soils had brD>#n and yellowish larcwn surfaces
with Munsell hues of lOYR, and gray, light gray, or pale brown subsoils with
loany textures.
Hie culturally altered soils of the site developed in loany, siliceous
sediments. Hiese soils were readily distinguished lay the thick, dark reddish
brown, humus-stained epipedons vdiich resulted from prolonged occupation and
cultural activities. Past occupation of the site had drastically modified
normal pedogenic features of color, horizonation, organic matter ccxitent, and
other parameters. Hie soils comprising the upper meter of the site were
distinctly different from adjacent undisturbed soils vdiioh had thin surface
horizons and Munsell hues of lOYR.
Large populations of earthworms, crawfish, rodents and other diverse
microfauna and microflora thrived in the organic-rich mound. The site
elevation above the adjacent floo^lain of Mackey's Creek provided an
excellent habitat for diverse biota. Extensive bioturloation and recent
digging by artifact hunters had tended to mix the upper meter (3.3 ft) of the
site and had retarded normal pedogenic development.
173
SOIL LEGEND
SYMBOL NAME
1 Harleston fine sandy loam
2 Kirkvilte>Mantachie
3 Lexington sill loam. 2 to S percent
4 Mantachie loam
5 Matniston silt loam
6 Smithdaie fine sandy loam. 5 to 6 percent
7 Smithoaie fine sandy loam. 8 to 1 7 percent
6 Smithcaie association, hilly
Figure 36 Soils in the vicinity of 22lt590.
PHYSICftL DESCRIPTION
Uie i:¥)per meter (3.3 ft) of the site soil was dark reddish brcwn and dark
brown (Table 69) with moist hues of SYR and 7. SYR, respectively. These colors
differed from adjacent undisturbed soils vbich had thin surface horizons with
lOYR hues. The mound soil epipedon had a Munsell vcLLue which changed at least
one unit from wet to dry, vbich was in sharp contrast to adjacent soils that
did not exhibit this color change. The dark reddish brcwn ^ipedon had a
distinct "greasy" or slick feel when ruUaed between the fingers. The vpper
horizons had greinular structure and friable consistence. Subsoil colors of
the site soils were brownish yellow to light yellowish brown with lOYR hues.
In places, horizontal lamellae occurred in the deeper strata parallel to the
surface. Tbe lamellae had higher clay contents and were usually brighter in
color. These layers may represent periods of landscape stability and
pedogenesis, or represent water table fluctuations.
TAB[£ 69
Pedon deacription of repreaentative profile, 22It590. _
Depth (cm) Description (moist colors) _
0-15 Dark reddi^ brown (5YR3/2) loamy sand; weak fine granular
structure; friable; sand bridging by hvmus; ccmnon fine and
medium roots; conmon charcoal fragments; gradual wavy boundary.
15-30 Dark reddish brown (5YR2.5/2) sandy loam; weak fine granular
structure; friable; intergranular bridging by humus coating;
conmon fine and medium roots; conmon krotovinas and worm casts;
few charcoal fragments; gradual wavy boundary.
30-50 Dark reddisii brown (5YR3/2) san^ loam; weak fine granular
structure; friable; conmon fine roots; gradual wavy boundary.
50-59 Dark reddish brcwn (5YR3/2) sandy loam with few fine faint
mottles of light yellowish brcwn (10YR6/4) ; weak fine granular
structure; friable; conmon fine roots; clear wavy boundary.
59-69 Dark brcwn (7.5YR3/2) sandy loam; weak fine granular structure;
friable; few fine black fragments (10YR2/1) in lower part of
horizon.
69-79 Dark brown (7.5YR4/2) sandy loam with conmon fine faint brcwn
mottles (7.5YR4/4); weak fine structure; slightly firm in place,
friable vdien disturbed; conmon vesicular pores oriented
perpendicular to vertical direction; common black charcoal
fragments; gradual wavy boundary.
79-116 Dark brown (7.5YR4/2) sandy loam with cortmon fine and medium
faint mottles of strong brcwn (7.5YR5/6) , and reddish yellow
{7.5YR6/8); weak fine subangular blocky structure; firm in
place, slightly friable vdien disturbed; many vesicular pores
oriented perpendicular to the vertical direction; common black
charcoal; gradual wavy boundary.
I
A
I
J
176
^ ■ ■
brown (10YR6/6) loamy sand; massive in place
breaks to single grain v^hen disturbed; loose; sand grains
cxninonly strij^ed of coatings; slightly firm in place; clear
wavy boundary.
144-203 Brownish yellow (10YR6/6) loam with cannon medium distinct
mottles of yellow (10YR7/8); yellowish brown (10YR5/8) and
strong brown (7.5YR5/8); massive in place, parting to weak fine
svbangular blocky structure vihen disturbed; no evidence of
cultural alteration; gradual wavy boundary.
203-232 Mottled yellow (10YR7/6) and p2Q.e brown (10yR6.3) loany sand;
ain; loose; no evidence of cultural alteration.
Particle size distributions (Tables 70 and 71) reflect textured,
stratification resulting from different fluvial depositions and p rhaps
reworking action. Textures ranged from sand to loam. There was some textural
variation across the site eis shown hy data. Sand contents were higher on the
southern part of the site adjoining the steep sand^ slqpes of the Pleistocene
valley wall. The sand fraction was daninated by fine sand (0.25-0.10 mn) and
median sand (0.5-0.25 mn) . The fine-sand fraction was higher* cxi the southern
portions of the site adjacent to the inlands. Silt contents tended to be
greater in the center and northern portions of the site. Clay contents were
variable with maxinum levels ranging to 15%. There was little evidence of
illuviation.
TRBUB 70
Particle size distribution of selected soil
Lve, 22It590.
Fraction
locntiiMBai .
Sand
Very Coarse
Coarse
Medium
Fine
Very Fine
Depth
(2-1 mm)
(1-.5 nin)
(.5-. 25 nn)
(.25— .10 mn)
(.10— .05 Rin)
on “
50-59
.04
.20
18.50
42.90
7.60
59-69
.04
.20
22.40
45.50
6.00
69-79
.10
.30
18.70
41.70
7.20
79-116
.03
.30
17.40
39.60
7.60
116-144
.04
.40
24.20
50.40
6.40
144-203
.10
1.40
14.10
25.20
5.20
203-232
.03
.30
41.70
38.90
2.60
1MLE 71
Barticle size distribotion of repci
tentative soil pedon at tlie edge of Sits
Sand
Hit
Clay
Depth
'
cm
0-7
88.4
6.4
5.2
sand
7-12
89.3
5.7
5.0
sand
12-20
88,7
6.5
4.8
sand
20-75
86.8
3.4
10.2
loamy sand
75-97
84.5
7.9
7.6
loany sand
97-140
85.6
11.0
3.9
loany sand
Sand Fraction
Very Coarse
Coarse
Medium
Fine
Very Fine
(2-1 nin)
(1-.5 nin) (.5-. 25 m)
(.25-. 10 m)
(.lO-.OSiiin)
an —
— percentage —
0-7
0.6
0.3
15.2
63.3
9.0
7-12
0.4
0.4
16.3
63.9
8.3
12-20
0.2
0.3
14.7
64.6
8.9
20-75
0.1
3.1
18.1
62.4
3.2
75-97
0.2
0.6
25.7
52.5
5.5
97-140
0.1
0.1
23.7
53.4
7.8
CHEMICAL DESCRIPnON
Calcium and hydrogen (acidity) were the dominant exdiangeable cations
(Tables 72 and 73) . Calcium contents varied across the site vdth lower values
occurring cxi the southern part of the site. In the represoitative site pedon
(Table 72) , calcium levels varied with depth, and maxinun values occurred at
depths of 69-116 cm (27.2-45.7 in). Ihe calciun levels were several hundred
fold higher than undisturbed adjacent soils. Ca/Mg ratios varied with depth
and across the site. Potassiim levels were hipest in the surface layers.
Trace levels of sodium were present. A large increase in eiodiangeable
alixninixn occurred at depths of 20-75 cm (7.9-29.5 in) in the pedon on the
southern portion of the site (Table 73) .
an
onol
0-15
4.9
2.58
0.37
0.26
0.04
9.93
0.61
13.18
24.6
15-30
5.7
5.37
0.59
0.07
0.02
8.09
0.01
14.14
42.7
30-50
5.8
6.34
0.42
0.05
0.03
6.49
0.01
13.33
51.3
50-59
5.8
5.99
0.55
0.04
0.03
6.49
0.00
13.55
48.7
59-69
5.6
6.72
0.66
0.05
0.04
5.79
0.01
13.26
56.3
69-79
5.7
8.72
0.80
0.07
0.03
6.79
0.01
16.41
58.6
79-116
5.6
9.47
0.56
0.05
0.03
8.27
0.07
18.38
55.0
116-144
5.7
3.40
0.21
0.04
0.02
3.06
0.03
6.73
54.5
144-203
5.5
8.10
0.72
0.14
0.04
3.90
0.42
12.90
69.7
203-232
5.6
5.90
0.47
0.11
0.03
2.33
0.27
8.84
73.6
TABLE 73
riiMtlnil dUKacteristicB of zeixesentative soil pedcn at the ed^e of
Site 22It590. _
Exdiangeable Cations B2ise
Dep^ pH _ Ca _ Mg _ K Na _ H _ Al Tot£d. Satiaration
(an) - - - - onol (p4-l kg-1 - - % -
5.3
2.63
0.74
0.15
5.23
8.76
7-12
5.5
2.36
0.74
0.12
4.83
iSI
8.07
40.1
12-20
5.3
1.76
0.71
0.10
0.02
4.35
iia
6.94
37.3
4.8
1.19
0.63
0.08
0.02
6.79
3.23
8.71
75-97
4.9
■SB
0.81
5.25
2.56
6.64
97-140
5.2
mSm
0.59
3.16
52.5
llie exchangeable alvmirnin ccxitents were similar to levels occurring in the
subsoils of adjacent upland soils, suggesting similar materials. Cation
exchange capacities (Table 73) were greater in the representative pedon cm the
site (Table 72) suggesting possible differences in mineralogy. Bcise
saturation exceeded 40% except for the surface 0-15 on (0-5.9 in) in the
representative pedcn.
Soil levels varied vdth depth and distance across the site (T^les 72
and 73) . Values were sli^tly hi^ier than adjacent undisturbed soils.
Organic matter contents were greatest in the surface horizons. However,
variations occurred with distance across the site (Table 74) . Organic matter
exceeded 1% in the v^jper meter (3.3 ft) and tended to coincide with the daricer
color. Total nitrogen levels decreased with depth. Ihe C/N ratios tended to
increase with depth, and decreased at one meter (3.3 ft) . levels
(soluble in 1% citric acid) exceeded 250 ppnn in the upper 60^cft (23.6 in) of
the pedon 6uicdyzed, with levels decreasing with depth to 184 ppm at 100 on
(39.4 in) (Table 75). Uie total and organic phosphorus distribution was
variable and difficult to interpret (Table 76) . Free iron-oxide contents
exceeded 1% in the v^aper meter (3.3 ft) with maximum values occurring at
69-79 cm (27.2-31.1 in).
74
Ooganic
Gnbon, nitrogen contents, oBKtxa/nitxogen ratio, and 1%
WliW 111* Ok
Level
Organic
Depth Matter
C
C/N
1% Citric Acid
Soluble P205
-cm-
— ^percentage -
Fpm -
2
10-20 1.10
0.64
0.047
13.6/1
509.6
3
20-30 1.13
0.65
0.032
20/1
508.6
4
30-40 1.17
0.68
0.025
27/1
467.9
5
40-50 1.27
0.74
0.024
30.8/1
428.5
6
50-60 1.48
0.86
0.022
39/1
388.9
11
100-110 0.38
0.22
0.018
12/1
184.43
TMUZ 75
Gcgnic I
Mtter, free iron aside, total and organic phoaphoms ocsiLents and pH
Depth
Organic Matter
Fe203
cm
- percentage —
— ^ppm -
0-15
3.00
1.23
353.5
117.0 4.9
15-30
1.25
1.65
230.8
21.0 5.7
30-50
1.33
1.51
427.7
158.4 5.8
50-59
1.53
1.65
277.2
5.8
59-69
1.13
1.65
250.8
10.7 5.6
69-79
1.20
2.17
184.2
5.7
79-116
1.40
1.97
286.7
109.9 5.6
116-144
0.36
0.67
267.0
178.0 5.7
144-203
0.10
1.08
110.5
20.5 5.5
203-232
0.07
0.58
58.5
5.6
TMUS 76
Organic i
Bitter, oazlion, total nitrogen contents and cazbon/nitrogen ratio of
raoresentative padcn, 22It590.
Depth
C
N
c7n
cm
percentage
■
0-15
2.75
1.59
0.172
9.2/1
15-37
0.73
0.42
0.047
8.9/1
37-60
0.96
0.56
0.035
16.0/1
60-100
1.38
0.80
0.032
25.0/1
100-150
1.03
0.60
0.029
20.7/1
150-180
0.46
0.27
0.023
11.7/1
180-195
0.09
0.05
0.012
4.2/1
195-250
0.06
0.03
0.011
2.7/1
250-275
0.08
0.05
0.014
3.5/1
MINEBALOGICAL DESCRIPTION
Kaolinite was the doninant clay mineral with lesser amounts of illite,
vermiculite-chlorite integrade, snectite and quartz. Trace amounts of
smectite were detected in the clay fractions of the materieds firm 0-30 cm
(0-11.8 in). However, smectite contents increased with d^)th and reached
179
naxiimn levels betMsen 69-144 on (27.2-56.7 in) depth. Sand and silt
fractions were daninantly quartz with tr2K3e amounts of mica, glauconite, and
feld^ar.
SIMAFY
Site 22It590 was located on a low terrace at the juncture of the southern
valley wall and flooc^lain of Madcey's Credc. The site was 2-3 m (6. 6-9. 9 ft)
alxve the active floodplain, and it had a slope of 0-3%. Ihe site soil was
distingui^ied by thick, hun^-rich dark reddish brown loany sand and sand
vesper layeze. Ihe dark-colored epipedcn had Mansell oolors of SYR which
contrasted with adjacent off-site soils vHtiich had thin surface horizons with
lOYR colors. Sand contents were hi^^r on the southern part of the site
adjoining the steeply sloping Pleistocene uplands. Little evidence of clay
illuviation was detected in the site soil. Ihe dark-colored surface layers
graded to lifter colored subsoils with 7. SYR and lOYR colors. Organic matter
contents were greater in the sxirfaoe horizons and esoeeded 1% in the upper
meter (3.3 ft) .
Cedciim and hydrogen (acidity) were the dominant exchangeable cations of
the site soil. Calcium levels were enriched in the site soil well above
levels of adjacent undisturbed soils. Calcium levels tended to decrease with
distance frcxn the center of the site. Potassium levels were hi^iest in the
surface l2yers of the site soil. Increased levels of exchangeable alizninxxn
occurred in subsoil layers in the southern edge of the site bounding the ste^
uplands, suggesting similarities in soil parent materials. Trace levels of
sodium were detected with no aocunulations. Soil pH varied with depth and
location in the site and was slightly higher than adjacent soils.
Total nitrogen levels decreased with depth. The C/N ratios varied with
depth and differed from adjacent soils. Soil horizons in the upper 60 an
(23.6 in) of the site contained greater than 250 ppm citric acid (1%) soluble
P-0_, vdiich is one criteria for anthropic epipedcwis. Total and organic
pno^horus distribution weis variable and difficult to interpret.
Kaolinite was the dcminant mineral in the clay fraction with lesser
amounts of illite, vermiculite-chlorite integrade, smectite, and quartz.
Smectite contents increased at depths belcw 30 cm (11.8 in) .
SITE 22IT606
Ihe site is located in an upland position on a Pleistocene terrace vhich
has loeen incised by the Torbi^see River and Mud Cre^. The mature landscape
had been partially isolated on the east, south, and west resulting in a
peninsula-sheped surface (Figure 37) . 'Hie incision resulted in steep-sided
slopes with a pronounced escarpment on the southern terminus. Hie surface was
located above the floo(^rone area of the Tatbi^oee River, and it was not
subject to flooding. Recent cultivation and ooci?>ation of the site had
resulted in sane sheet or rill accelerated erosion of the surface horizon.
Hie area had a slope of 0-2%.
Hie lower-lying adjacent soils in the flooc^lain were scmesdiat poorly
drained, loany textured with weak subsoil pedogenic development. Hie surfaces
were dark grayish brcwn with gray subsoils in Munsell hues of lOYR. Hiey were
subject to seasonal flcxxJing. Hie related upleuid soils of similar elevation
on the Pleistocene surfaces were deep, mature soils with definitive yellow-red
180
22IT606 I
Itawamba County, Mississippi /
Mr n I
jpfl>tl» . j 1 1 jl I
i ^ t I I
I \|22ITe06 / // I' h I
7 f I
Co. lutawamba lit i 1
i ^ / // (s ! I
Monroa^.^^^ ^ jj I
' '
1 1 I
' I
: !
I
I
I
I i
I X ,8ta. ^00
/ r '
i
\
'A)
Normal Pool 300 ft
I Beaver Lake /~-o>^Beaver Lake
[ Recreation Araai \\\ /
i
•- ' IW I •
• u y
i k
ly:
\y
-x.^320/>
Contour Interval : lOft
Note; Map from US Army C^'-rps of Engineers Tennesse Tombigbee Walefvray
Cana} Section General Plan (Design Memo e5), Sheet e 16. Pool D.
Figure 37 Site location of 22It606 and 22lt622.
subsoil argillic horizons esdiibiting strong peckigenic develcfxnent. The t^laiid
soils were very strongly acid, hi^y weath^^i and silioeous with low base
saturation (Ultisols) . Some of the soils contained fragipen horizons at
depths of 50-75 cm (19.7-29.5 in).
The floodplain soils adjacent to and below the site were daninantly within
the Kirkville (ooarse-loeny, silioeous, thermic Fluvaquentic Dystroohrepts)
and Mantachie (fine-loany, siliceous, acid thermic Aerie Fluvaquents) series.
These soils had minimal pedogenic development with loeny textures esdiibiting
stratification. They had cambic (Bw) horizons with little eluviation and
illuviation. They were subject to flooding during the winter and spring
months.
The soil at the site had developed in Coastal Plain sediments and was very
representative of mature soils of the area containing an argillic horizon.
The upper part of the sola had been culturally altered isparting a darter
color than occurs in such soils in nature. Itooont cultivation had resulted in
mixing the upper horizons. Cultural disturbance was limited to horizons
located above the argillic horizon. The site soil had a well-developed
paleosol at greztter depths.
PHYSICAL DESCRIPnON
A representative pedm at the edge of the site, with cultured, material in
the tpper 27 cm (10.6 in), had well-developed pedogenic characteristics (Table
77) . The x^per 27 cm (10.6 in) had darter colors in Munsell hues of 7. SYR
which differed fran undisturbed soils that had thinner surferm horizons with
lOYR hues. The natured. soils also had thin, lic^iter colored E liorizons in
areas that had not been plowed or culturally disturbed. A wll-developed
argillic (Bt) svhsoil horizon occurred at the site with most of the artifacts
and evidence of cultural disturbance located above and in the vpper part of
the horizon. This indicates landscape stability, and pedogmiic development
had resulted in the farmation of the mature soil prior to habitation of the
site. Subsequent habitation and cultural disturbance affected the horizons
above the argillic horizem, vhile the lower argillic frem 27 cm (27.6 in) and
deeper was not appreciably affected, and horizonation was readily disoemible.
The argillic horizon had subanguleu: blodcy structure with oriented clay skins
on ped faces.
ve oadOD. 22It606
Descr
XDwn (7.5YR4/2) sandy loam;
structure; very friable; cxuroDi'i cultural fragments;
ocnnpn fine and medium roots; few charcoal flakes;
gradual wavy boundary.
Strong brown (7.5YR5/6) sand^ loam; weak fine sub-
angular blocky structure; friable; patchy clay skins
on ped faces; few charcoal flalces and cultural
fragnents; cxintcm fine and median roots; gradual
wavy boundary.
Horizon Depth (cnu Description (moist oolor^
Bt:^
27-79
Yellowish red (5YR4/6) loam; moderate fine subangular
blocky structure; sli^tly firm in place, friable vrfien
disturbed; clay skins on ped faces; gradual wavy
boundcury.
Ij
Bt3
79-92
Strong brown (7.5YR5/6) semdy loam; moderate fine
subangular blocky structure; sli^tly firm in place,
friable vhen disturijed; cl^ slcins on ped faces;
gradual wavy boundary.
a
Bt4
92-115
Mottled reddish yellow (7.5YR6/6) and yellowish red
(5YR4/8 and 5YR5/8) sandy loam; firm in place, friable
vdien disturbed; patchy clay slcins on ped faces; gradual
wavy boundary.
2Btbl
115-140
Yellowish red (5yR5/6) sant^ clay loam with common
medium distinct reddish yellow (7.5YR6/6) , red (5YR4/8)
and pcile brown (10YR6/3) mottles; moderate coarse
prisnetic parting to moderate medium subangular blocky
structure; firm; gradual wavy boundary.
2Btb2
140-180
Yellowish red (5YR4/8) sandy clay loam with man^’
medium distinct red, reddish yellow, strcaig brown
and pale brown mottles; moderate cx>arse prismatic
parting to moderate ne^um subax^lar blodcy structure;
firm; continuous clay cxtatings cai ped faces; leached
seams between prisms form a polygonal network; gradual
wavy boundary.
2Btb3
180-250
Yellowish red (7.5YR5/8) san(fy loam with cxrman strcng
brown (7.5YR5/8) and pale brown (10YR6/3) mottles;
moderate medium subaxgular blocdcy strucrture; firm;
continucxis clay coatings on ped faces; abrupt smooth
boundcUY-
3C
250+
Strong brown (7.5YR5/6) loaiiy sand; Icose; single
grained.
A well-developed paleosol occurred at a depth of 100-150 on (39.4-59.1 in)
with a paleoargillic horizon of sandy clay loam textures. Ihe pcLLeoargillic
horizon had medivm-ooarse prismatic structure vhich parted to noderate medium
subangular blocky structure. Ihe prisms were separated by lighter colored
seams of fine sand and silt v^ch formed a pjolygonad network. Ihe -Jl
paleoargillic horizon rested upon loatiy sand.
Particle size distributions (Tables 78 and 79) indicate slight variations
acrcxss the site with rather uniform textural classes of sandy loam and loam.
Maxiiiun clay contents occurred in the paleocugillic horizon at depths of
140-180 cm (55.2-70.9 in). The pedeosol graded abruptly to loamy sand at
depths of 250 am (98.5 in) . The sand fraction was dcminated by fine and
medium sand with lesser amounts of very fine Scind. Ihe very fine sand
ocaitents were higher in the paleoargillic horizon. There was an abrupt
increase in the medium sand fraction in the underlying C horizon. j
183
IH9i
Sand Silt
(2-0.05 itm) (0.05-0.002 nin)
Clay
« 0.002 nm)
Texture
on
0-7
57.7
39.1
3.2
sandy loam
7-18
57.3
39.6
3.1
san<^ loam
18-47
38.5
48.5
13.0
loam
47-76
25.8
55.7
18.5
silt loam
76-100
39.1
44.8
16.1
loam
100-125
57.6
18.0
24.4
sandy clay loam
Sand FreK^tion
Very Coarse
Coarse Meditm
Fine
Very Fine
(2-1 mm)
(1—0.5 nm (0.5-0.25 nm) (0.25—0.10 imi) (0.10—0.05 nin)
on
0-7
1.0
1.1 9.3
35.7
10.5
7-18
0.3
0.9 11.3
34.1
9.7
18-47
0.1
0.5 5.7
23.6
8.6
47-76
0.4
0.6 3.8
15.0
6.1
76-100
0.3
0.6 7.5
23.1
7.5
100-125
0.0
0.0 5.1
40.0
12.5
TMttZ 79
Sarticle size distribution of reixesentative pedon, 22It606.
Sand
Silt
Clay
(2-0.05 irm)
(0.05-0.002 nm)
« 0.002 mm)
Texture
an
0-10
59.0
38.2
2.8
sandy loam
10-27
58.6
33.6
7.8
sandy loam
27-79
46.3
38.4
15.3
loam
79-92
62.8
22.0
15.2
sanc^ loam
92-115
65.5
15.8
18.7
sandy loam
115-140
60.3
19.2
20.5
semdy clay loam
140-180
56.4
17.4
26.2
sandy clay loam
180-250
71.5
10.6
17.9
sancfy loam
250+
87.1
4.7
8.2
loany sand
Sand Fraction
Very Cocurse
Coarse Medium
Fine
Very Fine
(2-1 nin)
(1-0.5 nm) (0.5-0.25 nm) (0.25-0.10 nm) (.10-. 05 nm)
cm
. _ _ • _
0-10
2.3
0.6 .7.2
32.4
6.5
10-27
0.2
0.2 8.7
43.0
6.5
27-79
0.1
0.3 11. 0
28.7
6.3
79-92
0.1
0.2 33.3
25.8
3.4
92-115
0.0
0.3 10.0
45.9
9.6
115-140
0.0
0.4 6.0
41.9
12.3
140-180
0.0
0.0 4.4
39.2
12.8
180-250
0.0
0.0 8.6
52.1
10.7
250+
0.0
0.3 42.4
41.5
2.9
184
The thickness of the paleoaarglllic horizon and the well-developed
structure and oriented clay skins on ped faces suggest the paleosol foxined
over a long period of landsce^ st£^ility. The prianatic structure and
polygonal nebn^iidc appear to be relic pedogenic features.
cHagcaL DEscRipnow
Calcivm and hydrogen were the doninant exohangeable cations (Tables 80 and
81) . Calciun levels were greatest in the surface and reflect liming or
cultural additions. Exchangeable Mg contents increased with depth and reached
maximm levels in the paleosol. The Ca/Mg ratios also decreased with depth
and were less than one in the paleoargillic horizon which reflects the greater
weathering and age. Poteissiutn levels were hic^ier in the upper lexers. Trace
levels of Na were detected. Exchangeable A1 increased with dep^ and reached
maximum levels in the paleoargillic horizon idiich cdeo indicates greater
weathering and age in the paleosol. Maxisun cation exchange capacity values
(Total, Tables 80 and 81) occurred in the paleoargillic horizon and coincided
with maxiiiun cley content. Base saturation levels were higher in the surface
layers and reflect cultural additions of CaCO^.
TMI£ 80
rtwriral ctMogcbetisties of roisjciieptative pedan, Tl^ Chit 106/»l# 22It606.
Organic _ Exchangeable Cations _ Base
pH
Matter
Ca
K
H
msm
tkTrT^^IIH
Saturation
cm
A
■■■■■■
*
0-7
5.2
2.23
1.80
0.67
0.49
0.04
7.90
0.54
10.90
27.5
7-18
5.2
1.24
1.61
0.59
0.29
0.02
6.07
0.48
8.58
29.2
28-47
4.8
0.47
0.92
0.75
0.37
0.10
7.08
1.92
9.22
23.2
47-76
4.6
0.23
1.09
0.74
0.50
0.03
9.35
3.97
11.71
20.1
76-100
4.4
0.11
0.94
0.97
0.18
0.04
7.63
3.52
9.76
21.8
100-125
5.2
0.10
0.47
1.14
0.09
0.14
10.82
6.55
12.66
14.5
TMU 81
Depth
pH
Organic
Matter
_ Exchangeable Bases
Base
Saturation
Ca
Mg
K
Na
H
A1
Total
cm
%
.1
a
0-10
5.2
3.48
3.59
0.68
0.37
0.01
6.84
0.12
11.49
40.4
10-27
5.1
0.76
1.13
0.49
0.28
0.02
3.65
0.60
5.57
34.4
27-79
5.0
0.32
1.30
0.79
0.20
0.04
5.71
1.93
8.04
28.9
79-92
4.8
0.09
0.69
0.86
0.14
0.04
6.25
2.66
7.98
21.6
92-115
5.0
0.09
0.57
1.09
0.21
0.05
7.68
4.18
9.60
20.0
115-140
5.1
0.11
0.33
0.69
0.11
0.10
7.71
4.16
8.94
13.8
140-180
5.0
0.10
0.44
1.04
0.10
0.14
10.33
6.05
12.05
14.3
180-250
4.8
0.08
0.39
0.73
0.05
0.07
7.59
4.76
8.83
14.0
25(H-
4.8
0.08
0.22
0.39
0.07
0.04
4.38
2.60
5.10
14.1
Soil ^ levels varied with depth and tended to decrease at the lower
depths. Levels did not differ greatly from natured soils that had not been
disturbed. Organic matter ccxitents were hi^iest in the surface horizons and
decreased with d^>th. Levels varied scncMhat across the site.
Total nitrogen decreased with depth and generally was related to the
organic matter contents (Table 82) . The C/N ratios were similar to levels of
natural soils of the area.
•EMU 82
Organic cazfaon, nitrogen, cazbcn/nitrogen ratio, and 1% citric acid soli±iIe
P205 contents of repreaentative pedon, 22It606. _ _ _
1% Citric Acid
C
N
C/N
Soluble P205
(on)
- ^percentage -
ppm
0-10
2.01
0.182
11.0/1
343.1
10-27
0.44
0.043
10.2/1
0
27-79
0.18
0.031
5.8/1
0
79-92
0.05
0.026
1.9/1
0
92-115
0.05
0.026
1.9/1
0
The surface layer contained more than 250 ppm P_0_ (soluble in 1% citric
acid) , but none was detected at lower depths (Table "^83) . The P_0(. levels of
the surface horizon eure related to occt^jatiOT and cultured additions.
MINERALOGICAL DESCRIPTION
Kaolinite was the dcminant clay mineral with lesser amounts of illite,
vermiculite-chlorite integrade, and quartz in the vpper sola from depths of
0-115 an (0-45.3 in) . However, a change was detected in the clay mineral
suite in the paleosol by X-ray diffraction and differential thermal analyses
vMch indicated the presence of smectite and gibbsite in the paleoargillic
horizon. The mineral suite in the paleosol between the depths of 115-250 on
(45.3-98.5 in) consisted of )caolinite with lesser amounts of illite, smectite,
gibbsite, and quartz. Gibbsite contents reached a maximum at depths of
140-180 cm (55.2-70.9 in) and cotprised 10-15% of the clay minerals at that
depth.
SimABY
Site 22It606 was located in an ipland position on a Pleistocene terrace
necu: the ccnfluence of the Tonbic^nee River and Mid Cre^. The site wsis above
the floo^lain and it was not subject to flooding. The site soil was altered
in the surface horizons by occipation and cultur^ additions which imparted a
darlcer color with Munsell hues of 7.5YR. The subjacent argillic horizcai was
relatively unaffected by cultural changes, and it had strong brown (7.5YR hue)
and yellcwish red (51® hue) colors. Soil morphology indicated eluviation and
pedogenic develcproent of the argillic horizon, and sola were ccnpleted prior
to hvman occipation and disturbance.
A well-developed p>aleosol with a thick pedeoargillic horizon occurred at
depths belcw 115 on (45.3 in) . The paleosol was yellcwish red with a 5YR hue
and had cotpound structure of prismatic parting to sul>angular bloclcy. The
prismatic structural units were separated by lifter colored, leached seams of
silt and sand vhich formed a polygcwial network. The polygonal nacro-structure
c^peared to be a relic pedologiccil feature. Pau:ticle size distribution was
rather uniform over the site with naxlmum clay ccaitents occurring in the
pedeoargillic horizon. An abrupt increase in the medium sand fraction
occurred in the underlying C horizon at d^>ths belcw 250 cm (98.5 in) .
Calciun and t^drogen (acidity) were the doninant exchangeable cations of
the site soil. Hi^ier Ca levels in the surface harizcn reflected agricultural
additions of lime and/or other cultural additions. 'Rie Ca/Mg ratios decreased
with dqyth reflecting greater weathering and age. Exchangeable aliminun
increased with depth and readwd roaxinun levels in the paleoargillic horizon
an indication of weathering and age. Soil pH levels varied with depth but did
not differ gireatly from non-site auijacent soils.
Total nitrogen decrecuged with depth and was related to organic matter
contents. Ihe C/N ratios were similar to natural soils of the area. Ihe
surface layer contained more than 250 ppn (soluble in 1% citric acid) ,
and none was detected at lower depths, indi&ting the cultural additions were
primarily in the surface. Ihe paleosol had a different mineral suite
ch2u:eK;terized by the presence of anectite and gibbsite in the clay fraction
v^ch was not detected in the upper sola.
'IHE MUD CREEK SITE (22It622)
The Mud Creek site is a small rise in the floodplain appraximately 500 m
(1,650 ft) south of 22It606 at the junction of the Mud Creek and The Tcrabi^aee
River floodplains (Figure 35) . Mud Cre^ flows noirth of the site
approximately 400 m (1,320 ft) . Ihe site rises over a meter above the general
elevation of the surrounding flcn^lain and is rou^ily circular in outline
plan. The flanJcs have a 1-2% angle of slope.
The lower-lying adjacent soils in the flooc^lain were scmewhat poorly
drained, loamy textured with weak subsoil pedogenic development. The surfaces
were dark grayish brown with gray subsoils in Munsell hues of lOYR. They were
subject to seasonal flooding. The related upland soils of similar elevation
on the Pleistocene surfaces were deep, mature soils with definitive yellcw-red
subsoil argillic horizons eidiibiting strong pedogenic development. The upland
soils were very strongly acid, hig^y weathered, and siliceous with low base
saturation (Ultisols) . Some of the soils contained fragipan horizons at
depths of 50-75 cm (19.7-29.6 in).
The floodplain soils €KijaK:!ent to and below the site were dominantly within
the Kirkville (coarse-loany, siliceous, thermic Fluvaquentic Dystrochrepts)
and Mantachie (fine-loany, siliceous, acid thermic Aerie Fluvaquents) series.
These soils had minimal pedogenic development with loany textures esdiibiting
stratification. They had canbic (Bw) horizons with little eluviation and
i.T' aviation. They were subject to flooding during the winter and spring
months.
The culturally edtered site soils are developed in alluvied loam to sanefy
loam deposits. These soils are readily distinguished by the thick, daric,
reddish brown, humus-stained epipedons from honan occupation. Soils in the
upper 50 cm (19.7 in) were distinctively different from adjacent undisturbed
floodplain soils which had thin surface horizons. The sediments had also been
severely disturbed by bioturbation. Most stratigraphic boundaries are gradual
and wavy. Six identifiable strata were reccjgnized. All of these strata were
above the dry-season vmter table.
PHYSICAL DESCRIPTION
Particle size distributions (Table 83) indicate textural stratification.
Sanity Icjam textures cxxnirred throu^iout the representative pedon, except for
the 50-108 cm (19.7-42.6 in) depth, vhich had a loam texture cxwitaining the
highest clay (14.8) and silt (46.3) contents and the lowest sand (38.9%)
content. OSie sand fracticxi vios daninated by fine (0.25-0.10 mn) and laediiin
(0.5-0.25 mn) sand, with lesser amounts of very fine sand, and trace amounts
of coarse and very coarse sand.
W 83
Rnrticle siae distribution of repp
Bssntactivm
psdon. 22Itfi22.
snt
(2-0.05 nm)
(0.05-0.002 mn) (< 0.002 nm)
Texture
cm
~'p6ru6nu£iige'
0-15
71.1
25.9
3.0
sandy loam
15-29
65.9
26.2
7.9
s^uxty loam
29-50
53.1
36.8
10.1
sandy loam
50-108
38.9
46.3
14.8
loam
108-122
72.5
22.5
5.0
sandy loam
Sand Fracticm
Very Coarse
Coarse
Medium
Fine
Very Fine
(2-1 nro)
(1-.5 nm)
(.5-. 25 mn)
(.25— .10 mn)
(.10-. 05 mn)
on
_
UCSi J - — —
0-15
0.30
0.80
25.90
38.40
5.80
15-29
0.10
0.20
25.00
36.20
4.50
29-50
0.03
0.10
16.00
32.60
4.30
50-108
0.03
0.10
12.70
22.70
3.60
108-122
0.03
0.40
29.40
38.50
4.20
CHEMICAL DESCRIFTIGN
Soil pH levels decreased with depth (Table 84) and ranged frcm 5.9 in the
surface layer to 4.6 in the 108-122 cm (42.6-48.1 in) dqith. Only the surface
layer (0-15 an or 0-5.9 in) had pH levels hi^ier than adjeK»it off-site soils.
CaQcivxn and hydrogen (acidity) were the dcminant exchangeable cations.
Cedcivin levels decreased abn:ptly with increasing depth with the surface
horizon containing levels severed hundred fold greater than the subjacent
layers, vAiich suggests liming and/or cultural additions. Soil magnesitm
ocxitents also decreased with depth. Ca/Mg ratios varied with depth and ranged
frcm 8.39 in the surface to 1.5 at deptim of 108-122 on (42.6-48.1 in).
Potassium levels were sli^tly hi^ier in the 29-50 an (11.4-19.7 in) layer and
decreased abruptly at a depth of 108 cm (45.6 in) . Trace levels of sodiixn
were present. Exchangeable aluminum had maximum valxies in the 50-108 cm
(19.7-45.6 in) layer vhich corresponded to the higher clay content. The
cation exchange capacity v^dues were low indicating a doninance of kaolinite
and 1:1 phyllosilicate minerals.
TMES 84
ClwW fil characteristics of reggesaatative pedop, 22It622.
_ Exchangeable Cations _
Depth
rfl _ Ca
Mg _ K
Na H A1 Total
Base
Saturation
84
nwrinwl cfaaracLeristics of rcfTroBentative pedon, 22It622 (cxwtinnBd)
Exchangeable Cations
Base
Depth
pH
Ca
K
Na
H
A1
Total
Saturation
cm
■ Qnol
(pt)
kg-1
percentage
29-50
4.5
0.64
0.10
0.22
0.03
6.23
1.87
7.22
13.71
50-108
4.6
0.90
0.25
0.17
0.03
7.58
3.73
8.93
15.12
108-122
4.6
0.24
0.16
0.06
0.05
3.22
1.48
3.73
13.67
Organic matter content exceeded 1.3% in the surface horizon and decreased
abixptly with increasing d^>ths to levels less than 0.3% in subjacent layers
(Taddle 85) . Total nitzxsgen levels etlso decreased with depth, decreasing from
0.081% in the surface layer to 0.012% at a depth of 108 cm (45.6 in) . Uie C/N
ratio decreased with depth to very low values in the de^er strata v^iere only
trace levels of C and N were detected.
85
(kganic matter, carbon, nitrogen contents, cazben/nitrogen ratio, and 1%
citric add aolifcle P205 contents of representative peden, 221t822.
Organic
Matter
C
N
C/N
1% Citric Acid
Soluble P205
cm
‘percentage -
ppm
0-15
1.33
0.77
0.081
9.5/1
121.20
15-29
0.35
0.20
0.028
7.1/1
121.30
29-50
0.32
0.18
0.029
6.2/1
40.01
50-108
0.16
0.09
0.019
4.7/1
4.07
108-122
0.03
0.01
0.012
0.8/1
0.00
Citric acid extractable P2(}c levels were much less than the minijium level
of 250 ppm hic^iest contents which were slightly higher than levels of
adjacent off-site soils (40 ppm) .
THE BEECH AND OAK SITES (22It623 and 22It624)
Ihe Beech and Oak sites are situated on adjacent levee remnants in the
flooc^ledn of the Tombi^^Dee River Valley. The two sites etre two foagnents of
the same geomorphological and archaeological entity. Located some 8 km (5 mi)
north of Fulton, Ms, the sites are approxiinately 1.8 km (1 mi) east of the
main river channel and 250 m (820 ft) west of the eastern valley escarpment.
The lower^lying adjacent soils in the floo^lain were somewhat poorly
dradned, loany textured with weak subsoil pedogenic development. The surfaces
were dark grayi^ brown with gray subsoils in Munsell hues of lOYR. They were
subject to seasonal flooding. The related inland soils of similar elevation
on the Pleistocene surfaces were de^, mature soils with definitive yellow-red
subsoil argillic horizons exhibiting strong pedogenic development. The upland
soils were very strongly acid, hi^ly weathered and siliceous with low base
saturation (Ultisols) . Some of the soils contained fragipan horizcxis at
d^>ths of 50-75 cm (19.7-29.6 in).
The floodplain soils adjacent to and below the site were dcminantly within
the Kirkville (ooarse-loaity, siliceous, thermic Fluvaquentic Dystrochrepts)
and Mantachie (fine-loany, siliceous, acid thermic Aerie Fluvaquents) series.
These soils had minimal pedogenic develorment with loany textures exhibiting
189
O-IO/I Dark reddish brown (5yR5/2) forest hums or topsoil averaged about 10
cm thick.
10-50/ Ihis zone was divided intx) two substrata based primarily on color and
IIA root content at both sites. Stratun IIA, a dark reddish brown
(5YR3/2) to dark brown (7.5YR3/4) sandy loam, averaged 35-40 cm
thick. It had very weak, stibangular, blodqr, friable soil with many
rootlets and approximately 2-3% ch2u:cx>al flec^.
45-80/ Sli^itly lifter color - dark reddi^ brcwn (5YR3/4) to very dark
IIB brown (10YR3/2) - sancfy loam averaging about 35 cm thick; very weak,
subangular blocky, friable soil with few rootlets and less than 2%
charcoal flecks.
L
80-100/ Transitional zone with reddish brown (5YR4/3) to brcwn (7.5YR4/4)
III loany sand with a small amount of light yellcw (10YR6/4) and light
gray (10YR7/1) mottling due most likely to natural disturbances.
110-120/ Relatively unaltered fluvial sand stratum quite variable in thickness
IV and £f:pearance; generally, it was ocnposed of yellowish brcwn
(10YR5/8) to very pale brown (10TO7/4) loamy sand with sane lifter
yellowish or grayish (10yR7/l, 8/4) mottling due to bioturbation. It
ranged from about 10-50 cm and banded in the middle with a particilly
developed illuvial zcme 5-20 cm thick of dark yellowi^ brown
(10YR4/4) to strong brown (7.5YR4/6) sandy loam. This band was
identified substratum IVB, and the paler semds above and below it IVA
and IVC, respectively.
120-140/ A well-developed illuvial deposit formed by the heavy oonoentratioi
V of fine iron-rich sediments in an exceptionally clearly defined layer
directly atop the impermeable paleosol, marking the level of the
water table. This stratum is ccmposed of firm dark yellowi^ brown
(10YR4/4) to strong brown (7.5YR4/6) sanc^ loam with a hi^ier colloid
ccxitent because of the perdied water table. It contained a high
amount of manganese and ferrv^inous sandstone nodules concentrating
above the inpervious paleosol.
190
22It623 and 221t624 (cCTitinoBd)
Deecription of repecacntative pcofile.
Depth/ ~
Strate De«3:jtf?tixai _ _ _
14(>f/ this buried soil horizon is made vp of a yellcwi^ IsroMn (10YR5/8)
IV sand/ loam matrix heavily mottled and streaked with yellow (10YR7/6,
2.5YR7/4), brown (7.5YR5/8), li^t gr^ (5Y7/1) sandy loam and clay^
_ sand. Subanqular, blocky stoucture with clay skins on the ped faces.
PHYSIOVL DESCREPTIOM
The particle size distributicn (Table 87) reflects proninent textural
stratification due to episodal fliivial sedimentation. Sanfy loam and loamy
sand textures occurred in the representative pedon analyzed. The loany sand
layers at depths of 47-122 cm (18.5-48.1 in) reflect hi^ier aiergy fluvial
events. Highest sand oontents occurred in the loany sand layers. Silt
contents were greatest in the surface l^ers, while clay contents were
greatest at depths of 122-162 cm (48.1-63.8 in) . Fine and meditsn sand classes
doninated the sand fraction with lesser amounts of very fine sand
(0.10-0.05 mm) and traces of coarse and very coarse sand.
rasiE 87
Particle size distribution of reprenentative pedtan, 22It623. _
Sand Silt Clay
Depth _ (2-0.05 nm) (0.05-0.002 ran) (< 0.002 mm) _ Texture
cm - percentage - - -
0-28
66.2
28.9
4.9
sancty loam
28-47
65.7
26.6
7.7
sandy loam
47-84
79.4
17.6
3.0
loany sand
84-122
84.5
12.7
2.8
loany sand
122-135
72.9
14.3
12.8
sandy loam
135-162
63.4
23.2
13.4
sancfy loan
Sand FTacticn
Very Goarse
Qoeurse
Mediun
Fine
Very Fine
(2-1 urn)
(1-0.5 ran) (0
.5-0.05 ran)
(0.25-0.10 ram)
(10-0.05 ran)
cm
-percentage
0-28
0.80
1.00
.28.20
31.00
5.20
28-47
0.04
0.20
5.60
51.10
8.80
47-84
0.03
0.10
29.50
43.50
6.30
84-122
0.00
0.03
6.90
69.50
8.10
122-135
0.00
0.03
6.30
58.60
8.00
135-162
0.10
0.10
6.60
45.20
11.40
The upper 110-120 cm (43.3-47.3 in) of the profile was dark brown
organiccLLly stained with weak subangular blocJq^ strvicture. A transition zone
Stratun IVB is tlwu^t to have developed because of the concentraticxi of
iron-rich sediments at the level of a perched water table vhich must have been
in existence for some time. It is a weaker, paler version of Stratum V. Its
transformation into a different substratum most probably took place after
cultural occx?)ation and fluvial deposition had covered it. Stratum VI was a
paleoeol esddbiting subangular blodcy structure vdth skins on the ped
faces. It is highly ««eathered and has been eroded vdth the A and part of the
B hcrizcxis reacved. Polygonal cracks are also present. This strata has the
hi^iest clay ocxitent on the site.
CHaccaL DEBaopncw
Soil pH levels (Table 88) were extremely acid with levels increasing
sli^itly with depth. Soil pH levels were similar to those of adjacent
off-site soils. Hydrogen and eOuninun were the dcminant exchangeable cations.
Hydrogen values decreased with depth, %diile exchangeable alvnlnun levels
varied considerably. Exchangeable calcitxa increased with depth reaching
nexiiun levels at depths of 122-162 an (48.1-63.8 in) . The increased Ca
levels were acocnpanied by proportional increases in percent base saturation.
Exchangeable Mg and K also tenM to increase %idth deptfi. Trace levels of Na
were detected. The Ca/Mg ratio also tended to increase with increasing
dq>ths. Very low cation exchange c^pacity values (Total, Table 89) occurred
in the loamy sand layers. Base saturation values were Id^iest at the greater
depths.
22It623.
Base
Saturation
Organic matter contents were hi^iest in the 0-28 cm layer (0-11 in) (Table
89) . Abrcpt decreases in the organic matter contents ooctrred at depths below
47 cm (18.5 in) , suggesting less cultural additions fron habitation. Nitrogen
contents decreased abr\;ptly with depth with the C/N ratio following a similar
trend.
P205 oonftants
ratio, and 1%
22ItS23.
trie Ac
Solthle P205
cm
—percentage—
ppm
0-28
1.80
1.04
0.117
8.8/1
0.0
28-47
0.59
0.34
0.045
7.5/1
141.6
47-84
0.25
0.14
0.012
11.6/1
120.7
84-122
0.03
0.005
2.0/1
0.0
122-135
0.09
0.05
0.016
3.0/1
40.5
135-162
0.09
0.05
0.018
2.7/1
60.8
192
Citric add solvdsle P^Oe contents %dere very erratic. No P-O. (extractable
in 1% citric acid) was detected in the 0-28 on (0-11 in) and 89-122 on
(33.1-48.1 in) layers. Maxinun levels occurred in the 28-84 an (11-33.1 in)
layers, but values were far below the 250 pfin required for anthropic epiped^
(Soil Taxonomy ) . Ihe lower levels of P^O. may be associated with less
cultural additions. These layers also have ^e^ low clay contents vdiich could
retain the P20^ against leaching.
mxxBac iNF!!aation by man effects the soil bcd^ in a unique, singular manner,
or produces a conposite synergistic result. These resultant features tend to
be masked by progressive weathering and farces of nature, including subsequent
pedogenic di^lQpnent, erosion and dqx>sition of soil materials, and intense
biological ekctivity from plants and arganiams.
Soil profiles reflect the different stages of horizon developnent that
correspond to the different time lengths during which the soil materials have
been eiqposed to soil forming processes (Thorp ) . Wider variations may be
expected in the characteristics of more recent soils than in older, mature
soils (Harradine ) . Human cultural alteration and natural fluviatile
actions can destroy previous soil development and result in relatively
juvenile soils over a wide chronological period. Siperitiposed soil profiles
in fluviatile deposits may be formed on discrete deposits due not cmly to
climatic variation, but also to aggrading, swinging streams depositing
sediment loads in various places, with superposition spaced widely eixxigh to
allow soil formation (Stephens ) . These fluviatile pedons may have
differences in texture and mineralogy associated with their original position
on levee banks of the floodplain.
SOIL DEVELOPMENT IN ABCHAEOLOGICMi SITES
Several studies have addressed soil development in archaeological sites to
relate the degree of pedogenic development with elapsed time. McOcnb and
Loomis () studied Indian mounds in forested regions of the DesMoines River
in Boone County, Iowa, and reported a lack of eluviation or aocunulation of
clay to form B horizons within a 1,000 year time period. Th^ concluded at
least 1,000-2,000 years were required to produce detectable soil changes under
forest in Iowa. Parson et al. () ocnpared soil developmmit of seven
prehistoric Indian mounds in northeastem Iowa. They reported mound soils
reached their present horizonation in a period of no more than 2,500 years.
They concluded that Al horizons vhich farmed under deciduous forest vegetaticm
in northeastem Iowa attained a maxiiiun degree of expression within 1,000
years or less and remained relatively constant in oonposition with increasing
age. They also concluded clay translocation may be evident between 1,000 and
2,500 yecurs of age.
Holliday () reomitly studied soil development in sites located in late
Holocene depx^sits at Lubbock Lake in the Southern High Pledns of Texas in a
restricted rainfall area. Soil ages ranged &oni 5,000-200 years before
present. He reported the presence of argillic and canbic horizons in the
Lubbock Lake soil (4,500-5,000 B.P.). Holliday concluded canbic horizons
formed in sediments of 450-800 years.
193
Ifae influence of huDoan activities such as fixes, waste disposal and
burials on the soil chendcal characteristics has been docunented
(Hoidenreich et al. ; Lutz ; Griffith ). Priaary asphasis has
been directed to elevated phosphorus levels attributed to prolonged hunan
activi^, with less attention given to other chasioal attributes. Recently,
Griffith () reported that nagnesiuni, organic phoephorus, and inorganic
phosphorus relative levels were sufficiently different to be useful in
distinguishing formerly occqpied soils from off-site soils. Galm ()
retorted that neutral pH levels, clay content, and hi^ iron oodde contents in
soils of the Curtis Lake site in Oklahoma %#ere useful indices of human
cultural activity.
The inorpl^io9ical expressions of buried soils vhich have remained
umodlfied reveal something of the conditions prevailing during a discrete
period in contrast to soils esqx^eed on residual features vhich reflect
emulative effects. Buried soils represent the product of soil-forming
processes over a time interval between cessation of deposition and its
reocninenoement .
GBONORFHIC SEITINS AM) SOIL PARQIT MATdlAL
Sites 22It539, 22It576, 22It590, 22It622, and 22It623/22It624 occupied
topogrephic highs of the Tenhighse River floodplain, rising 1-3 m (3.3-9. 9 ft)
above it. Site 22It606 was located on a Pleistocene terrace vhich had been
incised by the Ttahi^see River and Mud Credc, and it was not subject to
flooding. The Ttehi^oee River is an underfit stream which has meandered from
the eastern to the western part of the floo^lain. It is currently an
aggrading stremn. The river has the cenroon fluvial features including
oK-bowB, natural levees, and point bars. The river fills the floodplain
channel only during flood stages which usually occur in the winter and spring
months. The off-site floodplain soils are saturated nuch of the year, and
they eodiibited little pedogenic development with bedding planes persisting.
The soils had reduced conditions vhich were reflected by gleyed colors in the
subsoils.
The Tenbi^xe Hills of the study areas were oon|a:ised of unconsolidated
marine sediments of Upper Cretaceous age. The Eutaw and Tuscaloosa formations
outcropped in the area and provided the parent material for the vqpland soils
and edluvial deposits (Stepheson and Monroe ) . The TUsoaloosa formation
is characterized by irregularly bedded sand, clay, and gravel. The Butaw
fozmatlcn is generally oenprised of croes-bedded glauconitic sand and clay.
The Tcnbi^aee Hills have apparently been eroding and redepositing in the
Ttaobi^bee River floodplain since Pleistocene and perhaps late Pliocene time
(Stepheson and Monroe ) . Current Holocene sediments in the active
floo^lain are heterogeneous and related to current erosion and deposition
processes. The floodplain had a dense cover of deciduous vegetation.
Precipitation in the study area ranged frero 125-140 an (49.3-55.2 in) with a
frost-free period of 200-240 days (Petiry ) .
The Entisols and Inc!q>tisols ooiprising the floodplain exhibited little
profile develcpnent and horizonation. Typically, these soils had thin surface
horizons underl2dn by stratified materials or cznbic horizons with little
evidence of translocated cley. The deeper horizons had gleyed colors vhich
reflected the saturated, reduced conditions. The water table in these soils
ranged frem at or above the surface during winter and spring to depths of 2 m
(6.6 ft) or greater in the sunmer and early fall seasons. Periodic overflow
resulted in surface scouring in the hi^ energy floodwater areas and
deposition in slack water areas.
194
Sedimentation rates calculated frcm the sediment thickness and ages of
major cultural, oonponent segments of 22It576, 22It539, and 22ItS90 indicate a
progressive decline in sedimentation rates from the early Holocene (Kirk:
9,500-10,000 B.P.) to the present (post-Benton, 5,000 B.P. - present) as shown
in Table 90. Pre-5,000 B.P. sedimentation rates ranged from 8 cm (3.2 ft)/
century at 22It590 to 17.6 cm (6.9 in) /century at 22It576 vihich was located
about 16 Ion (9.9 mi) downstream at an elevatiim 6-7 m (2. 4-2. 8 in) lower.
Both sedimentation rates and sediment thickness increased with increasing
distance frcm the headwaters. Site 22It590 was located near the headwaters of
Macke's Creek with a smaller %«atershed than 22It539 and 22It576. Po8t-5,000
B.P. sedimentation rates averaged less than 2 cm(.8 in) /century for all the
sites. Calculated annual sedimentation rates may not be an accurate depiction
of yearly aocunulation events, but the data clearly reflect much greater
sedimentation during the early Holocene period. Gecmorphic stability was
probably less during the early Holocene period and suggests lower artifact
densities may be ejipected in these strata.
TMU 90
SM— gy of pndimraitation rates by cultural onmCTient and elevation. _
I _ 22It576
Elevation
Sed.
cm/
Cultural Ccnponent
(m)
(cm)
century
Kirk
79.0-t9.88
88
17.6
Bm/Morzow Mountain
80.2-80.4
20-40
3. 3-1. 5
sykM/White l^xrlngs
♦•not definable**
Benton
80.4-80.6
20
2-4
post-Benton
80.6-81.28
68
1.2
22It539
Elevation
Sed.
cm/
Cultured. Ccmponent
(m)
century
86.5-87.25
75
15
Eva/Morrow Mountain
87.15/25-87.5
25-40
2. 2-3. 5
%kes-4tiite Springs
87.5/7-88.0
30-50
Benton
88.0/1-88.3
30-40
6-8
post-Benton
88.25-89.03
78
1.2-1. 6
22It590
Elevation
Sed.
cm/
(ra)
(cm)
century
Kirk
96.1-96.5
40
8
Eva/Morrow Mountain
96.4.5-96.7/8
30
1.5-3. 3
S^es-White Springs
**not definable**
Benton
96.7/8-96.9
10-20
2-4
post-Benton
96.9-97.3
40
0.7
Note: Calculated using best defined cultvured. zones Band/ or natural zones.
was conservative with data - used best-^fined zones and ocnpcmients.
Ccnponents best defined are Kirk, Benton, and post-Benton. Ccnponents
least defined are Eva/Morrow Mountain (vpfser portion likely eroded
_ away) and S^tes-White Springs (may be part of Benton at 22It539) . _
The incranent£d additions of fresh cdluvixxn over time to the sites in the
floo^lain resulted in cumulative soil development. The A (surface) horizons
of cimlative soils such as in sites investigated developed continual
deposition of fresh sediments on the surface.
The soils on the Pleistooene uplands bounding the Tcehigbee flood^>lain
mere tn^oally acid, hic^y leached, ii«ell-dBveloped Ultisols with low base
saturation levels and organic matter contents. The mature rpland soils had
definitive argillic horizons of illuviated clays with Nunsell colors that
ranged fzon yellowHarown (10YR5/4) to zed (2.5TR4/6) . Firm, dense fragipan
horizons occurred in some upland soils on broad, gentle slopes.
PmOBCPlC EPIFH3GNS
Anthropic epipedons are diagnostic horizons that form at the soil surface.
They axe darX-oolored horizons rich in organic matter vhich have resulted iron
long-continued use of the soil fay man eis a residence or site for irrigated
crops. According to Soil Taxonomy () , requirements for anthropic
epipedons incltide the fbllowiing:
a. Soil structure is strong enou^ that the horizon is not both massive and
hard, or very hard «hen dry.
b. Have Munaell color value darker than 3.5 vhen moist and 5.5 \hen dry and
chroma less than 3.5 \dien moist; the color value is at least one Munsell
unit darker or the chroma is at least two units less than the underlying
1C horizon.
c. Organic carbon content is at least 0.6 (1% organic matter) .
d. The ^ipedon has more than 250 parts per million (ppn) of P.Oc soluble in
1% citric acid.
Sites 22It539, 22It576, and 22It590 had epipedons vihich met the criteria
for anthropic with (tinsell value and chroma colors darker than 3.5 aocxapanied
by organic matter contents exceeding one percent and contained greater timn
250 ppn P-O. extractable in 1% citric acid. Site 22It539 had layers occurring
at depths^oz 0-20 on (0-7.9 in) and 50-80 cm (19.7-31.5 in) which met the
anthropic criteria. Site 22It576 had layers from 0-80 cm (0-31.5 in) which
met the criteria, vbile 22It590 had an anthropic epipedon extending from the
surface to depths of 69 cm (27.2 in) . Base saturatim was about 40% in the
anthropic epipedon of 22It539, and exceeded 50% in 22It576 except for the
0-10 cm (0-3.9 in) leyer. Site 22It590 had base saturation levels near 50%
except for the 0-30 cm (0-11.8 in) inczenent.
Sites 22It622, 22It623, and 22It606 did not have horizons udiich met the
neoessazy anthropic requirements due to low P.Oe contents and/or colors vbich
were too li^t.
The anthropic epipedons are distinctive features of the sites and serve as
pronlnent indicators of the long-term habitation. The additions and
incorporations of waste, plant materials, and other ocnponents over the
duration of habitation synergistically prodooed the unique hcnrizonation which
diff^s markedly from naturcil, undisturbed soils. Soil organic components
result from inocxporation of the bicmass of the site. Typically, the organic
fraction may be divided into the part ybere plant parts can be recognized, and
the ocmpleti^y decoapoeod materials. The organic fraction is generedly
crwpoewl of 1) carbohydrates, 2) amino acids and proteins, 3) lipids,
4) nucleic acids, 5) lignins, and 6) humus (Thn ) . TSfm organic fraction
of the sites studied was dominantly well deocaposed with very low content of
recognizable plant parts. The soil polysacxharides have a profound influence
on soil physical and chemical conditions and affect pedogenesis. Interaction
196
of soil polysaocharides with soil particles prcnctes soil aggregation and the
formation of granular structure (Greenland et al. ) . The stabilizing
effect on soil structure has been attributed prixnarily to cementatic^ effects.
The organic oonpounds interact with clay surfaces and oaipete with water
molecules for adsorption sites and reduce wetting and swelling, thus enhancing
cementation (Tan ) . Well-developed granular soil structure and the
tendency to repel water were proninent characteristics of the anthropic
epipedons.
The hig^y deocnposed or hunified organic matter is referred to as huius,
and it is ocnsidered to be the end product of deccnposition of plant materials
in soils. Previously, hunus was classified into 1) himic acid, 2) fulvic
acid, and 3) insoluble hunin, the inert part. Currently, humic ccnpounds are
defined as amorphous, colloidal, polydi^^ersed substances with yellow to
brown-black color and high molecmlar weights (Tan ) . Analyses of the
anthropic ^ipedons indicated the humus was dominantly hunic acid carpounds
with hi^ molecular weights (Tan and Giddens ) .
Soil hunus content in the southern region of the United States seldom
exceeds 3.5%, and it is concentrated in a relatively thin surface horizon.
Very low contents typically occur with increasing depths. The carbcai/nitrogen
(C/N) ratio usucLLly declines in the hunification process from values in excess
of 20 for fresh naterial to values of 8 to 20 for hunus (Tan ) .
Fitzpatrick () reported subsoil middle horizons may have C/N ratios as low
as 4 that may be due to a hi^ content of amnoniun ions fixed ly the clay.
Carbon/nitrogen ratios tended to be erratic with increasing depths, increasing
to 44/1 at depths of 50-60 on (19.9-23.6 in) in 22It576, 40/1 at 110-120 on
(43.3-47.3 in) in 22It539 and 30.8/1 in 22It590. Despite the large C/N
increases, no intact organic matter was detected in these strata. The erratic
C/N levels with depth appear to be a diagnostic reflection of discrete
cultural components. The hi^ier subsurface C/N levels of 22It539, 22It576,
and 22It590 generally coincided with the Benton oomponents.
The extent of organic compound movement fron cm strata to another, or
removal from the profile is a very coiplex issue which is greatly ooiplicated
by physical perturbation of the sites during and after oocn^tion. However,
in natuTcil soils it is theorized that mobile organic materials are leached in
acid environments and move downward to sane depth were they are rendered
imnobile by chemical combination with alunixun and iron oanpcunds
(Holzhey et euL. ) . The hunic/fulvic acid ratio reportedly increases with
depth in such ocmditions (MoKeague ). Holzh^ and co-workers ()
reported the thick, black organic accumulations in spodic horizons were
generally absent vAien clay contents were above 10%. They postulated migrating
organics aocuoulated and apparently polymerized to hunic acids in sands until
some maximum organic content attctined. They surmised that b^rond that
maxiraun little accumulation occurred, althcu^ migration continued. Organics
moving with percolating waters were dominantly fulvic acids, whereas the
water-insoluble organics were dominantly hunic acids. Laboratory tests
oonducted on the epipedons of 22It539, 22It576, and 22IT590 indicated
dominantly water-insoluble hunic acid caipounds.
The well-developed granular stiructure, resistance to wetting and
solubility, and uniform coating of siliceous sand and silt grains suggest
orgzuiic-ininercil bonding to the silica or quartz surfaces. Based on visual
changes during chemical extraction with bases, the organic-minered bonding
appears to be quite strong. The organic coatings tend to resist removal with
H2O2 treatment.
Questions exist jrelative to the permanoioe of the huaus rich, dark'-oolored
anthropic epipedons. Will they graAially lose organic matter and dark color
throu^ microbizil degradation over time, since eidvanoed carbon inputs frcnt
tudditatlon have ceased?
The ^ipedons way represent a climactic characteristic resulting from the
synergistic inputs of various carbon and phosphorus ocnponents due to
prolonged habitation by man.
Ccsparisons may be made to carbonr-nitrogen balances in natural emulative
soils %^ch have not received enhanced additions of organic matter and mixing
due to prolongsd habitation, emulative soils oosnonly contain sane allogenic
organic matter vdiich was present in the eroded sediments. Recent studies
(Mmel ) indicated the ratio of total nitrogen in eroded materials to
that in the source soils (Enrichment Ratio) decreases logarithmically as the
amount of sediment increases in agricultural weteraheds. This relationship
indicates the allogenic organic matter present in the parent materud would
have a decreased rate of breakdown as burial proceeds and time increases.
Jenny () quantified the edlogenic organic matter hrealcdawn rate of burial
sediments of the Nile River in the following manner: dN/dT > A - Ko N, vdiere
A was the nitrogen gain by fixation and Ko N were the losses due to micurobial
debonposition. After d^xmitions, both nitrogen fixation (additions) and
microbial deoonposition (losses) occur, with losses exceeding additions.
Thus, nitrogen (organic matter) tends to decrease with increased lime and
burial depth. Other studies (Hole ) have indicated the dark-oolored,
huoaus-rlc^ spodic horizens degraded after removal of the hemlock (Tsuqa
r-wniVwru^is) foTOst oover vhlch was providing the organic replenishment. Hole
(11^75^ estimated the half-life of a spodic horizon was alxut 100 years
foUotdng removal of the hemlock forest oover.
The dazk-oolored anthropic epipedons in 22It537, 22It576, and 22It590 have
persisted at least 5,000-6,000 years. However, replenishment of the organic
oespounds by hz^itation ad^ticxis have been maintained over the 5,000-6,000
year period until abandoonent within the past few hundred years (500-300 years
B.P.).
The P.O- levels extractable in 1% citric acid are supposedly depictive of
the phospnoCus additions by man during habitation with a threshold level of
250 ppm necessary for anthropic ^ipedons. The P^Oe levels of 22It539,
22It576, and 22It590 far exceeded 250 ppm with highest levels detected in
22It576 and 22It539, reflectively. Values in these two sites were several
fold hic^ier than were detected in 22It590. The hi^mst citric acid
extractable P20e levels in 22It539 and 22It576 generally occurred in the
Bentcn oenpeni^ strata. In cenparisen, the extractable P^O^ levels in
adjacent indistxirbed off-site soils were far less than 250^p^, which
indicates that the much greater levels present in the sites were due to
prolonged habitation.
Althou^ organic phosphorous contents were much hi^ier in the site pedons
ooipared to adjacent off-site soils, the data were erratic and difficult to
interpret. Total fractionation of the phosphorus (soil) into the discrete
ooiponents for a representative pedon from Block D of Site 22It576 indicated
the distribution with df>th. Total phosphorus reached maxinun levels at
df>ths of 60-70 an (23.6-27.6 in) (Benton). The non-occluded phosphorus
fraction (Pnoc) ootprised the largest proportion of the total phoqphorus with
highest values occurring at depths of 70-80 on (27.6-31.5 in) and lowest
values in the surface horizons. The non-oocluded phofhorus has been
associated with the al\ininum and iron oxide surfaces is availadale for
plant use (Meixer and Singer ) . Walker and Syers () reported in a
study of hvmid to dry soils of different ages in New Zeedand on P
transfarmaticns during soil development. They found the transfomations
consisted of the dissolution of calcium phosphates and the formation of
organic and alunoinun- and iron-oxide associated forms. Initially the
altminiiD- and iron-oxide associated forms were non-occluded, and later all
forms of soil phosphorus transformed to organic and occluded forms. Recently,
Meixner and Singer () found organically bound phosphorus did not diange
systematically with time, and non-occluded phosphorus generally decreased in B
horizons in California soils ranging frcni 300-250,000 years of age.
Beised on the data of 22It576, non-occluded forms of phosphorus oonprise
the dcminant part of the complex, followed by oocluded phosphorus, calciun
phosphorus, and organic phosphorus (Pnoc < Poc < Pea < Po) . Due to the acid,
inert nature of the original sediment oonprising the parent material very low
levels of calcivn phosphorus were present vAien the sediment was deposited.
The elevated levels of total phosphorus vdiich were dominated by non-occluded
form were characteristic of the sites. In an idealized pedogenesis model of
pho^horus transformation, calciun P in apatite or other parent materials
woild weather and be transfomed to organic P by plants with some of the
organic P transfarmed to non-occluded P in the presence of iron and aluminun
oxides which would then be transformed to oocluded P in the crystal lattices
of iron and aluminun oxides and not be available for plant \:qptake. The
munomorphic amorphous organic matter resulting from the dissolution and
precipitation of humified organic matter (DeOonick et ed. ) oonplexes with
iron and alixninum oxides and provides the mechanian for the persistence of the
nan-occluded P. The monomorphic organic coating of mineral surfaces gives
rise to the persistent darlc color and the elevated phoqphorvis levels (greater
than 250 ppm P extractable in 1% citric acid) . The hunus and sesquioxides
present served as a very large "sink" for the formation of non-occluded P.
Calciun-P could be provided by bones of animals and/cn: burials. However, the
very lew levels of Ca-P suggest contributions by beme were either minimal at
the sites, or the Ca-P introduced was transformed readily to non-occluded P
and organic P. The intermediate levels of organic P (hi^ier than adjacent
off-site soils) reflect the cultural additions of organic P from organic
sources and contributions by plants present on the sites, i^parently,
sufficient time has not occurred (10,000 years) or environiental conditions
changed sufficiently for transformation of the non-occluded P to oocluded P.
PALE306QLS
Well-developed paleosols, with illuviated clays and oenpound structure,
occurred in 22It539 and 22It576 at d^sths below 1.25 m (4.1 ft). The
paleosols had yellowish brown colors in Munsell hues of lOYR in contrast to
dark-oolored superjacent anthropic qpipedons and gleyed, reduced colors in
adjacent off-site soils. The presence of well-developed paleosols containing
argillic horizons with enrpound structure was uneoqjected in the floodplain
setting. A preminent mDr^hologiced feature of the paleosols was the
occurrence of a polygonal network, with polygons separated by hi^ily leached
seams of silt and very fine sairi. The presence of the paleoargillic horizons
indicates advanced pedogenesis occurred over a prolonged period of stability
prior to subsequent burial by sediments. Diagenetic alteration of the
paleosols after burial a^ipears to be minimal other than possible cciip>action
and chemical infusions. The absence of a detectable antecedent surface
horizon (A) and the presence of an unconformity with the anthropic epipedon
overlying the paleosol suggests truncation by erosion. The high sedimentation
199
(
I
I
rates of the early Holocene suggests a period of instability and widespread
j lateral erosion and deposition in the floa^lain during that period.
Assigment of a chronological age to the paleosols of 22It576 and 22It539
I presents an enigna in the absence of materials in context for radiocarbon
; dating. HoMever, based on the occurrence of axgillic horizons of oonparable
pedogenic developnent occurring only on stable Pleistocene surfaces in the
region, a late Pleistocene or early Holocene placement seems prudent. Other
than formation of the anthropic epipedons, relatively little pedogenic
I develofinent was detected in the sites above the paleosols, other than lamellae
(particularly 22It590) and sporadic occurrence of caobic horizons. The
episodic sediment depraitions and intense biopedoturbation due to habitation
and intense microfloral and microbiotic activity have apparently retarded
pedogenic developnent in the sites.
The extensive polygonal macxoetructure present in the paleosols occurs in
fragipan horizons of mature upland soils and has also been reported in buried
I paleosols of selected rivers in Mississippi (Abu-Agwa ) . The presence of
fragipans in soils has been attributed to ocnpaction due to weight of glacial
ice, permafrost, periods of desiccation (Grosanan and Carlisle ) , and to
buried strata (Buntley et al. ) . Soils with polygonal surface patterns
have Iseen reported in semiarid soils (Hugie and Passey ) . The vesicular
pores associated with polygonal surfaces and fragipans, and vhich %iiere oontnon
in the paleosols of 22ItS39 and 22It576, have also been associated with dry
conditions (Lphan ; Springer ). Micro-morphological examinations of
the paleoargilllc horizons indicated ocamon bisection of existing structural
dooiaLis by the polygonal cracks vhi(^ suggests formation of the argillic
horizon prior to development of the cracks and polygonal morphology.
Based on the pedological data obtained in this stui^, the following are
reasonable postulates;
1. Sites 22It539 and 22It576 were stable gecmorphic surfaces such as outliers
of the Pleistocene valley wall or prominent levee/bar features which
developied cungillic horizons during the Pleistocene period.
2. Xeric conditions developed vdiich resulted in the polygonal morphology with
deeply incised desiccation cracks pemeating the argillic horizon and
bisecting existing structural fabric.
3. Hixnid conditions occurred vrtiich resulted in severe flooding and erosive
truncation of the loose surface horizon down to the ocxisolidated argillic
horizon.
4. Progressive incremental sedimentation of the sites occurred over time with
the aggrading stream conditions becaning less proncunoed.
5. Habitation of the sites enhanced carbon inputs, mixing and associated
bicperturication, v^ch synergistically affected soil genesis.
Grissinger et al. () reported the occurrence of paleosols in stream
channels near the bluff area of northern Mississippi. The paleosols had
distinctive polygonal morphology and were ocnsidered to be Holocene age.
Paleoclimates of a hotter and drier nature have been suggested for the region
of the stuc^ curea during early Holocene to roid-41olocene bemed on pollen and/or
plant micro-fossil studies. Drier ocxiditions have been su^r^ted in middle
Tennessee J&xm 8,000-5,000 years B.P, (Delocurt ), and in scutheastem
200
Missouri fran 8,700-5,000 years B.P. (King and Allen ) . These findings
tend to support the hypothesis for the sites containing well-developed
paleosols.
Similarities exist in pedogenic parameters of the buried pcdeosols of
22It539, 22It576, find the Pleistocene inland 22lt606 which contained an
argillic horizon. The argillic horizcxi of 22It606 tended to serve as a
"basement” for habitation activity, with cultural disturbance and artifacts
limited primarily to horizons located above the argillic horizcsi and in the
vfjper part of the horizon. Coloration of the argillic horizon was not
e^preciably affected by prolcxiged habitation, suggesting that the
well-developed argillic horizon with defined structure and illuviated fabric
was existing and in place prior to habitation. A similar analogy may be
^plied to the paleoargillic horizons of 22It539 and 22lt576. However,
chemical inputs of a nonrvisible nature resulting from habitation left
imprints in the paleoargillic horizons. Cations and phosphorus levels in the
paleoargillic horizons %4ere apparently enriched by leachate fran superjacent
horizons containing cultural activity. Analyses of the interior of a polygon
in the paleoargillic horizon of 221t576 at a depth of 1.6 m (5.3 ft) had the
following P levels: non-occlvided P = 926.9, occluded P = 369.6, calcium P =
12.3, and organic P = 18.2 ppm, respectively. The elevated phosphorus content
and fractional distributicxi are depictive of the cultural disturbance and
habitation.
A oonoeptual model of soilscape stability for habitation based on
soil/georoorphic parameters for the sites analyzed in this stucty may Ise
esqpressed as: 22It606 > 22It539 > 22lt576 > 22It621 > or < 22It622. However,
other factors obviously have a bearing an the population and longevity of
habitation and the resultant artifact assemblies and densities.
Althouf^ each river system differs markedly in hydrological properties and
sediment characteristics, sites in ether river systems in the scutheastem
U.S. should exhibit similar pedological features of paleoeols and polygonal
roorpholcigy if the proposed hypothesis has validity. If other sites are
detected vhich have similar soil morpholc3gical features, cxtrelation of
artifact cxnponents and radiocarbon dates should clarify the chronological
sequence.
SaiftRY FOR AJCHAEOLCXSISTS
The soil studies fcxrused on basic ejuestions: were the midden mounds
natural or cultural phenomena, «hy the thiok midden was still so dark after
5,000-6,000 years of percolation and almost 300 years since active site use,
arxi vhat is the description of the paleosol buried below this midden? In
order to deal with these issues, a firm baseline of desenriptive data had to be
cxxipiled from a r^nresentative sanple of midden mound profiles and from an
vpland site as well. Five floodplain sites were studied, but most analytical
work was done on the Poplar (22It576) and Walnut (22It539) sites vhich were
typicul midden mounds. The ethers investigated on a less intensive level were
the Ilex (22lt590) , Oak and Beech (22It623 and 22lt624) , and Mud Creek
(22lt622) . The upland site studied was 22It606.
Fieldwork wzus conducted in several visits to each site. The first visit
exxurred vhen a profile had been eoqjosed to sterile soil either in an
excavaticxi unit or by bacldtie. Chrcsnolcsgy of the soil profile based on the
archaeolc}gi(oal information and other relevant site information was presented
to the soil scientist in a briefing session pricsr to each fieldwork sessicxi.
Fieldwork at each site included a detailed desenription of the soil profile
201
(e.g. Table 50) with saraples of each stratum retrieved for laboratory
analysis. Off-site sampling was conducted by hand with a three-inch (7.5 an)
bucket auger or by machine with a backhoe. Off-site sanples were collected on
systematic transects or positional locations. In addition, ocsiferenoes were
held on site with the soil scientist and the archaeological staff concerning
site stratification, distributicxi, and sequence of depositional units and the
oorrelaticm of cultured oonponents and strata. Backhoe trenches were often
exc:an;ated during these conferences to cinswer questions, and the tracing of
strata on unit profiles for scele drawings was reviewed. Usu£dly the soil
scientist visited the site several times.
The labesratozy work including both physical and chemicel analysis was done
in the soil science laboratories of Mississippi State University. X-ray
diffraction analysis was cdso performed on sane samples.
The results of soil field and laboratory analysis will be presented below
follcwed by an integration and interpretation. Supporting data are contained
in the previous secticxis of this chapter.
THE WAI2Wr SITE (22It539)
This site was loceted at the headwaters of the TOtibigbee near the
confluence of Mackey's and Big Brown Creeks (Figure 1) . It was a preminent
flcx)^lain feature approximately 80 m (264 ft) in diameter, 1.5 m (4.5 ft) in
hei^t and surrounded by creeks and wetlands (Figure 3) . The deposits of the
site were wall-drained, reddish brown san^ loams vhicdi contrasted sharply
with the adjacient pc»rly drained silt and cl2Q^ d^xasits and the well-develcped
upland soils. The site soils were characterized by 1.25 m (4.1 ft) thick
stratum of organically stained loatry soil \^ch rested on the truncated
surface of yellow-brown paleosol (Figure 11) .
The dark loamy mddden had been extensively altered by burrowing animals
such as crawfish, earthworms, and rodents, by roots from the thick vegetation
cover, and by long-term human exxupation. These forces had affected admost
every measure of the upper meter of the site including color, structure, soil
development, texture, and chemistry. For example, in the 5,000-6,000 years of
buildup, no soil horizons had developed in this zone. It it and was
essentially an "A" or surface horizon.
This kind of organiccdly stained midden has been identified and defined by
soil scientists (Soil Taxonemy ) as an anthropic epipeden, a unique soil
fonted by the interaction between human and natural forces. This zone at the
walnut site was characterized by a high amount of organic matter which had
coated the sand grains with hunus and cemented them together into granules
which water could not penetrate. It took extremely caustic bases and physical
shaking finally to free the sand grains of this humic acid coating. Due to
the strength of this chemical bond between the humic acid and sand grains, the
organic matter could not be moved down by water percolation, and the zone
remained dark in color and greasy to the touch. Bands of fine material
(lamellae) were present in the lower half of the dark midden zone probably
caused by poor percolation of water through this organic zone and a frequent
high watertable in the floodpleiin. The carbon: nitrogen ratio, vhile varying
widely, was especially concentrated in the base of the midden zone (50-150 an
or 19.7-59.1 in) . This ratio may be related to the presence of fired clay and
corresponds to the 2u:^ in the profile of many fired clay hearths (Chapter V
and Chopter X) .
The paleosol beneath this midden had been eroded, and the old "A" or
surface horizon had been removed esposir^ the "B" horizon, but otherwise it
202
was well preserved, nils paleosol was as well developed {d^sicadly as the
paleosols of Pleistocene age in the adjacent inlands. Ihe paleosol Indicates
a long period of developoent during which the landscape was stable. Ihe
structure is crosscut with polygoned cracks outlined by white seans yahich
indicate that the soil developed before the polygonal cracking took place.
Chendccdly, the paleosol contained the hi^)est amounts of exchangeable
adunimm in the site profile vAiich indicates extreme age and weathering.
Ilhe size of the particles in sediments of the site all reflected that it
had formed naturadly from discrete episodes of fluvial deposition. Ihere was
the expected sorting of coarser particles toward the bottom of the landform
and finer particles toward the top. From these data and the lack of any
artificially deposited strata, the formation of the landform was determined to
be from natural forces. It was a "parallel bar", tear drop in cross section,
with a steep side farcing upstream flow and other sides tapered (Figure 3)
vAiich formed during episodic flooding. 'Ihese bars are not formad along the
mainstream but cue scattered in the flnndplain. ohe coarseness of the
sediments and the lack of a mainstream channel are the primary indicators of
these phenomena, both of vhich were present at the Walnut site.
All the characteristics described above contrast sharply with the
surrounding soils of the flcxx^lain or inlands. 13ie dark midden is unique to
sites intensively cxxnjpied by hunans, and the pcdeosol is cus well developed as
those in the uplands vhich have developed in sediments which are tens of
thousands of years older than those in the floo^ledn.
THE PQPIAR SITE (22It621)
The Poplar site was located in the flcxx^lain 24 km (14.9 mi) downstream
from the Walnut site (Figure 1) . It was smaller than the Walnut site,
cpprcDciroabely 40 m (132 ft) in diameter and 75-100 cm (27.6-39.4 in) in
hei^it, and was in the same physiographic position surrounded by creeks and
wetlands. The stratigraphic sequence and {^sicad and chemical
characteristics of this site mirrored that of Wednut with a few vcuriations.
The two prinary strata were a 120 cm (47.3 in) thick organically stained
midden resting unocmformably on the ercxied surface of a paleosol (Figure 4) .
In this midden between 50-80 cm (19.7-31.5 in) below surface, a zcxie of
particularly hig^i citric acid soluble P2OC defined vAiich cxxdd be
indicative of greater site population at that time, incnaased length of
ocxn:pation duration, or other acrtivities during this time. This agrees well
with the archaeological material recovered from this zone. The cdiemical tests
also identified two zones (50-70 cm/ 19.7-27.6 in and 90-100 cm/35.5-39.4 in)
v^ch had greater additions of phosphorous \^en they were exposed surfacas,
perhaps from increased site pc^pulaticxi and animal prcxassing. At the base of
the midden zcne horizontal bands or lamellae had formed in the midden (Figure
4) , likely dve to perciiing of water on top of the relatively impermeable
paleosol and to the depxaition of fine particles. The similarity between
these two sites at even stall levels of detail is striking and documents that
the same forces of site formaticxi and site use were in acrticn.
The Poplar site landform was also a parallel bar formed from naturad
d^xaition of materiads from episodic flcxxis. It was teaur drop in cross
secrtion with a steep flat vpstream faca (north) with cxarser and more
ccxisistent distribution of particles than the surrounding flcxx4>lain.
Discrete depositions could be identified, and edl strata had the same parent
material .
203
THE HTCKOOr SITE (22It621)
ttiile intense soil studies were not perfarmed at the Hickory site, it is
mentioned here because of its similarity to the Poplar and Walnut sites. It
was located in the floodplain between these two sites e^jpcoodmately 5 km south
of Walnut and 14.5 km (9 mi) north of Poplar (Figure 1) . It was characterized
by a 100-110 cm (39.4-43.3 in) thick, dark organically stained midden resting
unocroformably on the truncated paleosol (Figure 17) . This particular area of
the f looc^edn was extremely low and wet due to many seep springs in the
vicinity, and this ocmbination of a relatively hi^ier water table and the
iiipezmeable paleosol caused the ijpper midden to be saturated longer than any
other midden mound investigated. This probably caused a 50 cm (19.7 in) thick
zone of manganese concretions to form within the midden above the paleosol.
Althouc^ no intensive soil studies were performed, the archaeologiccd
investigations at this site indicated that the deposits likely were chemically
and physically very similar to the Poplar and Walnut sites. The Hickory site
landform was also a floot^ledn parcillel bar similsu: in shape and orientation
to Walnut and Hickory.
THE ILEX SITE (22lt590)
The Ilex site was located in the floodplain on a terrace of Mackey's Creek
cpproximately 5 km (3.1 mi) upstream from the Walnut site (Figure 1) . The
site was in a different position than the others investigated in that it was
tangent to the valley wall and had formed from overbank deposition as well as
colluvium washing down the ste^ valley wall. The sediments contained more
sand, and there was no paleosol beneath the 100-120 cm (39.4-47.3 in) thick
organically stained middten (Figujre 14) . The S5uicty strata umderlying the
midden also contained lamellae similar to those at Poplar and Ilex sites.
There «ms a lateral gradient (north-south) to the chemical testing results
indicating perhaps greater age at the southern end of the site adjacent to the
valley wall and greater cultural ailteration at the northern end of the site.
Seme rewDr}dng of the sediments was also indicated. It diould ]3e understood
that at least ten lateral meters of the site at the the northem end site had
been removed prior to our investigations. The chemical and physical
characteristics of the organically stained midden were the same as the Walnut
and Poplar, including the humic acid coated sand grains and the hic^ier amounts
of calcium, organic natter, and potassium. The absence of the paleosol at
this site may be due either to erosion or to the lack of suitable environment
and materials for formation.
SITE 22IT606
This was the only non- floodplain site included in the soil studies. It
was on a hi^ (15-20 m/49.5 ft) terrace outlier overlooking the floodplain
within 1 km (.62 mi) oi the Mud Creek site and 2 km (1.24 mi) of the Hickory
site. It is different fron the floo^lain sites stratigraphically,
chemically, and physically. The soils of this site are typical of mature
upland eu?eas and include a ]:)uried Pleistocene paleosol. Only the u^per 27 cm
(10.6 in) had been culturally altered, and this may have been throu^ the
addition of agricultural fertilizers rather th^m prehistoric hunan occupation.
The pedeosol was buried 100-150 cm (39.4-59.1 in) and vms fully developed
prior to human habitation. Below 27 cm (10.6 in) , the uqpper zone weis not
measurably different fron the surrounding area.
'SttB paleosol at this site was cxxnposed of a differ^t par^t material than
the zones above it and was the product of a different depositioncd.
enviromient. As with the flooc^lain sites, however, the paleosol had been
eroded with the "A" horizon removed, and younger, unrelated sedinents were
deposited lat^.
THE MUD C3?EEK SITE (22It622)
This was a small mound in the Tcnbigbee floodplain approodmately 6 km
(3.7 mi) north of the Beech and Oak sites (Figure 1) . It was approodmately
40 m (132 ft) in diameter and 50-75 an (19.7-29.6 in) in height and tear drc^
in cross section. The stratigraphy revealed an organiczdly stained loatny
midden vdiich had severe bioturbation (anijial burrowing) . The stratigrc^^hic
sequence did not include the pcileosol. However, this site was oily tested and
did not include intensive investigations. The chemical and physical
properties of this site reveeded less alteration than any other site studied
in the floo(%)lain, revealing only a 30 an (11.8 in) culturally altered zone or
midden. This documents that sane well-drained elevations in the floodpledn
were not intensively occv^Jied.
THE BEECH AND CAK SITES (22It623 and 22It624)
Although these two sites were originally separated vhen identified, they
are adjacent fragments of the same landfom in the floo^lain of the Toxbigbee
and are considered caie site for anedytical purposes. This site was located
between the Walnut and Popleu:, approximately 6 km (3.7 mi) north of Poplar
(Figure 1) . The site landform was a levee remnant which has been bisected by
a small tributary stream. The levee was fanned by overbahk deposition from
the east \*ere there is an abandoned channel segment of a major stream, likely
the Tcnbi^see \diich today is ca. 500 m (1,650 ft) to the west.
The primary strata are in a 120 on (47.3 in) thick organic-stained midden
zone vrtiich rests on the yellow^brcwn truncated pedeosol (Figures 18 and 19) .
The midden at this site is dark organically stained for only 50 cm (19.7 in)
with a lifter brown lower half. This differs from the previously discussed
sites, and it does not meet the criteria for an anthropic epipedon. The
ccxitact Isetween the peileosol and the midden was characterized by manganese
ocmcretions, especially at the base of organically enriched pit features. As
at the Hick^ site, only 9 km (5.6 mi) north, the concretions are likely the
result of orgsmically stained soil saturated much of the time a fluctuating
water table perched on the impermeable paleosol.
There were no identifiable zones of chemical anomalies in the organically
stained midden such as those identified at the Poplar and Wednut sites,
indicating that hunen occupation had been less intense at this site than those
previously described. This ccxiforms with the cultural matericd and features
reoovered at the Beech and Oak sites.
DISOJSSICW AND INIEHPKETATION ^ SOIL ANALYSIS
The results of soil, chemical, and physical analysis are sufficient to
address specific research questions posed in this study as well as more
general issues of soil development in archaeological sites in a floodplain.
New information produced in this study will enable both the arxdiaeologists and
soil scientists to understand the interaction between natural and cultxiral
farces in the floo(^lain.
Ihe first question about how the midden mound landfozms developed has been
firmly answered. are natural and were formed by fluvial deposition of
sediments. Four sites known as midden mounds (Poplar, Walnut, Hickory, and
Mid Creek) were parallel bars which formed away from the mainstream during
floods effecting the entire floodplain. All fit the classic description of
tear drop cross section with a steep v^tream face and tapered edges,
consistent coarse sediments and gradation of fine particles \3p the profile.
IVro others were formed by overbank deposition (Ilex and Beech/Oak) . Ihese
sites eidiibited classic ovexbank sedimentation characteristics such as
thickness in the strata adjacent to the stream and a linear sh2pe paralleling
the stream. No eutificial site constrxiction wsis detected in any of these
studies.
Die resecuoh into the questicxi of the midden which is so
characteristic of these sites still was still so dark after up to 6,000 years
produced a wealth of infometi(»i vAiich not only addressed this issue, but the
larger issue of the characteristics of the midden zone (or anthropic
qpipedon) . For the past several decades, soil scientists have realized that
prolonged human habitation on a site alters the soil development process and
can inpart unique characteristics vAiich persist for long periods. These
characteristics are primarily related to the introduction of organic material
into the soil ifAiich masks and alters the natural processes of soil
development, weathering, erosion, and biological activity. In addition,
prolonged habitation by humans can destroy previous development in the soil.
The organic materials introduced into the soil by people were well deocnposed,
and the resulting organic oonpounds, especially hunnus, cu^tually formed a
mineral bond with the soil particles which prevented natural wetting and
swelling and enhanced oementation. This well-developed granular soil
structure and the tendency to repel water were the proninent characteristics
of the midden zones. This strong bonding inhibits the movement of humic
organic ccnpounds down the soil profile. This promotes an artificial buildup
of a "surface" horizon, and the midden zone is actuedly an "A" soil horizon
vhich is dark in color due to the presence of organics.
The unusually thick buildip of an organic zone in the midden mounds has
been enhanced by the two factors: 1) continual deposition of new sediments to
the flocx^lain landforms each year and 2) the bonding of organic ocmpounds
coating the sand matrix of the soil profile. These factors ccmbine to
produce a growing midden zone that is excluded from the natural foroes of
leaching and percolation. These middens are "greasy," ever-growing "luitps"
resulting from intense habitation ooccpation in the flooc^lain, v^ch, as far
as soil science can detect, cu?e edmost permanent fixtures in the landsc::ape.
Erosion has been the main effective force in decreasing the size of these
sites.
The chemiccil aneilysis of the midden zone documented discnrete
cx^noentrations of some chemiceds in sane of the midden profiles. This
included a high caurbon/nitrogen ratio at the walnut site between 50-150 an
(19.7-59.1 in) and two zones of high 0K«phate at the Pedlar site. Due to the
poor movement of organic ccnpounds through this midden zone, these zones of
cdiemical aixmalies cxxild vrell be related to an increaise in the site population
emd/or length of cxxnpation time.
Some movement of fine particles through the midden zcxie was documented,
however, by the presence of lamellae or bands of fine particles in the lower
half of at the bottom of the midden zone at several sites. These were were
also noted at several sites downstream in the vicinity of Psiory, Ms (Bense
) . The c:ause of the lamellae is not well understood, but it is suspeerted
that they are caused by dropout of fine particles (silt and cl^s) moving down
the profile due to peinched hi^ water tables in the soil profile, or they
represent periods of stability and pedogenic developnent.
One of the most surprising findings of the soil studies in this project
was the paleosol buried beneath the midden zcne in the floo(^lain. This
phenonenon had been recently discovered in some archaeological studies in
floodplains in the Midwest and Southeast. However, this infonnation was not
yet well known or in the genered literature at the initiation of this stu^.
Actually, soil scientists had considered the floodplain far too active and the
sediments too young to have ai^ well-developed soils. So, vdien the "yellow
brown" zone below the dark midden zone proved to have structural developnent,
illuviated clays, and chemiccd weathering equal to upland Pleistocene soils of
much greater age than the Holocene floodplain deposits, and this zone
cOTitained cultural material no older than 10,000-12,000 years, it was cause
for great interest by the soil scientist.
Sane of the first questions posed by the soil ooisultant related to the
integrity of the cultural material and the age of the tenporal markers
contained in this zone. Unfortunately, there was no reliable datable material
in the paleosol, althouc^ severed attoipts were made to recover charooed for
radiocarbon dating. Whole 1x1 m (3.3x3. 3 in) mits were fine-screened, and
the charcoal flakes were removed and sent for dating, but the dates obtained
varied widely and did not correlate with the consistent dates received for the
midden zone. Ihe best docunentation of the age of the paleosol is from the
consistent radiocarixm dates of 6,500-7,000 B.P. at the base of the midden and
the inclxision of stratified tenporally sensitive artifacts. Uie truncated
upper 20-40 an (7.9-15.8 in) of the paleosol contained Eva-Morrow Mountain
projectile point/)aiives vrfiich have been dated in the adjacait Tennessee Valley
drainage to 6,500-8,000 B.P. (Chapman ; Cridld^au^ ; Hofinan ).
Below these markers, usually in the central portion of the paleosol, Kirk
projectile point/knives predoninated. Tlus marker has been consistently dated
ais occurring in the mid-South between 9,000-10,000 B.P. (Chapman ) .
Between the Kirk and Eva-Morrow Mountain zone usually were transitional forms
such as cypress Creek and Wade projectile point/)cnife types which probably
were made between 8,000-9,000 B.P. There usually were no in situ projectile
point/knives below the Kirk, and the lower half of the paleosol was culturally
sterile. Fran these indirect data, the date of the sediments making vp the
upper half of the paleosol are approxiuBtely 10,000 years old, and the Icwer
heilf must therefore be older, perhs^Js dating to the late Pleistocene.
To document the integrity of the artifactual naterial in the paleosol,
piece plotting of all specimens was performed at the Walnut, Poplar, and
Hickory sites. This documented that the specimens were not size-sorted from
secondary fluvial deposition and that specimens crosscut the structured
features, indicating that the soil developed with the artifacts in situ. In
addition, the eu>gle (or dip) of the specimens in the pedeosol was relatively
flat indicating they were not laid down by running water. Therefore, the
specimens were incorporated into the sediments by people as they built-up
during flooding.
Soil development follows sediment deposition, and the extent of
development reveals certain characteristics of the environment in vAiich it
developed. Fran the study of the paleosol structure beneath the midden
mounds, it can be inferred fron the advanced level of develc^ment that the
landscape was stable during development, i.e. after the sediments were laid
down and before they were tnincated (post-7,000 B.P.) . In addition, the
network of polygonal cracks intersected the soil structure and therefore must
have oocurxed after development tocdc place. This reflects a period of
envircnnental desiccation. This information suggests the following scenario:
1. The sediments of the paleosol were deposited on a swanpy and possibly
eroded floo^lain or eroded stranded surface of Pleistocene outliers in a
relatively hi^ energy fluvial envirorroent during the late Pleistocene to
approadmately 7,000 B.P.
2. D^xssiticnaLL rates steadily decreased, and a well-developed soil formed
before 6,500 B.P. This is probably the onset of the AltithemiELL (or
Hypsitheimal) climatic episode.
3. A period of desiccation occurred in which the soil actually cracked
forming the polygonal network in the soil. This probdoly correlates with
the xeric maximum of the Alti thermal climatic episode.
4. Catastrophic flooding took place approximately 6,500 B.P. and removed the
A horizon of the soil followed by regular deposition of fluvial deposits
vrfiich continues to the present day. This liJcely correlates to the onset
of modem conditions (Meditherroal) and the end of the xeric Altithermal
climatic episode. Human oocrpatirxi of these landforms increeised in
intensity at or before this time and continued throu^iout the prehistoric
period, greatly altering the soil and inpeding soil development.
Fran the information produced in this detailed analysis of the soil from
six archaeologiced sites in the floodplain of the Toibi^see Valley, much has
been learned about soil development processes, landscepe evolution, late
Pleistocene and Holocene climatic episodes, and the effect of intense human
oocwpation on the soil. The three~part Holocene climatic sequence is
ocnsistently reflected in the floodplain sequence of depositicm and
development of soils. Since the Altithermal climatic episode, hvmans have
occupied these hi^, well-drained knolls in the floodplain with consistent
intensity, drastically altering the soil.
f
208
CHAPIER VII BCmmCKL STUDIES
As part of the interdisciplinary approach to the stuc^ of the remains frtxn
these eleven archaeologiccil sites, botanic2d studies were edso performed.
These studies included: 1) an aneilysis of existing vegetation on or near
several sites investigated to quantify the "modem" vegetation ocmposition in
the study area; 2) an assesanent of the vegetation history in the mid-South
frcm the late Pleistocene to the present to better understand plant connunity
(fynamics; and 3) an analysis of plant remains from the archaeological sites
investigated. Vegetation history was derived largely from palynological
studies performed in wetlands near the stufy area. Botanical information
complements the geomorphological studies to produce a holistic per^iective of
the effect of the climatic changes of the Holocene on the landscape and plant
ccnnunities in the tapper Tombic^see Valley.
Plant remains research addressed questions relating to subsistence, site
envirc»inent, and relative density of occupation. Most of the botanical
remains vgere charred plant fragnents. Several attenpts were made to recover
pollen; however, it was generedly not preserved in the well-drained midden
soils. Pollen was found in one site (11^: 22It590) , and it produced
interesting information on the Early Amhaic enviroment at this localify.
Most of the information produced by the botanical studies concerns use
patterns of plants. Iftifortunately, charred plant remains r^mesent a skewed
sanple of the range of plants \ised in the past and they may be limited to 10%
of the total utilized flora. Although these remains represent only a small
fraction of the utilized plants, analysis indicates that there were patterned
changes throu^ time.
VEGETEAngCVL HISTOFBf OF TOE MID-S(XJrH
The present climate of northeast Mississippi is temperate and humid with
mild winters, warm svnmers, and abundant precipitation. Rainfedl distribution
is usxially hitter in winter and spring months and lowest during the fall
season.
Past environmental conditions in the mid-South have been studied primarily
through pedynology. At Nonconnah Creek, Tennessee rear Menphis (ca. 160 km:
100 mi northwest of this study area) , Delcourt and Deloourt () indicate
that the climate was colder during the late Pleistocene. A series of samples
representing 23,000-13,000 years B.P. shews spruce (Pioea spp.) deminating,
with fir (Abies spp.) and larch (Larix spp.) present. Continuous
representation of ircxiwcxxi/hcp-hornbeam (Carpinus caroliniana - Ostrya
virqiniana) , ash (Fraxinus spp. ) , bircii (Betula spp. ) , beech (Faqus spp. ) ,
maple (Acer spp.), cottonwood (Populus spp.), willow (Salix spp.), elm (Ulmus
1^.) , vibumam (Vibumyi spp.) , and walnut (Juglans spp.) pollen svpports
their hypothesis that dissected terrain adjewjent to north-south trending
rivers throu^xxit the southeastern United States served eis refuge arecis for
deciduous tree species during the full glacial period.
The later glacial eind post-glacial history of south-oentral United States
vegetation may be traoed from peilynological data obtained from numerous sites
south of the glacier's edge. As the climate ameliorated, cool temperate mixed
Mesophytic forest species spread north £d.ong the Appalachian Mountedns and the
Allegheny and Cixiherland Plateaus. By 5,000 years B.P. the wanning and drying
trends of the fypsi thermal had their maxiimm effect, and mesopl^ic species
became restricted to northern latitudes and high ailtitudes, while the preiirie
spread as far east as eeistem Missouri, and xeric oak-hickory-ash forest was
present in central Tennessee.
209
After 5,000 years B.P., southern pine (Pious) species bec£Bae abundant cn
the Coastal Plain due to the increeiaed doninanoe of the tropical inaritiine
ainssss from the Gulf of Mexico. At the aouthem end and west of the
^jpalachian Mountains, ho%iever, a mosaic of deciduous and ccxiiferous forest
developed and persists today. Near Coltmbus, Ms, for exanple, sweetgum
(Liquidariaar styraciflua) became more inportant after 2,500 B.P., tupelo
(Nyssa sylvatica^ and black gtxns (N. aquatica) increased iintil 2,300 B.P. , and
pine has increased continuously since 2,500 B.P.
Today, in the study area, the x^lan^ vegetation is doninated by
short-leaf (P. echinata) and loblolly (P. taeda) pines vdth blackjack (Quercus
marilandica) , post (Q. stellata) , ^>ani^ (Q. fadcata) , and white (Q. alba)
oaks OT the loMer slopes. CcRinon associated on the ridges are the rock
(diestnut (Q. prinus) and black (Q. veltitina) oaJcs, dogwcxid (Comus florida) ,
and hickorTes
Qy
m
•
rs
• 4
fO
•
m
•
rH
«
•
•
n
•
in «-4
in
•
oo
•
cs
•
fO
tN
(N 00
r-
Ov
IN
0
vo
m
w w
in
00
m
CO
i
j<
^ V
^ ^ 0-0
m ^
in
4J
4-1
4.1
CO
yo
^ A -H
VO *H
VO •H
14
cn ^
q
Ui
■e4
«H
•H
4J o
HI
4J tjO«
M U
M
4-1 04
M
3
n ied from at least the Early Archaic (Kirk) through the
Late Woodland periods.
The focus of the investigations at this site was the Early Archaic
ccnponent, and botanical materials from 1,482 liters of soil frcm three
features and 27 midden samples weire analyzed in this study (j^pendix II:
Table 4) . The midden samples were large volume of soil ranging from 2-68
liters with an average of 43 liters and were taken at 5 an (2 in) intervcds in
the control block in Block D. The increased vol\xne mxifaer of sanples studied
and the closeness of sampling reflects both the focus on the Early Arch6dc and
the integrity of the deposits at this site.
Unfortunately, in all the midden sanples studies, a total of oily 7.3
grans of floral material was preset, cdthough 11 sanples were only partially
analyzed. The presence of floral material while low (averaging only 0.2 grans
per sample) , it vas very consistent, and no sample was devoid of material.
In the midden sanples, hickory nutshells were not the most frequent floral
remains (42.5%). Wood remains were the most frequent (56.2%) floral type with
only a few aoom husks, six fern spores, and ten unidentifiable seed
fragments. The floral material was most abundant in the upper 30 cm (11.8 in)
of the Early Archaic deposit.
All features analyzed from the Hickory site were pits, however, only cm^e
(Feature 4) could be affiliated with the Early Ardiaic ccnponent. The other
two pits (Featxues 3 and 6) could not be associated with any specific
ccnponent. In the features, hickory nxitshell ocnprises an average of 90.7%
vAiile aoom shell and wood comprise 0.7% and 7.8%, respectively. Five fern
spores and one indeterminate seed were edso recovered from the features. It
should be noted, however, that the percentage of hickory nutshell frcm the
Hickory site is Icwer than frcm the other sites studied. Oversdl, the
concentration of plant remains from the Hickory site ranges from 0.-0.08%
wei^^it/vDlumB (i^pendix II: Table 5) . In an effort to obtain more floral
sanples, sediments from the enlarged control block units (1x1 m: 3. 3x3. 3 ft) ,
accept four liters of sanples to be preserved in perpetuity, were processed.
However, this did not yield proportionately more botanical remains. The
paucity of plant remains may have resulted frcm soil and flcxx^lain loc:ation,
vAiich may have accelerated the rate of organic decenposition. Alternatively,
short ooctpations and/or the function of the site mey accxxint for the paucity
of plant nateriads. This scarcity obviates any postulation concerning
seasonality or the prehistoric vegetation at the site.
THE BEECH AND OAK SITES (22lt623 and 22It624)
The Beech and Oak sites were located on two adjacent fragments of an
abandoned flooc^lain levee of the Temibi^see River (Figure 17) . Althou^ the
sites have separate numbers, they are considered as one for analytical
purposes. This locality was cxxxpied &xm the Early Archadc through the Late
Woodland periods. The focus of the investigations at these sites was on the
Late Archcdc conpcxient, for in eill other sites investigated it had been
disturbed. In addition, this was one of the few flooc^lain sites investigated
that was not a madden mound with the dark organically stained midden zone.
This site had only a moderately dark midden zone with eeisily detected
features.
A total of 413.6 grams of floral material frcm 3,588 liters of soil was
identified frcm the Beech and Oak sites (App^idix II: Table 6) . Plant remains
are deminated by carbcxiized hickory nutshell. Other minority plant remains
221
include caxtxmized aoom nutshell, ring-porous hardwood, pine, grape, fern
spares, pokeMeed, unidentifiable seeds, and hardwoods.
The midden sanples were obtained from Block A of each site intermittently
frcm the surface to the base of excavations and were a standard four liters
each. The amount of floral remains in the midden samples was quite low
(40.4 g) and they usually were dominated by wood material (75.2%) , with
hickory nutshells amounting to 24.3%, and aoom to 0.5%. The majority of the
wood (93.4%) came from one sanple only 20 cm (7.9 in) belcw the surface at the
Oak site, and it is not representative of the middens. With this exception,
the average florziL material was low (1.1 g) per sanple. TWo seeds were
reooveired, one pokeweed and one fern spore, both fom the Beech site.
Of the nine features from which floral remedns were scandned three were
well-defined Late Archaic pits (Features 9, 11, and 14) and one was a pit of
unknown cultural affiliation from the Beech site (22It623) . Five features
were studied from the Oak site (22It624) : one Late Archaic pit (Feature 1) , a
Middle Archaic (Benton) ocnpound pit (Feature 7) , a mixed Late Archaic and
Mississippian pit (Feature 9) , and one pit of unkncwn affiliation (Feature 4) .
A total of 373.2 grams of flor2d renains were identified from these nine
features. This represents 90.2% of the floral materi£d. &om both sites and
iidicates the concentration of plant material in features in general as
opposed to the midden. Over half the feature floral material, however, cane
from the Benton ocnpound pit (Feature 7 at the Oak site) . Samples of 16
strata within this pit were analyzed, and 2d.l but one were docimented by
charred hickory nutshells at an average density of 93%. Aoom busies were
present in six strata (5.4%) , and wood fragnents, edthough low in frequency,
(1.7%) were present in all but one. The samples frcm this feature also
contained a grape seed, four fern spores, and two unidentifiable seed
fragments. The four Late Archaic pits exmtained a total of 73.9 grams of
plant remains, but most (75.8%) were in twa pits (Feature 14 at the Beech site
and Feature 1 at the Oak site) . All of these features were deminated by
hickory nutshells (96.5%) , with traces of acorn husk in three pits, and small
amounts of wood in all. One pit at the Oak site (Feature 1) had a possible
geranium (Geraniim carolinianuro) seed. The other pits followed the same
pattern of plant material, except the mixed Late Archaic/Mississippian pit at
the Oak site (Feature 9) had 618 fern spores in it - far more than any other
feature and 97.2% of all recovered frcm the samples. A grape seed was also
identified in a pit of unknown cultural affiliation (Feature 2) at the Beech
site.
The relative densities of floral remains varied frcm 0.08-0.004%, except
in Feature 2 at 22It623 (0.3%) and Level 2 at 22It624 (0.5%) (Appendix II:
Table 7) . The concentrations from the Walnut, Poplcu:, Ilex, cuid Aralia
samples varied frcm 0.001% to greater than 4% for feature fills and was less
than 1% for general levels (Sheldon ) . These differences may be
attributable to a nrmber of factors, including site function, duration of
occipaticxi, se2ison of ocxnjpaticxi, and soil type.
An inference can be made on the vegetation at the Late Archaic Beech and
Oak sites based i;pon the identified wood fragnnents and the abundant fern
^x}res. The sites were probably covered with a mixed hardwood forest with
heavy canopy, low herbaceous undergrowth, and probably few clearings.
SITE 22It606
Site 22It606 was Iccated on a Pleistocene terrace remnant overlooking the
Tenbi^oee floodpleiin and was the only "vpland" site investigated in this
project, nils site had been occi:pied fxtxn the Early Archaic through the
Iteoent, however, the Late Woodland/Mississippian period dqxseits were intact,
e^eciidly the features, and this %«as the focus of the investigations. A
total of 293.2 grans of botanical material focm six features was identified
from this site (Appendix II: Table 8) . Since midden samples lacked integrity,
th^ were not analyzed.
Samples from six pit features were analyzed from this site: four Late
Woodland/Mississippian (Features 18, 19, 20, and 45) , one Middle Archaic
S^ces-ftiite Springs (Feature 28) , and one mixed Late Ardiaic, Woodland, and
Historic (Feature 30) . A total of 293.2 grans of carbonized plant material
was identified from 1,167 liters of soil and was ocnposed of 41.7% vrood
fra^nents, 30% hiokory nutshells, 10.9% aoom liusks, and 17.4% unidentified
material, along with 537 seeds.
Most of the material (78.7%) was in two Late Woodland/Mississippian pits
(Features 18 and 20) . With only one exception, all plants identified are
wild, vAiether from Archaic, Woodland, or Mississippian features. Most (41.7%)
of the sample is wood, looth piive and hardwood, and hickory pericarp. Ihere
are some aoom pericarp fragments and a variety of seeds of grasses, weecty
plants, and a few fruits, including persiitnon. Miscellaneous specimens
include fern spores, possible fruit skin fragments, pericarps, exines, pine
resin, acorn fragments, and unidentified seeds of various The one
donesticated plant is represented by four maize (Zea mays) cvpules in
Feature 20.
The low amount of hickory in the Late Woodland/Mississippian pits is
different from other ccRponents. The average amount of hickory nutshells is
21.5% with a range of 1.6-44.6%. The one Middle Archaic pit (Feature 28)
contained 96.3% hickory nutshells, vhich agrees well with other Archaic
sanples. The hi^ amount of wood charcoal (average 27.3%, range 0.3-62.4%) is
also unusual and ation at
this site was during the Henson Springs phase of the Late Gulf Formational
stage radiocarbem dated at about 460 B.C.
The 81.2 grams of floral material identified frem 134 liters of soil were
primarily from two major features (Features 10 and 11: 35.7 g and 45.5 g) and
midden in each of the two major excavation blocks (A and B) . The 27 midden
sanples were taken at 10 on (3.9 in) intervals throughout the profile. More
volume of midden (116 liters) was analyzed than for the feature, which
prcbably accounts for the fact that the majority of specimens (56%) recovered
were from the midden (Appendix II: Table 10) .
While a relatively diverse sample of floral remains was recovered, the
interpretation of these materials is clouded somewhat by the identification of
modem contaminants (predoninantly uncaxbonized specimens) in various contexts
within the site. Although it is unlikely that ccurbonized plant remains in
this sample have been contaminated through the inclusion of recently
carbonized specimens floating in the air, this remedns a possibility that will
require further stuc^.
Hie plant remains from both the features and midden were similar. Most
(57.3%) of the identified material is charred wood; hickory nutshells account
for 40.5% and acorns account for 2.2% of the total.
Hie plant remains from the Aralia site contained more carbonized seeds
(1,691) than any other site. Identifiable seeds include pokeweed, chenopods
(Chenopodium sp.) , persimmon, and grape, however, most seeds were
uniden^f iable . A single wild bean species fragnent, apparently of recent
origin, was recovered. Oak, pine, and unidentified hardwoods made up the wood
samples.
Hiis sample from the Gulf Formational stage is also different from the
Archaic samples in the low percentage of hickory and high percentages of wood,
acom, and weeds. Since these materials are in good context, the ccnpaurisons
with other sites is more significant.
SimABY OT MACROBOTANICAL INFORMATION
A total of 2,727.4 grams of macrobotanical remains and 3,894 seeds from
ei^t sites were identified during this project. Hie range of species
identified was relatively narrow, probably as a result of poor preservation in
acidic soils, coupled with differential preservation of porous and dense plant
parts.
The plant materials were preserved through charring, vdiich effectively
slowed the rate of deoorposition in these acidic environments. However, the
process of charring plants was probably selective, depending in part on
cultural preference in methods of food preparation, choice of firewcxid, and
modes of deposition. In addition, post-depositional factors such as pit
digging and refilling also affect the locaticxi of plant remains. Hierefore,
the sample of charred plant remains recovered from any site represents a
relatively small portion of the total plant resources once utilized by the
site occupants (Wing and Brown :147) . The charring of seeds, hickory and
acom parts from these sites may have occurred during processixig, through
secondary use as a fuel (nutshell) , or accidental chatting in a forest fire
after deposition. Roasting nuts or parching seeds to facilitate storage is
well-docnanented ethnographically and may account for their conditicai and
preservation (Sndth :109; Yamell ).
Irrespective of the preservation prc^lems, the archaeologically recovered
plant remains revealed much information and patterns which were not aveiilable
by any other means, and the data are internally cxnparable. As Tables 91 and
92 indicate, 90.3% of the plant remains were recovered from feature context
and 9.6% from the miciden samples. This is li)cely related to the fact that
81.3% of the soil volume came from features, and only 19.7% came from madden.
The density ratio, however, indicates that more plant remains were contained
in macJden soil (6 grams per liter) than in the feature fill (3 grams per
liter) . This result was unexpected, but it probably reflects the larger
volume of feature fill that was procassed.
Mhen vieMed at the site level, the sanples from Walnut, Hickory, Ilex,
Poplar, Oak and Beech were aloaoBt all from Middle or Late Archaic contexts.
Ihis sanple amounted to 6,990.1 liters of soil and produced 2,353 grains of
plant ranains, and 1,666 seeds or fern ^ores. This sanple is from the best
possible context at these sites and indicates a st3nng pattern of hickory
nutshell dcminance (98.4-83.8%) followed by minor amounts of wood and acorn
nutshell with a ratio of 90:9:1. Seed patterns appear to be site specific.
Indication of seasonality of the occvpaticns represented in the plant
remains are generally restricted to simner and fall. Nuts and acorns were
generally harvested during October and Novenber (Hudson ; Swanton ) .
This is clouded, however, by the probability that these archaeologiceLL samples
reflect both collection efforts occurring during occupation of the site and
long-term storage and transport. Evidence documents that there is a fall-bias
of plant assaiblages collected from archaeological deposits. The
under-r^mesentation of winter-spring ocxnpations must be anticipated, eilong
with the absence of many plant resources that are usually not preserved in
archaeological deposits (leaves, small fruits, roots, and fungi) . Therefore,
vdiile nuts and seeds were iaportant fcxsd resources diuring prehistory, their
role in the mhole economy must be interpreted with caution.
The ubiquity of hicdoory nutshell in the Archfdc period sites and the
strong presence at the later sites suggests its impcgtance in the subsistence
laase. The nuts mature in October and Novenber. Ttney were sometimes eaten raw
by the Indians, but more ccmnonly the oil, Jaxwn as "hickory milk” (Bertram
:57; Hudson :301) , was extracted. The nuts were pounded, and the
cracked pieces were put into a pot of boiling water. Afterward, the shells
sank to the IxTttom, and the liquid was passed through a fine strainer which
preserved the seasoning (for hominy and com caikes) . Bartram reported seeing
more than 100 bushels of hickory nuts stored for one family (Bartram :57) .
Seme patterns of usage can be seen in the plant remains identified with
respect to secxmd-line foods like acorns and seeds. Amounts of acorn husks
appears to fluctuate and pea]cs during the Benton and Mississippian periods.
Although small quantities of acorn were recovered, this is probably due to its
secondary use and/or its fragile shells. Consequently, Ch£^inan () has
suggested that the wei^t of acorn most be multiplied by ten in order to
compare it directly to the denser hickory nutshell. If the wei^ts are
multiplied by this factor, acorns represent 8.0% of Early Archaic remains;
10.5% of Middle Archaic; 31% of Benton; 25.2% of Late Archaic; 7.5% of Gulf
Formational; 21.5% of Late Woodland-Mississippian; and 88.7% of Late
Mississippian. This likely reflects a more realistic proportion of this food
source through time. Acorns ripen throu^iout the fall; Indians preferred the
sweet, white oak groi:p fruits (especially Q. virginiana) , but they usually
extracted oil from all the species (Hudson ; Swanton ) .
Only small quantities of walnut (JUglans nigra) were recovered from
cultural contexts. Diis is probably due to the fact that walnut trees were
widely dispersed in the natural forests.
Althou^ 3,894 seeds and fern spores were recovered, many were
unidentifiable. Only 40 identifiable seeds were found in 2.7 kg of carbe^zed
plant material. They eppeared, however, to occur more frequently in later
time periods. In the Late Ard^c an average of one seed was found in each
146 grams. During the Gulf Formational the oonoentration increases to one
seed per 37 grams. Many of these seeds identified are opportunistic species
cannon to forest clearings and edges of paths.
The presence of so many fern spores in the midden samples from Eeurly
Archaic through Late Archcdc corponents probably reflects the primary forest
225
environnent. A similar habitat is present today at flooc^lain elevations in
mature secondary forests. Hie canopy is closed, effectively shutting out
direct sunlight, eliminating understory shrubs and herbaceous plants and
encouraging the growth of ferns and other shade-tolerant vegetation.
Staall quantities of cane culm fragnaits were recovered from Middle and
Late Arch2dc features. Cane grows zdong riverbanks and in swanps, often
forming canebrakes. Its culnns are available throu^iout the year. Seeds were
sometimes used eis food, but most ocnmonly its culms were utilized as raw
material for baskets, mats, arrcMS, fish traps, and backing for wattle Wcills,
^nlong many other things (Hudson :287; Swanton :244) . Wood was found
in every sanple that was sorted. Pine, other gymnosperms, oak, other
ring-porous species, sweet gun, and other diffuse-porous species are
represented. Most pieces could not be identified further because of their
stall size.
Analysis of plant remains from the Archaic period indicates that the
subsistence base of the people was partially dependent upon gathered wild
plants, especicilly hickory ntitshell. Hie increase in quantities of nutshell
and concentrations of charred remains in the Sykes-White Springs/Benton
coiponent midden coincides with other evidence that the use of the midden
mounds was intense during that time. Hie large nurber of features, the
presence of structures or at leeist eictivity centers, hearths, burials, and
nuttbers of cu±ifacts also point towards long-term use of the sites as base
canps for several residential groqps.
Gulf Fozmational plant remains from the Henson Springs oompcsient at
22It563 indicate an economic reliance on gathered foodstuffs dominated 1:^ nuts
and seeds. The identified plant resources occur in the Toebi^see bottoms even
today and presumably could have been obtained with little difficulty
throughout most of prehistory. It is possible to infer a subsistence base
dependent on the scheduling of gathering-hunting-fishing activities. Hiis
proposed eoonony is comparable to subsistence strategies documented at a
similar time depth in sites throu^iout nuch of this region and the
southeastern Ikiited States (cf. Dye ; Morse ; Galm ). The origins
of this pattern lie in the Archaic period, and it apparently continues with
only minor changes until the widespread adoption of agriculture and an
attendant shift in economic enphasis.
The botaniccLL assemblage fron the Late Woodland-Mississippian
(A.D. 100-1,500) , represented by 22It606, is clearly similar to previous
assemblages. However, the large quantity of acorn suggests a widening of the
food base. In conbination with maize, large amounts of acorn are typical of
Late Mississippian eissenblages further down river >^iere acorns represent a
sv^jplementary food souISa in areas of low fertility or years of decreased
domesticated crop productivity. Settlement must have been repeated,
intermittent, lew-density, and short-term in nature. Perhaps agricultural
grovps spent brief periods at gathering/hunting stations sipplementing their
maize diet, or perhaps there was less enphasis upon intensive agriculture in
this hinterland area.
In conclusion, the roacrcbotanical information, while hampered by poor
preservation, heis demonstrated that plant resources were remarkably similcur
through time and were present in sufficient quantity to preclude any major
shifts in procurement strategy. Hiis must have been an impor1:ant factor in
the cultural continxiity docunented in other facets of this stucty.
226
ANALYSIS OF BIOSILICAIIES FRGM ABCHAEQLOGICAL SITES
As noted in the previous section. Botanical Analysis of Archaeological
Material, the preservation methods of cdiarred plant remains leaves a biased
and inoonplete record of the use of plants by the past ooctpants. In an
attempt to recover other plant remains for a more representative sannple to
stud^, a biosilicated (^diytolith) analysis vms undertaken.
A {d^ytolith is a dq)osit of opaline silica that forms in a plant cell and
subsequently is deposited in underlying sediment upon death and decay of the
plant. Phytoliths have many shapes and range in length from less than tvio
ndcrons to one millimeter. Plants are not represented by a siaple piytolith,
but rather by an assemblage of phytoliths (Mooc^ ) . The deposition of
ptytoliths for the most part is local, vihich enhances their value as
indicators of subsistence paleobotcuiical ocmnunities, and potentially,
paleoenvirontents .
A test anedysis was conducted, with a concentration c»i the Middle Archaic
deposits, on samples from 22It539 and 22It576 to determine if biosilicates
were present and if they could be Identified from both feature and midden
context. Ten sanples from each site from features and midden were examined
for opal phytoliths. Plytoliths were extracted using a modified version of
the techniques first developed by Rovner () . Sediment samples were dried
aiKi successively treated in solutions of sodium hexametaphosphate and
distilled water, HCL, and distilled water rinses, and floated to facilitate
extraction in a solution of tetrahronoethane and absolute etlyl alcohol.
Following extraction, samples were mounted on microscope slides and subjected
to both scanning electron microscope (SEM) and Ncmarski optical study for
identification of fdytoliths.
Phytoliths were present in all samples frcm 22It539 and 22lt576 and are
presented in Table 11 of Appendix II. IMfortunately, they could not be
identified, and further analysis is required using modem plant species as a
gxiide to the identification of plants fron archaeological sediments. At
present, a key for the identification of biosilicates of plant ^secies from
the southeastern United States does not exist. A more detailed analysis of
samples derived from the Upper Tcrobigbee Valley, therefore, must await further
study and the developmmit of a specific plant ki^ for this area.
It was logical to initiate the phytolith study, since the charred material
recovered represents only a fraction of the plant material once present. As
phytoliths are not subject to decay as pollen is, and are present in all plant
parts, hence it was likely that they would be preserved in both midden and
features. Piytoliths could, therefore, give a first view of the roots,
tubers, leaves, fibers, and other currently "invisible" plant resources on dry
sites that must have been used and have been documented at wet sites.
Although the abundance of phytoliths varies considerably annong the
sanples, they contained sufficient phytoliths to warrant a detailed
investigation using modem plants to develop an identification k^.
Paleoenvironmental and paleoagricultural reconstruction should be possible if
ocnparable modem plant taxa can be utilized.
POLLEU ANALYSIS FRCM ABCHABOLOGICAL SITES
Sediment sanples from four sites (22It539, 22It576, 22It590, and 22It621)
were subnnitted for analysis during the course of this project. Initially,
sanples were sent from both midden and feature contexts frcm the first two
sites excavated in Phase I, 22It539 (Walnut) and 22It576 (Poplar) to determine
227
Possible e}q)lanatlons of the Ilex pollen data may Include the sanpling of
older redeposited sediments, or, possibly, the persistence of a relic stand of
boreal forest on this Pleistocene outlier.
In Phase III, this explanation was tested. Samples of sediment fran the
Eeirly Archcdc deposits from the three other midden mounds containing this
component were sufcmitted for analysis. In addition, sanples of sediment from
eadti of the sites from beneath the Eeirly Archaic material were submitted to
determine the vegetation pattern.
228
No boreed pollen was present in ai^ of the other site sanples, and only
typical Early Holocene plants were identified. Diis leaves the boreal taxa
&om 22lt590 as an anomaly, and suggests that a relic stand in a cold pocket
^cisted there during the Early Holocene.
sLWARy OF BcnaNicaL studies
Botanical studies were a major part of the research conducted during this
project, and much infonnatian was produced from them that otherwise would not
have be^ avzdlable. Ihere were several aspects to these studies which
included: 1) a quantification of the present vegetation on several sites under
stuffy as well as nearby areas; 2) research on the vegetational history of the
mid-South during the late Pleistocene and Holocene; 3) identification and
analysis of charred plant remains from archaeological context of the ei^t
sites which were intensively excavated; 4) phytolith identification and
analysis of a sample of midden and feature soil; and 5) pollen analysis of
systematic samples from profiles, qiecific micro-preservation environments,
and specific midden/profile samples.
The studies of present and past vegetation in and around the stufy area
have provided a perspective vhich complenants that provided by the
gecmorphological research presented in the previous chapter. Essentially, the
late Pleistocene vegetation was characteristic of a colder envirorment than
today with a dissected terrain and a forest cover dcminated by boreal
coniferous trees but also containing some deciduous species. By ca.
5,000 B.P. the climate had wanned sufficiently so that the boreeLl coniferous
trees had migrated north out of the mid-South, and the forests were dcminated
by pine. By ca. 2,500-2,000 B.P. sweet gun and tupelo became major parts of
the forest, and pine had increased continuously. The wetlands ocxitinues to
oontedn a deciduous forest, unlike the X5>lands. The present-day vegetation of
pine forest uplands and deciduous wetlands was in place, therefore, ca.
4,000-4,500 yeajrs ago.
The pollen studies conducted in this project, while generally unproductive
in cultural terms, did add new information to this vegetation soeneurio. The
pollen from the Ilex site (22It539) in Early Archaic (ca. 9,000-10,000 B.P.)
context confirms the presence of boreed. trees (^ruoe/fir) in the site area at
a time later than that projected by a specialist in this field. It appears
that the Ilex site was and still is in a cold pocket of Mackey's Creek and
supported a relict stand of cooler-loving trees. Similar islands of boreal
^»cies exist elsewhere in the region. For example, at "Natural Bridge" in
northwest Alabama, ca. 20 mi (32 km) eaut of Fulton, Ns a relict stand of
hemlock still exists today in a cool, moist enclosed depression formed by
underground and surface streams.
The stufy of plant remains from eux^aeological context ccmprised the bulk
of the botanical work. A large volxxne of soil (8,290.8 liters) was floated,
cleaned, and the plant material hand-picked for Isotanical studies. Samples
from 47 features and 12 excavation blocks produced 2,727.4 g of chcurred plant
fragments and 3,894 charred seeds and fern spores. The analysis of the Icind,
distribution, and density produced interesting results. One major discovery
was identification of a possible Archaic "botanical signature" of 89:10:1
(hickory :wood: acorn) in both midden and features. This pattern changes during
the Gulf Formational and Late Woodland to 29:64:7 and appears to reflect more
use of second-line resources, probably due to population increase. The use of
fruits also appeeurs to increase in the post-Archcdc deposits.
229
The seasonal ocx:vpation for most sites was sparing throu^ fall, although
winter was hard to docurnent. Ihe ccninon problem of food storage and bi6uses
preservation plague the time-of-occu^ticn issue.
Another eispect of the information available from the archaeobotanical data
was the "intensity" of site ooci^tion. The amount or density of plant
material in the midden sanples was tanevenly distributed during the Archaic
periods. There was a definite peak in the amount present during the Middle
Archcdc Sykes-White ^arings/Benton period. In all sites examined, increases
of up to 6,300% were documented during this time. It is hypothesized that
this represents a longer term \ise by more people. Ihese data correlate well
with other indicators of increeused site use during this period.
Site floral environments were also reflected in the charred seeds and fern
spores in the sanples. VAiile most samples fron all time periods reflected a
closed forest cancpy, the post-Archedc deposits had many more seeds from
disturbed habitats, perhaps reflecting more use of the sites, old field
vegetation, and more clearing within site areas.
Ihe ejperiment with phytolith analysis was both interesting and
frustrating. The phytoliths were abundantly present in all contexts examined
and likely are from roots, tubers, fabric, cordage, other plant products and
foods vAiich have decayed beyond common recognition. Uie absence of an
appropriate identification key, at this time, precludes the identification of
the plants vrfiich produced the pJ^rtoliths.
230
CHAPTER VIII LITHIC ANALYSIS
Ihe most abundant material recovered from the eleven sites discussed in
this report are stone tools and the by-products of their manufacture. While
this an enormous amount of material, most (ca. 60%) were recovered from mixed
context. Although lithic assemblages were recovered from Early Archcdc
throuf^ the Mississippian periods, only those fron the Archaic were
represented by both midden and feature material and in good context. A sample
of these Archaic assemblages were the focus of Phase III lithic analysis.
Since distinct Archaic components were superdrposed at several sites, and
conpcxients from the same time periods were found at several sites, both
diachronic and synchronic veurlation in lithic manufacture and use could be
studied. Heat treatment, replication, and functional studies were inplemented
to develop an vinderstanding of the manufacture ecnd use of stone tools and to
develop the variable schemes that were ^splied to curchaeological specimens.
Both macirosoopic and microscopic techniques were used to investigate the
assenhlages. Ihis chapter describes the lithic stuc^ in detail.
The first section presents the theoretical perspectives, appropriate
background, and research questions involved in the lithic stiufy. Descriptive
and analytical questions based on the preliminary assessment of lithic
cLSsemblages are posed; likely explanations for changes in technology are
e;q>lQred; and expectations for the composition of lithic cussemblages are
specified. The second section deads with sampling techniques and general
laboratory procedures of this large data set ^diich affect the interpretation
of these a^-^anblages. Ihe third section describes the eaqjerimental program
and includes the results of heat treatment, replication, and use-wear
e;q»riments. Ihe fourth section presents the variables selected for analysis
of over 6,000 chipped stone tools included in the analysis. These variables
were selected to answer the research questions posed and reflect the results
of the experimental work as well. The fifth section presents the results of
the analysis and addresses the specific questions posed in section one.
SECTION 1; THEORETICAL PEKSPECTIVE AND RESEARCH QUESTIONS
Since the recognition and acceptance of ceraunia "thunderbolts of Zeus,"
as ancient, man-made tools, relatively indestructible stone artifacts have
provided abundant evidence of past lifesrays. Traditionally, stone tool types
hctve been used to describe and ocrpare cultures, and to establish
chronologies. These goals assure that the ccmbinations of morfhological
characteristics used to define tool types reflect ideas about hew an ideal
tool should be made. Since shared ideas are the essence of culture, the
greater the resemblance of tool types between assemblages, the closer the
cultural affinities. Changes in tool types are visucdly ascribed to direct or
indirect diffusion of ideeis between groups, or to developnents in groi:p
stylistic preference through time. Using this approach, material (Ejects can
contribute to the description and history of cultural similarities and
differences. The initial examination of midden mound lithics was, in part,
designed and implemented with this traditional approach as a model.
Although description is a necessary first step in artifact analysis, it
can not provide a framework for expledning v*y similarities and differences
should exist among archaeological assemblages. Alternatively, a systems
approach which articulates material objects, human behavior, and envirOTiment
can provide ways to explore functional relationships among archaeological
remains and can provide a structure for testing causal hypotheses. If
archaeological ranains are viewed as the results and reflections of behavior,
relevant artifact dinensicns linking material items and behavior must be
identified and made explicit (Binford ; Schiffer ) . Ccnnections
between material objects and behavior are often esqpressed as schema for lithic
"life cycles" including procurement of raw materials, steps in manufacture,
use, maintenance, recycling, and discard (Collins ; Schiffer ; House
) . These mo^ls predict the products of manufacture at different points
in the cycle and their entry into the archaeological record. On aiK>ther
level, one can relate artifact variables to cultural subsystems (Knudscn
) : attributes of production, utilization, and style can be used to
elucidate eocxicmic (in the sense of systematic provisioning) , social, and
ideological behavior. Our aim in the Phase III lithic analysis was to use a
systems approach to explore possible changes in lithic technology during the
Archaic in the Upper Tcmbigbee Valley. To acconplish this, lithic technology
must be seen as only one aspect of a subsistence/settlement system. It is the
interrelationship of environment, both physical and social, and technology
that produces lithic assemblages.
LITHIC TEJCHNOLOCy
Data collected during the initial two phases of the midden mound project
raised other questions v^ch needed to be addressed in the examination of
lithic technology in the Upper Tcmbi^aee Valley. The questions involve both
descriptive and analytical levels of investigation. These data may be
discussed under three cispects of lithic technology: style, manufacture, and
use, although they cure all obviously interrelated.
STVLISnC VARIATION
Bifacial tools modified for hafting are usually the most refined,
patteimed tools found at Archaic sites. Whole tool shape in plan view and
cross-section, haft configurations, and pressure fleiking patterns are usually
thought to express time-sensitive stylistic information. The traditional
southeastern projectile point/knife typologies (Carabron and Hulse ; Ensor
; Futato , ) were used in the first two phases of this project and
allowed the identification of archaeological cotponents to broad sequential
cultural periods. Hie specimens in the stylistic type categories, though,
shewed considerable variation. Classification difficulties are due to many
factors - the polythetic nature of type definiticais (Hicroas ; Johnson
) , the plethora of names given to objects with very similar
characteristics, the level of familiarity of laboratory personnel with
regional typologies, the frequent occurrence of resheu^iening and reworking of
tool parts, and the incomplete nature of most archaeological specimens.
Therefore, one aspect of the Phase III analysis was a limited exploration of
w^s in v^ch to make the classification of hafted bifaces and investigation
of the correspcaidence of specific types with stratigraphic sequences more
cAijective. Four continuous variables (haft length, neck width, base width,
and haft angle) and three discrete variables (haft type, base configuration,
and haft treatment) were recorded for the hafting elements of bifaoes during
Phase III, and statistical techniques were used to evaluate types and show
their relationship to time periods (Davis et al. ; Johnson ) .
Concurrent examination of the manufacturing sequences and the function (or
functions) of these hafted bifaces were made to determine if differences among
types are indeed stylistic.
MANUFACTORING \QtflEmGN
Possible changes in lithic technology had been identified in the
preliminary analysis frcm the informaticn produced on lithic raw material,
heat treatment, and identification of stages of bifaoe manufacture.
Basically, the lithic assemblages at all the midden mound sites eQ:pear to be
the product of bifacial reduction. Diere are relatively few formal vnifacinl
tools, and utilized flakes are often bifcK^ial reduction flakes - the character
of these bifaoe assemblages differs throuc^ time and from site to site.
Variation is most obvious in t^pes of hafted bifaoes; however, specific types
require different manufacturing strategies, cores, preforms and ddoitage are
all potential souroes of information about the ch^acter of the finished tool
assemblage. The midden mound esqierimental program was designed, in part, to
document the manufacturing strategies fcnr different bifaoe types and to telp
relate the products of manufacturing processes to the b^iaviors that produced
them.
Ihe primary materials used for lithic manufacture throughout the Archaic
were the locally available cobbles derived from the Tuscaloosa gravels
(Camden, Yellow Chert, and Pickwick cherts) . During the latter part: of the
Middle Archaic ijifxnrted blue-gray Port Payne chert became the prcmdnent raw
material for tool manufacture. The presence of cache blades, most likely tool
blanks, nede fron Fort Payne chert and the lack of large pieces of Fort Payne
debitage indicate that initial stages of production \ising this material took
place away from the midden mound sites. Before addressing questions about v^y
Fort Payne chert should appear in the Tcmbi^see Valley or by what mechanism it
vms brou^t into the valley, it was necessary to document how nuch of this
material is actually present at sites at different time periods, hew its
frequency varied from site to site, and vhat kinds of artifacts are made from
this material.
Phase I and II analysis showed that approximately 90% of the chert used
for tool manufacture at these sites have been exposed to heat. Exposure to
heat ms^ have been intenticnal (heat treatment) , unintentional (heat
alteration) , or both. It was important to knew to vbat extent and at >Aiat
point in the manufacturing sequence both intentional and unintentional heating
occurred, and if these processes occurred to the same extent for all cherts at
all sites and during all time periods.
FUNCn(»iAL VARIATION
Althou^ several of the morphological categories used during Phase I and
II iirply function, i.e., scraper, knife, drill; no use-wear studies were
conducted, and no attempt was ma^ to reccm:d specifically functional
variables. Yet knowledge of tool functicmi or functions is critical to the
xmderstanding of tool manufacture, final tool morpbology, tool curation or
disposed, and finally site use. Are different tool types used for different
tasks? Are tools special purpose or resheupening and reuse rather than
manufacture? Are tools expediently or intensively used? Vfere sites used for
the same range of activities at all time periods? To understand slylistic or
roanufact\nring variation, it is necessary to be able to control for tool
function.
ARCHAIC SUBSISTEMCE/SETTIfl'Dn' ^fUrrEOi
E}q>lanaticxi8 for dianges in tedmology are only possible within the
context of a subsistence settlement systen. Analysis of features,
geoDtorphological and floral data, and paleoenvironnental reoonstruction, as
well as other archaeological studies of the Archzdc in the Upper Tcxnbi^^
Vialley, provided the framework for the lithic stu^. Several possible
scenarios or hypotheses were suggested in earlier cdiapters to explain the
changing lifeways during the Arch2dc and for the function of the sites during
this period. Given the information available at the time of the study, the
following factors c^jpeared most influential in affecting the nature of midden
mound lithic assemblages; i.e. resource availability, contact with outside
cultures, georooxphic development of the valley, and changes in mobility
strategies. Ihese will be discussed in order of presumed iaportance - from
the least conpelling to the most.
Ihroughout the Archaic period, the Tcmbi^see Vall^ was probably a
food-rich area with a wide variety of plants and animals available seasonally.
Although faunal remains from the midden mounds are poorly preserved, the
floral renains shows consistent use of hickory nuts, acorns, and seeds
throughout the Archaic. Environmental studies in the lower Illinois River
Valley (Asch, Ford and Asch ) and the Duck River Basin (K11{^1 ) , as
well as the Tcmbi^aee Valley (Muto and Gunn ) , indicate that river valleys
may have been less effected by the fluctuations of the Holocene climate, and
there is also no reason to suspect that ary one location in the research area
would be better for procuring subsistence resources than another. For the
present, it is assxxned that all stJibsistence activities could have been
performed at aiiy of the flooc^lain sites and at all time periods investigated.
Stone tool manufacture is obviously dependent on the types, quality, and
abundance of lithic raw materials available. In the Upper Tcmbigbee Valley
several sources of local raw materials for tool manufacture were, and still
are, available. First, chert cobbles (Camden, Yellcw Chert, and Pickwick) ,
ferruginous sandstone, and conglcmerate derived from the Tuscaloosa formation
were avedlable in the river valley from gravel bars deposited during the late
Pleistocene. One such buried gravel bar exijacent to the Ilex site (22lt590)
was documented, and it is possible that the other sites were located near
gravel bars which are now buried, althou^ there is no direct evidence for
this svpposition. 10(32^, floo^lain gravel bars in the OTV and major
tributaries are covered by alluvial deposits and are not e^qposed even in the
main channel. They were probably buried during the Early Archaic, a period of
rapid deposition on the flocx^lain. On the other hand, the buried gravel bars
may have been exposed periodically by shifting river or stream channels
throughout the Archaic.
In addition to mean valley sources, chert cobbles and other materials are
available in the floodplains of tributary streams. One such source is
22Itl026, a gravel deposit along Rock Creek 8 km (5 mi) north of the study
curea. Gravel deposits are also avedlable in the inlands. Pickwick and Camden
cobbles are major components of modem gravel quarries near Beldon, Ms, 30 km
(19 mi) northeast of the stucty area. Here ocbbles in a range of sizes and
quality cure found near the present-day surface. Both Camden and Pickwick
cobbles used in the Phase III experimental program were gathered fortuitously
along the v^LLl^ margins near Fulton, Ms. Although valley floor gravel bars
nay have been gradually buried, other sources of the most cormon stone used
for artifacts were available in the upland gravel sources vdiich lie between
the sites and the nearest blue-gray Fort Payne chert source area. It is.
234
therefore, tmlikely that Fort Payne chert was brought into the valley solely
as a subetitute for dwindling local cdiert supplies.
Increase in Fort Payne chert and the doninanoe of Benton hafted bifaces in
the sites investigated during the latter p£u± of the Middle Archaic indicates
increased contact with areas outside the Vpper Torbigbee Vcdley. Contact may
have been in the form of population movement between the Torbigbee and Middle
Tennessee valleys and/or trade in goods and ideeis. The reasons for increased
social interaction at this time are not leadily apparent, although it is
tempting to suggest movement of populations from the intervening vplands
during the Hypsithermal.
As the tapper Toibic^jee Valley attained its modem configuration, potential
habitation sites in the floodplain may have decreased. Alluviation would have
decreased vedley floor relief, and fewer high spots imnune from seasonal
flooding may have existed. Nonerous small, buried Archaic sites in low lying
areas of the modem flooc^lsdn have been identified. Many of these appear to
be Early Archaic, single coiponent sites (Bense ) . Floodpledn landforms,
s\x:h as point and parallel bars and levees vdiich continued to increase in
elevation above the general valley floor, would have been prime aresis for
occupation and may have been used more frequently and for longer periods of
time in the later part of the Archaic. The oonsequenoes of more intensive
site ooctp>ation are discussed below in conjunction with changes in mobility
strategies.
In recent yesurs mcbility strategies, schemes for moving hatan groips to
resources or vice versa, have been a major focus of hunter/gatherer studies
(Binford , ; Carlson ; Kelly ). Studies of Archaic settlement
patterns in the lower Illinois River Valley (Brown and Vierra ) and the
Duck River Valley (Amick ) indicate a shift from residential mobility in
which small groups of producers and consaners move as a unit through a
seasonal round to logistic mobility in which consaners remain at a more
permanent base canp for several seasons, vbile small groups procure distant
resources and bring them back to oonstmers. Generally, residKitial mcbility
is an effective strategy vrtien a variety of overlapping resource zones can be
freely exploited by small groups of hunter /gatherers, while logistic mobility
is more eclvantageous vben "a single resource determines site location as a
result of abundance or necessity," (Carlscm :118) or vdien ocnpetition
induces a group to concentrate its subsistence efforts in one locality (Vierra
:170) .
In the tipper Tonbigbee Veilley as in the lower Illinois River Vcilley, the
shift from one strategy to the other may have occurred as the result of a
resource "push" or "pull." A reduction in resource availability during the
Hypsithermal climatic episode may have pushed the pcpulaticai from the
surrounding uplands into the main river vedley. The early Holocene (Early
Archedc) subsistence/ settlement system undoubtedly enocitpassed sites in the
main vcilley, secondary streams, and uplands. The warming and drying effects
of the mid-Holocene Hypsitherrral need only to have made resources in the
uplands relatively less abundant to effect group movement through the seasonal
round. Populations may have utilized valley resources for longer periods of
time during their seasonal round.
Although the Tonbigbee V2LLley may have been a food-rich area throv:^hout
the Archaic, a change in the river regime may have increased productivity.
The shift in the river system from braided channels to a single, deep channel
with natural levees and bac)cwater lakes, created and replenished by flooding,
may have increased the abundance of fish and water fcw'l in the valley,
particularly during the Sprinc when plant and other animal resources are
.35
relatively scarce. Hie aHnv^annta of aquatic resources amenable to storage may
have created a resource "pull” providing the incentive for more permanent
habitations (BroMn and Vierra , ) .
Acx3ording to Carlson () a residential mobility pattern is oonposed of
a series of residenticd camps. Residential canpe are oocipied 1:^ "a single
band or by a microband for the purpoae of exploiting resources in the
vicinity." They will be oocupi^ for only a short period of time, perhaps
less than one season. Logistic mobility patterns (Carlson discusses three
such patterns) are ccnposed of base camps, residential, and/or extraction
camps. Base camps may be oocvpied by more than one band or a macroband. They
will be located near an abundant resource and occipied as long as that
resource is available, i.e., ooctpation is apt to be at least seasonal.
Extraction camps are limited activity, short-term camps occupied by producers
only.
Archaeologicially, residential canpe are best distinguished from base camps
by different distributions of debitage classes and by varieties of feature
types. The partitioning of activity space in base camps will be greater than
in residential camps. Evidence of more substantial housing ney be found at
base camps, especially if they are occupied during winter. Storage and
garbage pits are also characteristic of base camps. Sincse both maintenance
and subsistence taisks will be performed at Isoth kinds of canpe, a functionally
similar set of tools may be found, especially if a residential camp is
recxxnpied several times at different seasons. Variety of tool types
distinguishes residential and base camps from> limited activity extraction
camps.
Preliminauy analysis of midden mound eurti&cts and features indicate that
earlier cxjcipations (Early Archaic throu^ Eva/Morrcw Mountain period
components) at all sites were residential camps. A wide variety of tools are
present in the aussemblages from these occupations, but features are limited.
During the S^ces-White Springs and Benton periods, and perhaps later as well,
occupations were more likely base cmmps. At three sites (22It539, 221t576,
and 22lt590) , prepared area features may represent structures, or at least
specialized work areas. Hearths were present as well as a variety of pit
features. Therefore, it was at least plausible to consider the effects of
changing mobility strategies on the composition of tool assemblages.
RELATIONSHIP BETMEEN MGeiLITy STRATEGIES AND LITHIC ASS01BLA(£S
Archaeologists interested in using adaptive or evolutionary frameworks to
discuss pest behavior have become increasingly aware of the limitaticwis of
traditional stone tool typologies. Recently, an econcmdc perspective has been
applied to lithic analysis (cf. Lurie , n.d.). This perspective considers
the costs and benefits of choices made by prehistoric people in the
manufacture and use of stone tools. The variables used to discuss these costs
and benefits are independent of morphological categories. The follcwing
discussion identifies the factors that determine costs and benefits and the
behaviors that are likely respcmses to these factors.
Three interrelated sets of factors determine the costs and benefits
accrued from stone tool technology: 1) the nature of the resources eoqploited
by a population, 2) the availability of suitable raw materials and reduction
strategies for tool manufacture, and 3) the degree of group mobility. The
nature of resources e:qjloited determines the need for efficiency (the ratio of
input to output) in procurement. Technology may be beneficied, since it can
improve efficiency by reducing input in terms of energy, time and risk, or by
increasixig ou^Mt (Torrenoe ; Jcxddjn ; Joslin-Jeske , ; Earle
and Chrlstenaon ) . The suitidsility of ran xnateriAls far tools and methods
of tool manufacture are also related to the kind of items poocuied or
processed, but obtaining these rat; materials and applying various reducticm
strategies represmts major costs of technology. Hunter/gatberer mobility
inpingra on both these factors. Ethnographic acoounts docanent a range of
mobility patterns, from daily or weekly movement to fully sedentary (Lee and
DeVtre ; Biochieri ; Carlson ) , tut for hueristic purposes, the
effects of hl^ versus low mobility %n.ll be considered.
A hi^ degree of mobility affects stone technology in severed wsys. It
isposes a 'carrying cost' (Shott :3) . Although nonhunan means of
transportation (Binford ) and caching large items at sites in anticipation
of return (Gould ) can cut these costs, aspects of tool manufacture and
use are often more important. Assemblages may be coapoeed of small, easily
carried tools (Torrence ) , eiqpedient tools discarded imnediately after
use, or multipurpoee tools such as large blfaoes which can serve as
ocsfcination knives, saws, piercers and choppers, as well as sources of small
^larp flaioes (Binfcnnd ) . High mobility may restrict the time that can be
invested in tool manufacture. Needs can be filled by minimally modified
chipped stone tools rather than those requiring hours of work, such as dmped
ground stone tools (Boydston n.d.) . High mobility can have a positive effect
on the procurement of raw materials for tool manufacture. An adequate range
of materials may be obtained at low cost if access to resources is not
restricted and procurement is aiheddted within other activities performed
throu^nut a group's territory (Goodyear ; Binford ) .
Lew mobility poses other problems. Sedentary groups may experience
resource depletion within their ismediate living areas or face restricted
access to necessary goods. As populatim grows in absolute ranbers or
aggregates, a mobile life style becomes difficult. It becomes harder to move
consuners to resources. Both larger population and longer occupations at any
one location place stress on available resources. As subsistence resources
beocroe inadequate to meet population donands, new items may be inoorporated
into the diet. For example, during the late Middle Archaic in Illinois a
second line resource - small seeds - were added to the diet, althou^ their
procurement and processing required more time and effenrt than collecting and
processing a first line staple - nuts (Asoh, Ford and Asch ) . Additions
to the diet often require the introduction of new tools or the use of old
tools in new weys. New grinding tools and new cxeking vessels may have been
necessary to prepare seeds, or tools used to grind pigments may have been
adapted for grinding seed.
Another re^xxise to subsistence resource stress is the introduertion of
more efficient ways to proc:ure old food items. Efficiency is often
acxxnplished by producing more specific or more cxxnplex tools. Specificity
refers to the diversity of tools within a functional class (Torrenoe ) .
The more specific a tool is to its task, the greater the chance of success in
conpleting the task. Torrence points exit that specificity is hi^ in
situations where resource options are limited and the risk of failure is high.
It is not surprising that Eskimos have different harpocxis for different hi^ily
mobile acjuatic animals. Tool complexity refers to the number of
"configurationally distinct items" that make up a tool or f2K:ility (Oswalt
:31) . The greater the number of items, the more coaplex tlie tool.
According to Shott () cntplexity may inenrease tool specificity or
versatility. While specificity may reduce the risk of failure, versatility
may reduce the time necessary to make tools. Alternate parts c:an make a tool
suitable for several t^pes of tasks with little advance pce|iaraticn, and
bcdwn osr used elements such as stone projectile points or scrapers can be
replaced without the creation of a totally new tool. Die cost of an efficient
tool kit is in more careful, tue-oonsuning, and more scheduled tool
I manufacture.
Stress on resources need not be restricted to food items. Other
materials, such as wood for housing, fire, or tools, and lithic raw materials
can be depleted. As a location is used far longer periods of time, better
quality stone may beocme scarce, and knappers reaort to using inferior
material or altering the stone to iitprove its quali^. Thermal alteration of
I chert is one such ra^xmse. Rick (:53-^4) discusses the costs and
benefits of heating. The costs include procurement of chert amenable to
alteraticn, building a facility, collecting fuel, and risking failure. The
benefits are chert that can be Jcnapped with greater predictability and tools
with sharper edges. Alternatively, good quality raw materials can be obtained
throu^ special trips or trade, agtdn at a cost of time and energy. In either
I case, more eiqpensive material for tool manufacture should be used more
eoonaniced.ly. It will be used to make special tools, perhaps smaller tools,
to be knapped with less waste and to be xxaed more intensively (Jo8lin'>Tedce
) .
As population aggregates, access to resources often Iseocmes controlled or
restricted creating stresses similar to those msntioned above. In addition,
I larger populations are usually more socially ccaplex than smaller ones.
Greater complexity requires more information traimfer both within and between
groups (Nbbst ; W^ssner ) . Information encoded in tools can take on
the function of status maricers or group idsntifiers. Pmang hontar-gatheiers
this information is likely to be in the form of exotic materials for
^ manufacture, stylistic variatibh in tool form, and elements of tool
I decoration. In other words, greater effort is expended in tool manufacture.
Clearly a more sedentary lifestyle places dsmands on technology that require
increaWd investment of time and energy in tool manufacture and eccncmies in
tool use.
The preceding discussion makes it possible to develop esqpectations for
what lithic assemblages would look like in residential and base canps. These
I models deal first with raw material acquisition and use, then tool
manufacture, and finedly tool use.
FW MATERIAL ACQUISITION AND USE
Chert oobbles, the primary material for tool manufacture, would have been
I available in a range of sizes and quality. The costs involved in procuring
these materi2d.8 is low. Nonlocal raw materials such as Fort Payne and Bangor
cherts and Tedlahatta quartzite are also used for tool manufacture, but may or
ney not have a higher cost. If groups are highly mobile, and small quantities
of these materials occur in a ocnpcnent, we su^mct that the nonlocal material
were acquired iiKiidental to seme other activity, i.e. the procurement of that
I raw material was not systematic. In this situation the cost of nonlocal raw
material would be low. If mobility is lew, and if nonloced. raw materials are
acquired in large quantities, the costs would be higher vbether the inaterisd.
v«is acquired by special activity groups or trade.
Although chert resources are plentiful, oobbles do vary considerably in
qusLlity and size. When nrbility is high and gmp size small, the chancses of
I depleting supplies of larger and better qusdJ.ty raw material are small. Slncse
good queility materials are cheap and plentiful, we would expect that all
238
vacifltlM of tools, fam utiliaod flahM to zofiasd bifMM, tnoli bo aodo
£>■1 9Ded ippllrir ooborlal. liwi aibility is hm, amaam of tow sobKlal in
tkm iflwttMo mma of a site ssiy be *pin»ieil oiior.* fhe Isafar ooMUos end
orfMos of the htthsst quality laotsrlel wnoM be used flmb. As IsofUi of
ooQopKtian iiBtiaaes only mailer oabhles sot oM»les of leaser quality would
be available. In this situation, avail^le food quality lasterial would be
used for ante highly shaped tools, while lesser quality raw aateerials would be
used for more eapediently produoed tools. Alternatively, good quality
Twaterlal may be brouc^ in from farther awey, or heat treatment nay be tised to
iapeove the qu2dity of chert. Both of theee alternatives han/e associated
costs. Again, sini« these two sources of good material are more costly, we
would expect that they would be used more eoonomically, i.e. for tools that
require more shaping.
Ife asmaw th^ pcehistorlc knappers were well acquainted with the
materials available to them and techni^ies of thermal alteration, and that
they made oonacious choices about which materials were suitdale for tool
manufacture and use. Biqiectations for the acquisition and use of raw
materials at residential canp sites (hic^ mobility) and base canp sites (low
mobility) can be siamarized as follows:
In residential caape associated with hic^ mobility
a. local raw material used for tools %fill be of good quality, sijnoe the
supply is plentiful. Poor quality raw msterial can be rejected %idLthout
penalty. Ihe chances of esdiausting the aucply of good raw material is
minimal, since the length of oocnqpetion ie shi^.
b. Good qu^ty local raw material will be used for all types of tools and
there is little restriction on the size of tools that can be made.
c. Since good quality material is plentiful, heat treatmi^ may not be
necessary.
d. small quMitities of a wide variety of nonlocal cherts may be found in the
assenblage, probably in the form of finijAed tools.
In base canps associated with low mobility
a. Local vmr material exhaustion beocwes more likely. Iherefore, we would
th because there is
little need for specificity and because they are more portable.
In base camps associated with low mobility
a. More energy will be invested in tool manufacture. We would expect to find
more extoisively shaped tools, both chipped and ground stone, and less
utilized or simply retouched pieces in the assemblages.
b. We would expect that tools become more specialized, since tools designed
for one task are more efficient at that ta^, and portability is less
important. Ihere will be fewer multipurpose tools, unless multipurpose
tools themselves fill a particular technological slot, i.e. tools to taJce
on logistical forcys, or tools made out of e]q>ensive raw materials that
have to be used eoonomiccdly.
c. Tools will be more complex. Tools will Ise more stand£u?dlzed, and there
will be more hafted tools.
TOOL USE STRATEGIES
Tool use strategies can also reflect an increasingly sedentary population.
If tools need not be efficient, and if the time and energy invested in
manufacture is low, then tools will be used expediently. Tools used for a task
may be abandoned \dien the task is completed, since others can be made quickly
if the need arises. If tools break, they will most likely be abandoned.
Casually made and used tools need not Ise part of the hunter-gatherer's baggage
as he or she moves on to the next camp. Because camps are used for a
relatively short period of time, we may recover more whole tools that still
seem useful. Ihere will be fewer chances for abandoned tools to be trampled
or accidentally Idcloed into a fire.
Conversely, if resources are scarce and require more efficient tools that
take more time and enexgy to manufacture, then these tools will be more
heavily vised, resharpened, and curated. They will be reworked, or perhaps
simply used for the same or another purpose vhen broken. As length of
occupation increases and opportunities for trampling and cleanup activities
increase, the nvmber of brolcen tools should increase. Ihese broken pieces may
240
also be aroaller than than those found in a short-tenn occupation.
Bqpectations for tool use at residential canp sites and base canp sites can
sumarized as follows:
In residential canps associated with hi^ mobility
a. Tools will show ccnparatively little use.
b. Tools will not be resharpened or reworked vAmn broken as often in
residential canp occupations as in base canp oocipations.
c. Abandoned tools will be l^s broken and heat altered less often in
residential caap oocuqpatidis than in base canp occupations.
In base catops associated with low mobility
a. Tools will be used more intensively at in base canp occupations than at
residenti2d canps.
b. Tools will be resharpened and reworked mcnre often in base canps than in
residential canps.
c. Tools will be broken more oft^, the Inxdoen pieces will be awaller and
more often heat altered in base canps than in residential canps.
SBCmON 2: SAMETf
lABORATDBY PBOCEDUFES
'foe first decision nacle for the E^iase III lithic study was identification
and selection of sanple units vddch could be used to address researcii
questions. Ehphasis was placed on units with the hi^iest integrity ^mmiiig
the entire Archaic periods. Units selected had good stratigraphic dsfinition,
included chronologically diagnostic materials, and/or were radiocarbon dated.
Units were then grouped into six time categories based on these diagnostic
materials and radiocarbon dates, 'foe time categories were:
1 Early Archaic (Dalton/Greenbriar/Kirk points)
2 Middle Archaic 1 (Eva/Morrcw Mountain points)
3 Middle Archaic 2 (Eva/Morrow Mountain and S;^s/White Springs points)
4 Middle Archaic 3 (Sykes-Mhite Springs and Benton points)
5 Middle Archaic 4 (Benton points)
6 Late Archaic (Little Bear Creek points)
Sample units were selected from six sites (22It539, 22It576, 22It590,
22It621, 22It623, and 22It624) (Table 94). They vary in size, content, and
time periods represented, 'foese variations are due to specific site histories
and functions as well ais preservation and excavation strategies. A total of
6,391 chipped emd ground stone tools were included in the lithic stuc^. 'foe
Walnut site (22It539) contained material from five of the six time periods.
The only period not represented was the Late Archaic. This site provided the
best sanple of S^ces/White Springs and Benton period occupations and
contributed the majority of chipped stone artifacts in the lithic study (total
number from midden and features - 2,375) . While the samples were smaller from
the Poplar (22It576: ii=l,409). Ilex (22It590: rF970) , and Hidoory
(22It621: ns648) sites, they were particularly important because of the
v^ll-documented Early Archaic and early Middle Archaic assemblages. The Beech
and Oak sites (22It623 and 22It624) provided the only Late Archaic assemblage,
although the sample was small (n=332 and n=558, respectively) .
241
nu 94
Vkmm ni miililHi aad fieatoe auits. _ _
Periiod Utait Description
22It539
(1) Early Archaic
(2) Middle Archaic I
(4) Middle Archaic 3
(5) Middle Archaic 4
Block 0 Levels 18-21
Block A Levels 15-17
Block D Levels 16-17
Block D Feature 128
Block D Feature 131
Block D Feature 132
Block A Levels 11-14
Block A Levels 5-10
Block B Levels 6-10
Block B Feature 6*
Block C Feature 99*
Block C Feature 120
Block D Feature 142
6x8 m block
4x4 m block
6x8 m block
Prepeuned area - 2. 8x2. lx. 73
Prepared area - 1.48x1.03. .23
Prepared area - 1.31xl.49x.02
4x4 m block
4x4 m block
6x8 m block (20% sanple taken)
Pr^>ared area - 1.78x3.34x.35
Cache - 1.3x.l6x.06
Prepared co^ - 5. 08x6. 39c. 34
Pit - 1.33xl.38x.95
22It576
(1) Ecurly Archciic
(2) Middle Arch2d.c 1
(3) Middle Archaic 2
(4) Middle Archaic 3
(5) Middle Archaic 4
Block D Levels 14-25
Block D Feature 115*
Block D Feature 116
Block D Feature 118
Block D Feature 119*
Block D Feature 120*
Block D Levels 12-13
Block C Feature 90*
Block D Levels 9-11
Block A Feature 71
Block B Feature 10*
Block A Feature 49
8x12 m block
Pit - .68x.66x.39
Lithic conoentraticn
1.10xl.06x.07
Lithic concentration
1.05xl.l4x.ll
Lithic concentration
.93xl.05x.06
Lithic oonoentration
.20x.61x.01
8x12 m block
Pit - 1.81xl.94x.81
8x12 m block
Pit - 2.25x2.00x.27
Cache - .26x.37x.05
Prepauped area 2.65x2.65x.23
22It590
(1) Early Archaic
(2) Middle Archaic 1
(4) Middle Archaic 3
(5) Middle Archadc 4
Block W Levels 9-12*
Block Y Levels 9-12
Block E Levels 13-15
Block F Levels 10-14
Block J Levels 7-13*
TP 13 Levels 7-13
Block W Level 8*
Feature 6
Block B Levels 6-10
Block J Feature 8*
Block A Feature 34
Block B Feature 54*
Block D Feature 73*
4x4 m block
4x4 m block
4x4 m block
4x4 in block
2x2 m block
1x2 m block
4x4 m block
Pit - .96xl.30x.88
Pit - 1.35x.39xl.22
Pit - .43x.70x.40
Lithic cluster - .81xl.25x.ll
Pit - 1.20xl.l8x.42
2
22It621
(1) Early Arcdvaic Block A Levels 10~16* 4x4 m block
Block C Levels 10.2-23 4x4 m block
Feature 4 Pit .78 cm in diameter,
.15 cm deep
(2) Middle Arcbaic 1 Block A Levels 6-9* 4x4 m block
Block C Levels 6-10.1* 4x4 m block
Block E Levels 8-10* 4x4 m block
Feature 1 Lithic cluster -
.36 in diameter
22It623
(5) Middle Archedc 4 Blcxk C Levels 6-9 4x4 m block
Feature 14* Ccnpcund pit - 1.55xl.95x.84
(6) Late Archaic Block D levels 6-7 4x4 m block
Feature 11* Shallow pit - 1.17x1.35x7
Feature 12* Six ground stcxie tcx}ls
Feature 20* Cache
22lt624
(5) Middle Archaic 4 Blcxk A Levels 8-9* 4x4 m block
Feature 7* Ccnpcaind pit - 3.00x3.85x.75
(6) Late Archaic Blcxk A Levels 6-7* 4x4 m blo^
Blcxk B levels 6-8 4x4 m blcxk
Blcxk C levels 6-8 4x4 m block
_ Feature 15* _ Pit - 1.20xl.23x.75 _
* units rq>reaented by tools only,
A great deal of data were acxuiulated on lithic dd9itage during Phase I
and II. Data include information cn raw material and size grade (1 inch or
2.54 cm, 1/2 inch or 1.3 cm, and 1/4 inch or .64 cm), and cxxmts, and weights
within size grade (raw material grovf>ing8) . An evaluation of this previously
collected data frcm the selected units was undertaken, and a snail nonber of
units frcm each site was selected for additional stu^ in reqonse to research
cjuestions. These were selected based c» the presence of large nunbers of
flakes and on the presence of sorted fine-scareen nwnilc'n frcm at least a
portion of the unit. One hundred and fourteen dribitage isiits frcm several
blocks at each site and from all of the time categories %«ere selected. Units
were from the 2x2 m (6. 6x6. 6 ft) midden units cn: features. The fine-scareen
samples were usually ta]cen from the cx>ntrol blcxks (Appendix III: Table 1) .
SAMPLE RETRIEVAL
Uhfortunately there were unforeseen difficulties ]x>th in xecx>vering actual
specimens from storage and in recxjvering cxnputerized data. Ihese
difficulties, to sene extent, limited the range of analysis that cxuld be
performed. Although Kuoy of the problems enoountered were remedied during
Phase III, the time required to do so affected the time left fexr analysis.
The followijig discussion of these problen^ is designed to give an accurate
account of the san(>le and to make reocranendations for the processing and
curaticn of materials from conplex sites.
The most serious retrieval problem vets the recovery of tools stored by
artifact type and arbitrary type and arbitrary ID maober, rather than by
provenienoe. Since sanple units were chosen by provenience, and there is no
absolute relationship between ID nunber and provenienoe, the actual recovery
of the designated artifacts from the hundreds of boxes and the thousands of
bags and coin envelopes in vhich they were stored was a far greater problem
than anticipated. Several smaller problems oocplicated the retrieval. The
inclusion of utilized flakes with ddsitage necessitated searches throu^
nonerous debitage boxes and bags for specific ID numbers. Many artifacts
selected for lithic analysis from 22lt576 and 22It590 had been placed in a
selected type collection to aid laboratory assistants in classifying artifacts
during Ehase I and II. These too were stored by artifact type rather than
provenienoe, and ^lecific ID nuobers were difficult to find.
I The metted of sorting and storing ddsitage adso posed difficulties.
Lithic materials not considered tools had been sorted into flake size
categories, (1 inch or 2.54 cm, 1/2 inch or 1.3 cm, and 1/4 inch or .64 cm) ,
and several kinds of introduced rock categories, presmably unmodified lithic
materials. Each flake size category was also divided by raw naterial lypes,
and seme raw material types were subdivided liy the presence of heat treatment.
I Many levels of packaging had to be sorted and opened to retrieve the sanple
specimens.
Introduced rock was subdivided into 25 subgroi:^}e based primarily on raw
material. One of these subgroup, non-utilized, fire-cracked chert/chunks, in
fact contained any blocky or angular piece, heated or not, that was not called
a core fra^nent or utilized chunk. Many of these pieces are the products of
I early stages of cobble reduction and are essential to the analysis of
d^itage. Even heat-altered pieces may shew signs of intentional
modification. None of these other 25 subdivisions were reexamined during
Phase III.
Retrieval of information from oemputer files was also ccnplicated by
previously undetected errors in coding. During the course of the project,
I several, iniividuals were in charge of setting up end retrieving oenputer data.
None of these people were archaeologists actuedly engaged in collecting or
analyzing data, and none of the archaeologists collecting data were thorou^ily
familiar with the data management system (SAS) . Lade of continuity and
conmunication between archaeologists and oenputer ^lecialists made data
retrieval nuch harder than necessary. In some cases, levels were incorrectly
I stored in the conputer data. Therefore, when requests for identification
nurbers associated with selected sample units were requested, incorrect
listings were generated. Although these problems were detected and corrected,
much time was lost. The recovery rate of specimens averaged 90.8% with a
range of 0-100%. Recovery rates for tools and ddsitage by unit by site are
presented in Tables 2, 3 and 4 of Appendix III.
I It should be noted that the airtifacts or debitage not retrieved are still
in midden mound storetge boxes, but neither the time nor the personnel was
available to search fra: them. The problem of recovery is not unique to this
project; but similar difficulties plague large, ccmplex archaeological
projects (Lurie ) . In li^t of the experience gained in the midden mound
project, the following recatmendations are made for storing items from
j multi-oonponent sites containing thousands of artifacts and hundreds of
thousands of pieces of debitage:
244
1. store material by provenience. Major clasaea of materials, lithic,
ceramics, floral, aid faunal can be sttxmd separately, but each should be
stored fay maanijigful provenience units, sot by axbiticary ID nusbers or
artifact typae.
2. Lithlc artifacts and debitage can be stored separately, since these are
usually analyzed using different variables usi^ different laboratory
techniques as long as both are stored by provenience.
3. When pluming the curation of specimens, ocnsideraticn should be given to
ease of recovery and reexaninaticn. This is eepecially inportant \Aien
several stages of analysis are anticipated. Althou^ coin envelopes are
easy to write on and nultiple stales ensure that specimens will not
escape the storage facility, inspecting and retrieving their contents is
very time consuming.
4. Computer work should be done if at all possible by airchaeologists doing
the analysis. Those intimately involved with recording and interpreting
data are best able to pick out errors and inconsistencies in the data set.
A great deal of time can be saved over the project as a whole \dien
requests for ocnputer-generated data do not have to go throu^
intermediaries. Time should be budgeted at the beginning of any large
project for familiarizing staff with the ocnputer data management system
in use and the general structure of the data sets.
LABORATORY PRXSDURES
As the retrieval of the selected artifacts &cm storage proceeded, items
were r^>ackaged in 3x5-inch plastic transparent envelopes. Ihe engined, coin
envelopes oQnt2dning provenience informaticn served as a for the
artifact in its plastic bag. Artifacts were organized ^ provwMi^ieoB, and
those examined during Phase III were separated from the main oollaction. This
system of storage aids in cross-checking data sheets, idmatifying refitting
fragpoents, and pulling t^>ecimens for special treatment, e.g. photographs.
Artifacts from each site were examined as a unit. All items were laid out
by block and level on laboratory tables and examined for pieces that oould be
conjoined. Success varied according to site and tine period. Tb increase
consistency in recordiag variables, laboratory assistants ware asaigned
particular categories of artifacts to evaluate. One reoarded variables for
all ground stone and for all snail, broken tool firagnents. The otiier recorded
all utilized or poteatially utilized flakes. Variables for 2lL1 cores,
preform, whole tools, including all hafted bifaoes, were recorded by the
lithic specialist. Frequent conferences were held among the laboratory staff
to ensure that all personnel were aware of conventions used to record
variables and to discuss any problems that mi^t arise. Spot checks were made
by the lithic specialist on coding sheets turned in by laboratory assistants.
All computer data entries were checked for internal consistency and recording
errors.
SECneW 3£ EXPERIMEl/EAL PROGRAM
The lithic experimental program for the midden mound project addressed two
primary aspects of litl^c maniifacture: heat treatnoit and tool reduction
strategies. It also addressed tool function to a limited extent.
Epqperimental work conducted in the Fall of provided several kinds of
baseline information for those individuals who were to examine the sample of
archaeological lithic materied. Heat-treatment experiments were performed to
245
pcoduoe exnples of variation in color, luster, and warJcability for severed
kinds of chert heated to different tenperatures. The exendnaticn of both
unheated and heated materials not only identified the variables that vere used
to score heat treatment on the archaeological specimens, but it also
identified those situations in vdiich the asaesaonent of heat treatment might
prove difficult. The r^lication of several reduction sequences in the
manufacture of Archaic tools (performed by flintknz^iper Jeffrey Kalin)
scqpplied exenples of tools at different stages of manufacture as viell as
samples of ddsitage for statistical analysis. Tliis analysis of systematically
collected d^itage provided the framework for interpreting the hundreds of
thousands of pieces of debris from the six sites. A set of use-^wear
e9q>eriments generated a ocmparative ooliection of tools with patterned edge
modification from known and measured activities. These helped distinguish
used and unused archaeologicad items, even when the motion with v^ch the tool
was used or the material on v^ch it was \ised can only be assessed very
generally. The detedls of the e9q)erimental program are described belcw.
HEiBT TREATMEWT
Since the work of Crabtree and Butler in , controlled heat-treatment
of chert hzis been recognized as an important technique in the manufacturing
repertoire of prehistoric peoples. Presumably, chert was heated to improve
knapping quality, although it may have been heated for aesthetic reasons as
well. In general, heat treatment results in material with a more hcmogenous,
vitreous matrix vMch flakes with less force and in a more predictable manner.
Rick () and Anderson () suggest that heat treatment mEy have been used
for technologicEd reasons. Heat-treated cherts can be used to make larger,
thinner tools and to produce tools with sharper edges. But there are costs
involved in gaining these desirable qualities - costs associated with the
process of heating, such as gathering fire wood and constructing a heat
treatment facility, costs due to loss of stone during the heating prooess, and
costs in tool use life, because heat-treated tools with sharp edges are ixTt as
long lasting as tools made from unheated materials.
Field and laboratory experiments have been performed to identify the
critical tenperatures and lengths of time needed to cause alteration in
various lithic raw materials, and to define attributes that can be used to
identify heat treatment in archaeological specimens (Purdy ,;
Mandeville ; Mandeville and Flenniken ; Collins and Fenwick ;
Gregg and Grayfcush ; Melcher and Ziimerman ; Weymouth and Mandeville
; Rick ) . The visual criteria usually used to indicate heat treatment
include changes in color, changes in luster, the presence of pronounced
condioidal ri^^ling, and heat fracture scars of various sorts.
Color cdiange is related to mineral impurities in the chert and may be very
dramatic, but improvement in raw material quality can occur without color
change, and color change can occur at tenperatures belcw that needed to alter
chert quality (Purdy ) . In any case, it is necessary to have samples of
unheated material with vhich to cxxvpare heated ones.
A better indicator of heating is differentied luster on the exterior eind
interior of heated pieces. "On an artifact with flaJced surfaces produced both
before and after heating, a contrast will appear in the luster of the turo
surface types. Presence of such a Ixister oontrzist is near-certain evidence of
heat treatment" (Rick :57) . If a heated piece has been subsequently
flaked to the point that all of the preheated surface has beei removed, the
degree of luster may be used to indicate heat treatment, althou^ with less
246
certainty. Again, it is neoessaiy to have xnxnexxius sanples of unheated pieces
of chert for aoB|Mu:iscn.
Idmtlficatijm of heat tieatnent formed an iiqportant part of the initied
processing of chert tools and debitage in Phase I and II. At that time no
attenpt was "«<*» to distinguish intenticnad from unintentionad heating.
Ejqperinents on heat treatment were undertaken in the Fhame II lithic stuc^ to
provide fcmad documentation of the changes that occur during the heat-
treatment process, to identify those situaticns in v^ch an evaluation of heat
treatment is particularly difficult, and to distinguish betweoi intentionad
and unintentional heating.
Preliminary heating experiments with Camden chert indicated that this
chert esdiibits change in color at 500** F (260.0° C) and that changes in luster
occurred by 800° F (426.7° C) . A more ocnprehensive set of heating episodes
was then ocaiducted. Four types of chert ware used in these eagperiments;
Caraden (ei^t cobbles) , Pickwick (five cobbles) , blue-gray Fort Payne (20
flakes and chunks) , and fossillferous Fort Payne (eight sanples) . The samples
chosen for the heat-treatment e:q)eriments were selected to represent the
maxinun variation for each type of raw materiad within the sample of raw
materiads collected in the stucty area during Phase I and II fieldwork. While
only the Attr\ee>r vcuriety of blue-gray Fort Payne was avadlable, a wide variety
of Camden, Pickwick and fossllifexous Fort Payne was available for the
eaqjeriments. Several unheated chunks and flakes fron each sample were
retained to seirve ais baseline samples. A sample of each i^)e of chert was
heated in a kiln at 100° Fahrenheit intervads, beginning with 300° F
(148.9° C) and ending at 1,000° F (537.8° C) . Tknperature was raised slowly,
approximately 100° Fahrenheit per hour, until the desired tasperature wais
reatched. This temperature was maintain^ for one hour, and than the heat was
turned off and the unopened kiln allowed to cool, usually overnight.
After each heating episode, specimens ware eoeanined for changes in color,
luster, and workability. Color was reoordcxl using tihe ocnplete MUnsell Color
chart. Luster was recorded subjectively as dull, satin, semi-glossy, and
glossy. Color and luster for both the unheated and heated ^mcimnis ware
recorded by a laboratory assistant. Changes in knappability ware subjectively
scored by the flintknapper. In general, unheated Camden and Pickwick chert
were very hard to work; a great deal of energy was required to remove flakes
fron cobbles. After heating, the material was much easier to flake. The
detailed results of these experiments are presented in Tables 5-7 in
Appendix III. In addition to information oollected during these controlled
heat-treatment experiments, many informal observations were made on pieces of
chert heated for use in replication esq^eriments.
Wiell-defined changes in the color of Camden chert occurred at 500° F
(260.0° C) . At this temperature the color change, primarily fron creams and
yellows to pinks eind reds, affected the outside of the heated specimen but
often did not penetrate all the way into the interior. Archaeological
specimens heated to this temperature may be difficult to identify as heat
treated, d^mnding on vdiether the inside, outside, or all parts of the cobble
(preform or flake) is represented. As temperature is increased, the degree of
color change became more pronounced, and penetration was conplete. Conplete
color change usually occurred at 700° F (371.1° C) . It should be noted,
however, that cocpletarvess of color change was dependent on the thickness of
the specimen as well as temperature. In no case weis the pink or red cx^lor
resulting fron heat similar to unheated Camden with a naturally slight pinkish
cast. At 800° F (426.7° C) definite luster change oocnirred in all but the
most chalky specimen. Camden chert is fairly heat tolerant, but at 1,000° F
247
sane ^aecjmens became brittle and knapping quality declined. Below this
taqperature ^dien heat fractures occurred, they viere most oft^ in thick flakes
or bifaoes that heated unevenly. These fractures also seemed to develop from
small moisture-containing crystal pockets within the stone.
Pickwick chert changed in color and luster at the same tenperatures and
was equally heat tolerant. Pickwick ohert often has concentric bands of dark
gray, red, and yellow color. All of these dianged in color and luster with
heat. The gray and red colors became darker, vAiile the yellow turned to red.
The red color in unheated Pickwick was of a different hue than the red color
in heated Camden chert. Problems in distinguishing heated Pickwick from
heated Camden or other heated Tuscaloosa gravels arose \dien only one of the
three Pickwick colors was present on an item. The heated red Pickwick is
similar in color to heated Camden chert and may be similar to heated Yellow
Chert. In addition, a small pieces of dark gray Pickwick material may be
confused with the darker varieties of blue-gray Fort Payne chert.
Fossiliferous Fort Payne also responded well to heat. The gray material
developed a strong pink to red cast at 400-500° F (204.4-260.0° C) , and luster
develcped between 700° and 800° F. It is also a heat-tolerant chert. The
e^qperimental results are similar to those reported by Morrow for the Yellow
Creek Archeological Project ().
In ooitreist, blxie-gray Fort Payne chert from the Wilson Dam Tennessee
Valley cu:ea was not heat tolerant. It became brittle and useless for knapping
after heating to 500° F (260.0° C) . Material heated to this tenperature, or
lower, often exhibited expansion fractures or breaks with very granular
texture. At 800° F (426.7° C) , all pieces shattered in the kiln. Resulting
fragments were heavily crazed and covered with pot-lid fractures. It is
unlikely that this material was intentionally heated prehistorically.
REPLICATIC^ EXPERIMENTS
Preliminary classification of lithic artifacts in Phase I and II indicated
that uses of raw materials and reducticm strategies were not constant through
time. To understand the lithic reduction strategies used prehistorically,
replication sequenoes were outlined by S. A. Abler, J. Kalin, and R. Lurie.
Replication sequences were performed with the raw materials collected mainly
from the study euea (Camden and Pickwick cobbles collected near Fulton, Ns and
blue-gray Fort Peyne chert slabs collected from the Wilson Dam area) . The
products of the various reduction sequence stages incltided usable flakes,
bifaces at different stages of ccnpleteness, and the d^itage that results
from each manufacturing step.
Five major manufacturing trajectories were outlined;
1 reducticHi of medium cobbles of Camden and Pickwick chert for the
production of broad blade bifaoes (Figure 38) .
2 reduction of small cobbles of Camden and Pickwick chert for the production
of narrow blade bifaces (Figiure 39) .
3 reduction of Fort Payne chert blanks to produce broad blade bifaces
(Figure 40) .
4 reduction of small and medium-sized ccAbles for the production of usable
flakes (Figure 40) .
5 bipolar reduction of pebbles to produce flake and fleike-like pieces that
could be used as tools (Figure 40) .
248
Model of medium-sized cobble reduction sequence
FOAT PAYNE REOUCTUN
REDUCTION FOR FLAKES
FREE HARD HAMMER
Psrcuttlon
Rpductlon
not performed
REDUCTION FOR FLAKES
BIPOLAR REDUCTION
Figure 40
Model of Fort Payne reduction sequence, flake reduction sequence,
and bipolar reduction sequence.
t
Greenbricu:
2
Kirk
2
r
E/Md/CC
7
1
S-WS
1
1
Benton
1
lac
8
lltese sequencses were established on the basis of the knapping and anedyticcd
experience of the lithic specialists » on archaeological literature, and fron a
prelisninary examination of the recovered artifacts. Hie sequences were,
therefore, specifically designed to contain the variability present in the
archaeological assemblages to be studied.
The goed of the replication program %«as to produce at least ten examples
of all five reduction sequences vhich would would include 10 examples of each
intermediate stage of biface reduction, 10 examples of each hafted biface
roocphologiced type praninent in the Archaic assemblages, and at least 10
examples of debitage associated with each stage and lype of reduction. Hiis
goal was not met in all cases because of limitations in time and availability
of raw material. Hiere were far more examples produced of early stages of
biface manufacture than examples of finished items, because later stages of
reduction were usucilly more time consianing (Tables 95 and 96) . Hie production
of Benton bifaces was also limited by the availability of suitable Fort Payne
chert. Biface reduction consisted of four stages (vbcle cobble reduction,
initial edging and thinning, secondary or advanced thinning, and biface
trimming and finishing) so that intermediate products could be weighed and
measured and debitage collected.
Table 95
NOriber of H
in redbcticn
Gobbles reduced for flakes
Free hard-hammer
Bipolar
Cobbles reduced for bifaces
Fort Payne blanks reduced for bifaces
Flake blanks and nuclei produced from medium cobbles
Small bifaces paxiduced from small cobbles
Bifaces produced by edging and initial thinning
Bifaces produced by secondary thinning
Finished narrow bifaces
Finished broad blside bifaces
small medium
CXI=Eva/Morrcw Mduntadn/C^ress Creek S-WS=Sykes-White Springs
LBC=Little Bear Creek
* Whole unfinished pieces have been manufactured through the secondary
thinning stage.
** The size of the finished produced in most cases is dependent on the size
of the flake blank or nucleus used for manufacture.
BIT = Iteme brokai during edging and initial thinning.
*** ST - Item® broken during secondeuy thinning.
252
The follcwing procedures for collecting and recording information were
used on eacdi knapping ^isode performed in the experimental program. All
cobbles and Fox± Payne flake blanJcs were assigned numbers, measured, and
weighed before reduction. Ccbbles used for bipolar reduction and free-hand
h£u:d haniner reduction for flake production were not heated. All Camden and
Pickwick materials used to produce bifaces were heated in an early pcurt of the
reduction sequence. All knapping episodes were performed on large plastic
sheets to facilitate the collection of debitage. Ihe time taken for each
stage of reduction, the type of percussor used, and comments by the knapper
were recorded for all st^s in biface reduction. Each stage of reduction was
photographed. A mirror was placed behind the tool is that plan view and
cross-section would ^jpear the same. The reduction sequences are described in
order of conplexity.
1 . Bipolar Reduction - Because no raw materials of the correct size had been
collected from the project area, 12 small Jasper cobbles fortuitously
collected along the embankment of the Mississippi River in New Orleans, La
were xised and reduced \ising a bipolar technique. In seme cases, the
cobbles were wrapped in a piece of leather to facilitate holding when
placed on a sandstone anvil. The ccdijle was then struck with a
hanmerstone repeatedly until it Wcis reduced to flakes and flake-like
pieces.
2. Reduction of cobbles for flakes - Gobbles of various sizes were reduced by
free hard hanmer percussion to produce flates suitable for making a
variety of snail tools. Four replicators were involved in this part of
the experimental program, and the flake size, shape, and thickness varied
with the abilities of the knapper s. Each knapper was asked to select
those flakes that they thought would be useful for tools. These were
collected separately from the other debris.
3. small Cobble Reduction - Using a medium-sized hamroerstone, several small
cobbles were reduced into roughly shaped small bifaces and flake blanks
\diich were heated and used to produce narrow bladed bifaces similar to
Little Bear Creek points found in the Late Archaic period assemblages.
In most cases, the production of these bifaces was accomplished in two or
three stages rather than in the four stages used for the production of
broad blades. Ccbble reduction and initial edging and thinning were
cembined into one step to produce the small rough bifaces. Flake blanks
were reduced through initial edging and thinning. No secondaury thinning
was performed in the manufacture of ttese tools. Bifaces were finished
losing a combination of percussion, with a small hammerstone, and pressure
flaking, with antler tine pressure f laker. Serrations were added on seme
of the tools using the narrow conctyle of a deer mandible.
According to the replicator, the manufacture of these points involves
little more than striking off excess material from a suitable flake or bifaoe
nucleus using a non-marginal flaking technique until a digital biface, almost
as thick as it vas wide, was obtained (Kalin personal conmunication ) .
The cortical platfom of the flake blank often served as the base of the tool.
Occasional traces of cortex or flake blank surfaces were sometimes present.
Short, curved, thick-plat formed flakes with simple dorsal scarring made up the
majority of the percussion flakes produced. Some more advanced flakes with
smaller platforms and complex dorsal surfaces could also be produced. The
253
inteirsection of c^iposing dorsal scars on advanced flakes seldcxn exceeded
1-1.5 an (.4-. 6 in) from the flake's platform. Pressure flake finishing
could be used to smooth the previous surface irperfections. When thin flake
blanks v^ere the starting point for these narrcw-bladed points, the percussion
stage could be skipped all together and all further work dcme solely
pressure. Ihe stem may be finished tising either percussion or pressxire.
Ihese tools were ea^ to produce. Hiey used a minimal amount of raw
material and took little time to manufeKrture (7-12 minutes) . Because of their
width/thickness ratio (c^roximately 2.5/1), they are sturdy tools that seldom
break during manufacture or use.
4. Medium Cobble Reduction - Camden and Pickwick cobbles were reduced to
produce all but one of the major types of broad-bladed points found in the
Archaic assemblages of the midden mound sites: Greenbriar, Kirk,
Eva/Morrcw Mountciin, and S^es-White Springs. The manufacture of these
point types usually involved the following steps.
a) . . The raw cobbles were reduced by free hand hanmer percussicxi into
nuclei (crude bifaces) , cores, flake blanks of a size suitable for
bifacial reduction, other usable flakes, "practice pieces" (blocky
pieces without much potential for bifacial reduction) , and non-usable
debitage.
b) . Nuclei and flake blanks were then selected for biface production and
were heated. In seme cases edging and initial thinning were
performed before heating. After heat treatment edging and initial
thinning, secondary thinning, and finishing were performed. Flaiking
styles differed in order to replicate the particular type of point
being made.
The production of Eva/Morrew Mountain points involved greater flaking
oarplexity than the production of narrowH^laded Little Bear Creek type points
described above. The plano-convex or biconvex bifacial Eva/Morrew Mountain
points were replicated using a snail hanroerstone. They were manufactured with
minimal sheauring platform preparation, and like narrcw-bladed stem points,
they were manufactured with a minimum amount of flaking. The technique of
shearing platform preparation without platform grinding was chosen because it
produced flake scar attributes most like those found on the archaeological
specimens. This technique provided the proper angle for marginal thinning.
The lack of platform grinding was responsible for hinge fractures, and/or st^
fractures similar to those seen on archaeological pieces. Debitage often
exhibited crushed platforms.
Unlike the Little Beeu: Creek points, pressure flaking was kept to a
minimum and used prinarily for margin leveling. Eva/Morrew Mountains were
redimed as triangular bifaces, and the flake debitage produced included basal
flakes which intersect lateral thinning flakes and leave right angle flake
scars on seme of the flake dorsal surfaces. When a minimal amount of flaking
was used to produce Eva/Morrew Mountadns, the resulting debitage did not
exceed an intermediate corplexity in flake dorsal sceurring. Cortex and flake
blank surfaces also occurred on the finished tools and debitage. These
attributes are often present in many of the archaeological samples.
Once the triangular preform was produced, the tool was finished hy basal
or comer notching. The exact locaticxi of this notching has often lead to the
classification of these tools as different point types. The difference
between Eva/Morrew Mountain, and Cypress Creek points is little more than the
location of one or two notching flakes. Despite the specific location of the
254
notches, the basic manufacturing techniques for these three point types main
the sans.
Kirk point peodac±ion inoacporated an ailUHffiwal saries of flakes bsyond
tlwae mowed in the production of Eva/Morron llBsntain points, nils
additioBal flaking gave these points a flatter and ■oothei' cross section.
Mild platfocn grinding, in oonjuncticn %d.th shearing, and antler billet
percussion mere used during the advanced thinning stage. To produce a
finished Kiric point similar to those found in the archaeological saaple, it
mas necessary to maintain maxinun width at the distal as well as prcximal end
of tile preform. Bectangular preforms of this type were recovered in the Early
Archaic ociiyonents of the midtei mound sites. TliBre is a distinct difference
betwwn this preform shape and the triangular Eva/Morrow Mountain preform.
Thinning takes place at both the proximal and distal ends and prodtoad more
dshitage with rig^ angle intersecting dorsal soars. Additional thinning
caused an increase in dorsal ocscilsacity in the dkbitage produced. More
pressure flaking vns used for the finikdng of Kirks than for Bva/Norrow
Mountains. A deer mandible vms \iaed for serration and comer notching.
Distinct cone tiiaped notching flakes were often produced by this technique.
The percussion thinning of Greehbriar points was similar to that of Klidcs
and is presented in Figures 41^5. The biccnvexity and surface attributes of
Greenbriars were most easily attained through seri^ pressure flaking.
Several aeries of pressure flakes were rencmd in their production. To
replicate the archaeological specimens, a slic^itly wider^-t^pped antler
pressure flaker was used for Greenbriar finiadtiing than vms used for Kirks.
Percussion or pressure basal thinning (fluting) was used to finish the tool.
S^s-4diite Springs and Benton points viere mwMfactured in very similar
wiys. Mhile the Sykes-Mhite Springs points were produosd from oohble flake
blanks, the Benton points were replicated Iran flake blanks deriwsd from
blocks of blue-gray Fort Payne. These tools are perhaps the most ccrpleoc in
terms of flaking technology, and the reduction stage products are presented in
Figures 38-40. The advaxKsd level of thinning used in their proAa^on was
unique because of the critical level of isolated platforma preparation
naoBssary to produce a biface with a plano/plano cross section. Platform
grinding was used to prevent platform collapee vdiich would result in hinged
and stepped flake failures.
To attain a plano/plano biface it was necessary that the thinning flakes
reach more than half way acroes the biface. This enabled the redaction of
blfaoe thickness faster than biface width. This technique produced a high
percentage of flat flakes with apposing dorsal scars. While tianwBiuU:»wiJ may
be used to produce this tool type, an antler billet was preferred by the
replicator. Overshot flakes are a more ocswon type of flake failure in the
production of plano-plano points. It appears that the more advawed the
knapping that takes place the more unlfoonity that exists between the dtisitage
prodboed by the manufacture of individual piims. A distinctly different type
of pressure flaking characterizes these points. While the other point types
were finitiied tising an inward pressure flaking tecbnigue, these plano-plano
points were finished using a dowiMard pressure flaking technique. With inward
pressure flaking the flake removed skims over the surface of the biface.
Pressure is applied with the tool hand-held. Inward pressure flaking involves
the use of marginea platfocms. In contrast, downward pressure flaking is most
easily acoonplished by placing the tool on an eoivil and short, abrupt flakes
are snapped or popped off by non-marginal downward pressure. This prooess
produces the distinct beveling typiced. of Benton points.
Figure 43 Second thinning
of flake blan]
Figure 44 Trimming of flake blank for
r
5. Fort Payne Flake Blank Reduction - Ten Fort Payne flake blanks were
reduced for the production of Benton points described above. Large flake
blanks were removed from large blocks of Wilson Dam Fort Payne chert.
This step was not carried out in a controlled manner - no debitage was
collected because it is documented that this manufacturing procedure did
not occur at the sites investigated and because of the highly flawed
nature of the collected materials. dark blue-gray Fort Payne can not
be successfully heat treated and is a much harder stone to knap. The
flake blanks were reduced through the edging and initial thinning,
secondary thinning, and finishing stages similar to the production of
Sykes-White Springs points.
Several kinds of information were recorded for these reduction sequences
(Table 8 Appendix III) . Hiese included variables of the original vdiole cd±(le
and its reduction sequence: cobble raw material, length, width, thickness,
wei^t, amount of cortex present, reduction type (bipolar, free hard hammer
for flakes, narrow biface production, and broad blade production) , reducticai
stage, and cobble end product (s) (rejected material, flakes, or specific point
types) . Other variables were recorded for intermediate tool forms and
finished tools. Each product of cc±tole reduction that was used in biface
manufacture was assigned an item nuirber. Length, width, thickness, and weight
for each item were recorded after each manufacturing stage. The amount of
cortex remaining on a piece after each stage was also noted.
After each manufacturing stage, flakes that could be iised for other small
tools were removed from the debitage. The remedning debitage was collected as
a batch after each manufacturing stage and was sorted using 1-inch, 1/2-inch,
1/4-inch, 1/8-inch and 1/16-inch square mesh hardware cloth. The screens used
were the same ones used to sort the archaeological debitage during Phase I and
II. Within each size grade, debitage was counted and weighed, ai>d the number
of pieces with cortex was recorded. After removing usable flakes, the
debitage produced from bipolcir and free hard hartmer reduction for flakes was
processed in the same manner. In addition to the above flake information
collected for mass analysis (Abler ) , individual flakes were placed into
nine categories based on polythetic criteria developed by the replicator. The
definitions of these categories cure given in Table 9 of i^jpendix III. Raw
data files and SAS data files for the replication experiments are available on
t^)e.
Summary statistics for the intermediate products and end products of small
and medium cobble reduction were calculated (Table 97) , and several
observations can be made about the reduction of the experimental bi faces:
TRBIf 97
Mean dinensicns of original oobbles and bifaoes prodooed at stages in
redaction sequences. _
N Mean Standard Deviation
Reduction Sequence !_
Original Cottoles Reduced by Hard Hammer Methods for Flakes
Length
12
107.65
20.49
Width
12
76.26
13.17
Thickness
12
49.58
10.60
Weight
12
547.60
228.95
259
•dubu; 97
Mean dimensians of Griguial onhhles and hi faces pcodhioed at stages in
redaction sequences (ooptinoed) . _ _
N
Mean
Standard Deviation
Reduction Sequence 1
Exha;xsted Cores from Hard Hanmer Reduction for Flakes
Length
3
76.03
18.87
Width
3
55.10
6.46
Hiickness
3
33.87
4.92
Wei^t
3
126.40
50.34
Reduction Sequence 2
Original Pebbles Reduced ly Bipolar
Methods for Flakes
Length
12
58.67
7.17
Width
12
38.43
7.14
Thickness
12
26.87
4.56
Weight
12
71.58
18.99
Reduction Sequence 2
Exhausted Cores from Bipolar Reduction
Length
7
37.39
5.30
Width
7
26.21
6.82
Thickness
7
17.09
7.61
Weight
7
24.37
25.53
Reduction Sequence :
)
Small
Cobble Reduction for Narrow Blade Bi faces (Little Bear Creek)
Length
3
98.23
18.40
Width
3
57.50
5.58
Thickness
3
43.13
9.96
Weight
3
258.20
160.84
combined Small Cobble Reduction eind Initial Edging and Thinning Product
Length
10
59.03
9.44
Width
10
32.86
8.19
Thickness
10
13.33
2.81
Weight
10
28.86
18.80
Combined Secondary Thinning
and Pressure
Finishing - Finished Tool
Length
11
56.57
8.67
Width
11
21.74
2.67
Thickness
11
9.90
1.84
Weight
11
10.57
3.94
260
N
Mean
Standard Deviation
I
Reduction Sequenoe 4a
Medium Cobble Reduction End Product Broad Blade Bifaoe (Kirk)
Original Gobble
Length
4
95.60
5.47
Width
4
68.86
2.05
thickness
4
45.05
12.73
Weight
4
307.32
63.81
Flake Blanks and Nuclei
Length
10
70.67
17.39
Width
10
49.02
10.04
Thickness
10
18.12
8.25
Wei^t
10
68.14
69.29
Edging and Initial Thinning Products
Length
10
65.78
16.01
Width
10
40.88
7.48
Thickness
10
13.56
5.08
Weight
10
43.83
30.87
Secondary
Thinning Products
Length
8
56.91
16.26
Width
8
31.62
5.11
Thickness
8
7.91
1.14
weight
8
15.39
9.18
Finished Tools
Length
10
55.61
14.26
Width
10
30.99
4.41
Thickness
10
7.07
2.35
Weight
10
13.52
7.80
Reduction Sequence 4b
Medium Cobble Reduction Eiid Product Broad Blade Bifaoe (Greenbricu:)
Original Gobble
Length
2
126.10
19.52
Width
2
86.55
28.77
Thickness
2
59.20
10.61
Weight
2
691.45
536.91
Flake Blanks and Nuclei
Length
9
91.86
21.12
Width
9
69.10
13.85
Thickness
9
20.11
4.56
Weight
9
122.72
71.20
BIBIE 97
Mean dinensicns of oariginal ooK>1es and h-ifarit^g pcodooBd at stages in
redaction sequences (oontinoad) . _
_ N _ Mean _ Starxlard Deviation _
Reduction Sequence 4b -
Medium Collie Reduction End Product Broad Blade Biface (Greenbriar)
Edging and Initied. 'Running Products
Length
9
78.77
21.44
Width
9
49.60
13.98
Thickness
9
13.90
3.29
Weight
9
56.88
35.70
Secondary 'Thinning Products
Length
9
74.31
20.81
Width
9
34.47
5.35
Thickness
9
9.94
3.03
Wei^t
9
28.91
16.67
Finished Tools
Length
7
74.71
23.01
Width
7
29.49
4.47
Thickness
7
8.10
1.82
Weight
7
20.96
14.33
Reduction Sequence 4c
Medium Cobble ReductiCTi End Product Broad Blade Biface (Eva/Morrow Mountain)
Original Cobble
Length
3
99.97
18.07
Width
3
69.27
4.34
Thickness
3
48.90
5.30
Weight
3
379.10
84.63
Flake
Blanks and Nuclei
Length
11
75.17
10.49
Width
11
55.45
7.37
Thickness
11
21.69
4.79
Weight
11
75.70
36.17
Edging and
Initial Thinning Products
Length
9
55.90
13.58
Width
9
39.38
4.97
Thickness
9
12.58
4.99
Weight
9
31.21
21.73
Secondary Thinning Products
Length
4
55.12
6.31
Width
4
35.17
3.22
'Hiickness
4
9.67
0.97
Weight
4
17.00
2.74
262
raeu 97
Mean duensions of corigiiial onMales and bifacBs pcodoced at stages in
redoctiop sequenoes (oontinoed) . _
N Mean Standard Deviation
Reduction Sequence 4c
Medium Cobble Reduction End Product Broad Blade Biface (Eva/Morrcw Mountain)
Finished Tools
Length
8
52.61
5.87
Width
8
34.46
2.50
Ttiiclcness
8
9.07
0.75
Wei^t
8
13.87
2.99
Reduction Sequence 3d
Medium Cobble Reduction End Product Broad Blade Biface (Sykes-White Springs)
Origineil Cdable
Length
7
104.43
6.95
Width
7
82.97
12.56
Hiickness
7
46.44
10.28
Weight
7
437.67
126.46
Plate
Blanks and Nuclei
Length
21
80.09
14.38
Width
21
55.19
15.61
Thiclcness
21
23.47
8.53
Weight
21
110.53
108.75
Edging and
Initial Thinning
Products
Length
18
66.53
19.62
Width
18
37.34
11.89
Thicloiess
18
14.10
5.30
Weight
18
35.55
16.23
Secondary Thinning Products
Length
9
74.31
11.16
Width
9
36.24
8.75
IhicJaiess
9
10.86
5.07
Weight
9
32.23
22.17
Finished Tools
Length
6
73.12
10.54
Width
6
28.58
2.12
'Thickness
6
7.65
0.95
weight
6
17.47
6.00
263
mu 97
Mbob dinBBBlQBs of ocigiiial ootMea and bitaoes prodnoBd at stages in
rednctinn snquwioBS (oostiiMBd) . _
N Mean Standard Deviation
Reduction Sequence 5
Fort Payne Blank Reduction for Broad Blade Bifaoe (Benton)
Original Blank
Length
10
113.97
14.17
Width
7
76.10
7.37
Thickness
10
24.20
4.83
Wei^^ht
10
289.96
55.49
Edging and Initial Thinnirig
Product
Length
10
96.98
15.04
Width
10
55.49
14.13
Thickness
10
17.34
4.96
wei^t
10
101.29
50.21
Secondary Thinning Product
Length
8
88.89
13.11
Width
8
39.59
8.16
Thickness
8
10.42
5.29
Wei^t
8
40.20
22.46
Finished Tool
Length
5
81.26
10.89
Width
5
32.38
4.36
Thickness
5
8.34
1.30
Weight
5
24.54
7.51
1) A relatively small niirber of cobbles was necessary to produce material for
all the finished bifaces. On the average, each cobble produced at least
two flake blanks or nuclei for further reduction. The number produced
deponded on the size and quality of the original, cobble. Only nuclei and
flake blanks of suitable size and those without major flaws were selected
for further biface reduction. The largest cobble used to manufacture
Greenbriar bifaces yielded one large nucleus and four large flake blanks.
Many analler flakes suitable for the manufacture of patterned unifacial (
tools and edge retouched flake tools were adso produced from this cobble.
2) A large number of cobbles were needed to produce the five finished
Sykes-White parings bifaoes, because l^ucge nuclei or flake blanks are
necessary for their manufacture. These bifaoes were knapped toward the
end of the replication sessions. The mediixi^-sized cobbles remaining for
tool manufacture did not provide more than one flake blank of both
sufficient size and quality for the manufacture of these plano-plano
bifaoes. If raw material sv^jplies were in any way limited
prehistorically, suitable cobble material for their manufacture might be
difficult to obtain.
3) Not surprisingly, the greatest material loss occurs during the process of
cobble reduction and edging and initial thinning. The loss of materied
decreeises as tools become more refined. Variation among items generedly
decreases with refinement.
264
4) Regardless of whether nuclei or flake blanks are used at the j^inning of
the inana£ac±uring sequence, width and thickness are fairly unifomi for the
finished products. Lengrth is more variable.
Ihese data can be ocnpared with statistics archaeologiceil ^lecimens
during Chase III to provide insights about tool menufacture and use.
For exanple, vhen the mean length, width, thidoiess, and wei^tt of named
hafted bifaoes are compared, the archaeological specimens are sometimes wider
and thicker than the eaqierimentally produced items, but they are etLl shorter
on the average (Table 98) . Ohe difference is greater for ^kes-ftiite Springs
and Benton bifaces. Ihis suggests that the archaeological specimens mic^t
have been resharpened. If resharpening does not occur edong the entire edge,
i.e. if it is resharpened in the haft maximnn artifact width would not be
reduced.
widtii. tfaicknees.
265
Archaeological exan|)les of Greenbriar and Kirk bifaoes are more
standardized ^dGng these measured dimensions than their experimental
counterparts, vAiile Eva/Morrow Mountain, and Cypress Creek bifaoes are much
less standardized. Archaeologic2d exanples of Greenbriar and Kirk bifaces
appear to be more refined, v^le the others, especially the Cypress Creek
bifaoes, appear to be more irregular both in plan and side view. Several
possible explanations can be proposed for these differences: tool function may
differ among the blfaoe types; pr^iistoric knapping skills and the skills of
our replicator may be different; concerns vd.th standardizatil edge in cme case. Distinct ridges separated tool edge and
surfaces.
Use-Rodified edges and pzojecticms were created cm umcdified flakes,
flakes with retouched edges and cm bifacial heifted tcmls (Figures 46 and 47) .
All tools were made of Camden c^ert excapt in cases noted below. All Geraden
bifacial tcols, ewapt Little Bear Creek #2, were heat treated. Except for
those flakes with the letters "UF" in their identifying ncmber, flake bools
are not heat treated. All tools were sketched before use, and tool edges were
examined micnascxpically. All pre-use edge modificaticm were noted, fiumples
of recording sheets are given in Figure 1, Appendix III. Tools weia usuedly
used for tasks that were likely to have been done by prehistoric populations;
for example, hide preparation, making tool hafts and foreshafts ftrxn oak or
cane, flitting bone to create awl blante, and throwing. The motions
used in these tasks and the amount of time each edge vms used were recorded.
Repetiticm of taslcs was not cxmsistent and depended cm materials, tools, auad
time available. After use, tool edges were cleaned with weak solutions of
hydrochloric acid and sodium hydroxide and examined microsoopically
(Table 99) .
Figure 46 Hafted end-scraper used in use-wear experiments.
99
r of
So^t
■ "
€:esh deer hide
7*
1
chideen
(9)**
7
(1)
Medium
(10)
i
cane
6
1
1
oak
(7)
3
(1)
5
2
(1)
4
3
dry hide
Hard
(3)
(5)
3
(3)
(2)
(4)
(3)
i
bone
5
2
3
8
1
4
(8)
(2)
(3)
(17)
(1)
(4)
* niirber of tools
** ramber of functional units
In addition to the tools used in our experiments, 28 unmodified
flakes stored in protective envelopes Mere loaned to the project for
the litlac replicator. These include six wedges of Fort Payne chert
bone, three vedges of Fort P£^ne chert used to split wood, one flake
skin a fatty raccoon, 11 flakes used to cut deer sinew, three flakes
cut deer skin or meat, four flakes used to sa»r, plane, and chisel wo
last 19 items are made of Coxsackie chert, a dark blue-gray chert fic
England.
The following use-wear patterns were present on the ejqperimental
CUTTING/SMyiNG
BOWS; IVro hafted bifaoes (Benton #1 [Port Payne] and Little Bear Creek #2) and
three flakes (flake #31, #45 and #2) were used to saw bone. After a total use
time of approximately 20 minutes for two edges, one functional unit on flake
#31 exhibited anall flake rectangular flate scars with step terminations.
Wear was primarily on one side of the edge, since the tool was used at a 45°
angle to the bone. The other functicxial e^e is cortex. This edge became
rounded. No microflaking occurred. Flake #45 (total use time for 2 edges: 13
minutes) edges shewed small triangular flake scars with step terminations on
both faces, although flake scars were more invasive on one face than the
other. Ed^s wore down very quickly. One functional unit on flake #2 was
used in a sawing motion but at less than a 90° angle to the worked bone.
Flakes cme-third the size of retouch flakes were removed from one side of the
edge. Flake scars had step terminations. After a total use time of 26
minutes for both edges, serrations on hafted biface #2 were worn down, and
edges appeeu: crushed. Wear was most evident on the middle of the edge
segment. After about 13 minutes of use on each edge, edges became noticeably
dull. After a total use time of 57 minutes on caie edge only, Bifaoe #1
269
I r
developed an eOnost ground appearance on the very edge. All hi^ spots along
the edge were obliterated. Flake scars with step terminations extending onto
both faces developed, althcu^ it was scmetines difficult to separate flake
scars removed through manufacture and use. One edge that had been use on wood
and had developed rounding and poli^ was used on bone (Benton #4) . After 90
minutes of use, rounding and polishing were obliterated by crushing and flake
scars with step terminations developed.
Cft^; Ihree hafted bifaces (Benton #4 (Fort Paynel , Kirk #3, Little Bear Creek
#5) , one unhafted biface (Eva/Morrow Mountain #1) and two flakes (#5 and
#13UF4) were used to cut cane. After an average use time 45 minutes, biface
edges beceme blunt rather than crushed. Blunt edges were not as smooth or
regular as ground edges, but did not eidubit the more extreme removal of
material seen in crushing. At 50-70 power blunt edges are ocnposed of short,
overlifjping, rectanguleu: flake scars with step terminations. Flake sc^u:
ridges and st^ terminations due to manufactrure were anoothed over. Wear was
primaurily on the of the edges. After 90 minutes of extended use of
Kirk #3, one functioned, unit appeared rounded and had developed a sli^t
polish. The serrations on biface #5 wore down in the first five minutes of
use. Flakes #13UF4 (total use time 30 minutes for two edges) and #5 (total
use time three minutes on one edge) were hand-held. FleOce sceurs were removed
from both sides of \ised edges. Scars are scaler. Wear was more extensive on
the middle of the used edges.
OAK: Three tools were used for cutting oak branches, one hafted Bentrai (#1)
and two flakes (#45EB4 and #39) . The Benton edge became blunt after 30
minutes. Smoothing of manufacturing flake scar ridges occurred further i;p the
faces of the edge than occurred on cane cutting edges. The used edges on both
flakes appeared blunt or rounded under 40 power magnification after ten
minutes. Scalar flake scars with both feathered, and step fractures were
removed primarily from one side of each used edge.
FRESH MEAT/HIDE; Three hafted bifaces (Kirk #4, Bentcxi #1 [Port Payne] and
Benton |3 [Fort Payne] and four flakes (#100P1, UF5 and UF6, #221F2) ware used
partially to dianember nine chickens. No attenpt was made to avoid chicken
bone. A wooden cutting board was used, and after use the cutting board had
175 anall cuts and two puncture marks. The bifacial tools were used for a
total of 23 minutes. Both edges of each bifacial tool were used. After use
on the chickens and microscopic examination, these bifaoes were reshaupened
and used on wood and bone. The flakes were used for a total of 25 minutes.
Tools seemed to becane ineffective after a very short period of use. This was
probably due to the aocumilation of grease or fat on the edges, rather than to
edge modification. Brose () noted that flakes used in butchering are
usuedly used for short periods of time (three to four minutes) . They becane
ineffective because of the accumulation of animal fat. The short period of
xise may leatve no traces of wear. Unhafted tools were also difficult to hold
once they became gxeeisy. Bifacial tools shewed no wear that could be
attributed to use on the chickens. None of the bifaoes were used long enough
to edter in any way we could measure the flake sceu: ridges or the sli^itly
crushed edges produced during manufacture. Three of the four flakes used
(#55UF1, #22UF2, and #10UF6) showed wear in the form of ve^ shallow,
bifexrial, scalar or amorp^xjus flakes with feather terminations. Edge segments
may appear wavy and smooth. T\ro of the flakes also have small nicks removed
from edge segnnents.
LBK: lliree of the four flaioes maobered flakes (55UF2, 22UF1, and UF2) used to
skin a deer shewed traces of wear. Both sides of xiaed edges exhibit
irregularly spaced, shallow, amorphous flake scars and mall shallow scedlops.
Occasionally mall, shallow flakes of a triangular or rectangular sha^ have
been removed fixm the edges. Several of the edges also had small scattered
nicks. These are probably due to contact with bone. Three unmodified,
unnrnbered Fort Payne blades were also used on the deer. These were not
examined ndcrosoopically before use. Each of the flakes were used for about
five minutes. Although all three had excellent cutting edges, they were
difficult to hold. Vfear similar to that on the three Camden flakes developed
on edJ. five of the functional units used. *
All but one of the 13 flakes used to cut deer skin or meat and to cut deer
sinew shewed similar wear. Edges have shallow eraorphous flake scars with
feather terminations. Cltmps of mope defined flake scars, and isolated flake
scars with step terminations and nicks are probably the result of contact with
bone. These flakes are an. the vhole larger than those made of Camden or Fort
Payne, and a few are backed. They may have been ecisier to hold.
SCRAPING/PLANING
BOME: The fracture edge of a broken biface (#22N) was used to screpe bone.
The edge had an approximately 90** edge angle. After 35 minutes of use the
edge was obviously rounded and smoothed. Flakes were ranoved from the edge in
contact with the bone, but these were quickly smoothed over. One edge of
flake #37 was also used to scrape bene. After 30 minutes the edge showed very
little wsar. Fifty power magnification revealed only slight polishing on the
very edge.
DBOf HEMi; Three hafted unifacial scrapers (scrapers #2, #3, and #4) were used
to scrape dry hide. In all three cases extreme rounding and polishing
occurred on the edge and extended on to the dorsal side (the side with
greatest contact with the hide) obliterating dorsal flake scar ridges. The
rounding was apparent after as little as 15 minutes. The polish was matte,
but apparent at relatively lew magnification (25 power) . On two edges
striations perpendicul^u: to the war)ung edge developed. Only one microflaJee
was removed from all of the screping edges. Those scrapers with the most
regulcu: edges in plan and side view were the most effective scrapers.
CANE; One flalce (#5) was used to scrape cane. Staedl, scaler flakes were
removed from one side. No rounding occurred. The edge was used for 10
minutes.
OAK; Five flakes (#10, #19, 20, #30UF1, #13UF5) - average use time for
5 edges = 30 minutes - were \ised to scrape oak branch segments used to make
tool Ivandles. Three edges exhibit even, small, scalar flalce scars primarily
on one side. At 40-70 power, edges also 6ppeared to be rounded and smoothed.
One cortex edge and one edge of chaUcy material have few flake scars. Edges
were smoothed and rounded. The smooth appearance on one piece extend up one
face more than tlie other.
raESH HIDE; One bifacial hafted sersper (#1 [Fort Payne]) (Figure 46) was used
to flesh a &:esh deer hide. The tool was us^ with both a pushing aivi pulling
motion and was usually held at a 45-75° angle to the worked material. Wear
developed very slcwly. After almost two hours of use, sli^t rounding 2uid
271
n—i»«-h1ng (*— waring") of nasufacturing flake aoers was aotlcMhU at 40 power
Mprifiral 1wear on cuxihaeologiced. specimens. The two pieces stnx:k with
a hamnerstone hcKi menre extensive battering.
IMPACT
Seven biface points (Eva/Morrow Mountain and Kirk types) were mounted in
cane foreshcifts and shafts. Oak trees and a large noose antler plate were
used as tcurgets. Severed types of inpact fractures were produced. One point
oonpletely shattered on inpact with the antler plate. Fractures on the pieces
recovered are very irregular, and althou^ they sure not sis smooth, they sune
similar to heat-produced crenate fractures. TV«3 have staedeed st^ fracture
emanating from the tip, and three have single large flsdce scsirs originating
from the tip. The flake scars shew prcax>unced rippling. One point which was
thrown against wood had a transverse frsKi^ure with tongue. In the process of
hurling these spears, targets were missed and points struck the ground. No
inpact fractxires resijdted. Our target practice was of a very informal nature.
The type of fracture that occurred on points was rxj doubt due to the force
with v^ch the spear was thrown, the angle at iddch it struck the target, eis
well as the type of material struck.
MISCELLftNEXXB
One drill was used to pry ice off of a refrigerator freezer. The tip
snapped producing a transverse fracture with a smooth tongue. A flake used
for an hour to sldn a fatty raccoon slxws little wear. Only one small section
at the end of the flake shews shallow flake scars.
The wear patterns described above are similar to those sunmarized by Lurie
() and to those described in Brink () and Huckabay () . Sane
additional observations on peurticular raw materials are given below:
1. Cortex seems to collapse with use. There is little microflaking no matter
what lund of task is performed. Functional \anits become blunt or rounded
rather than flaked.
2. Scraping wood with low edge angles produces microfla)djig quidtly, but
edges then become rounded obscuring the flake origins.
3. Vfear is more apparent on heat-treated flakes. Used edges may be eeisier to
identify if there is luster oontreist between the flake surface and the
used edge, or heat-treated material may flake more easily.
SECTION ^ V7VRIAHLE SELBCTIC^t AND ARTIFACT RECORDING SCHEMES
The eurtifact recording scheme used for the 6,391 stone artifacts examined
as part of the Phase III anedysis contains three types of information for most
variables: 1) the rationale for selecting a particulcu: variable to record; 2)
a discussion of attribute states, and their definitions when aipropriate, for
each variable; and 3) the conventions enployed in assigning these attribute
states. Exaitples of the scheme with appropriate computer information and
exaitples of the artifact recording sheets eure provided in Table 10 and Figure
2 in ^pendix III. Raw data files and SAS data files for all tools coded in
the Phase III study are available on tape. Vcuriables will be discussed under
the follonang headings: provenience, raw material, heat treatment, technology,
function, hafting, and mcrjAiology. It should be noted that many of the
variables provide infocroation on more than one of these topics. 'Derms given
in capited letters indicate the name of variables. Terms in quotation marks
aie attribute states.
PROVENIENCE: Provenience includes the site number, excavation block letter,
excavation level, feature number v^ien appropriate, an arbitrary field number
eissigned to excavation units or sections of excavation units, and artifeict
catalogue number. Ihe site number, ID, amd catalogue number together
indicated a unique item. Both inter- and intra-site carparisons can be made
using these units. Units wezre aissigned to one of the six time categories
(Eeurly Archedc, Middle Archeiic 1, Middle Archaic 2, Middle Archaic 3, Middle
Archaic 4, and Late Archaic) . Ihis facilitates testing expectations about
changes in tool manufcicture and xise.
RM? MATERIAL: In the preliminary analysis of all lithic material from the
midden moun(l sites, 35 types of raw materials were defined. Of these a
relatively small nutber make up over 95% of all the eurtifacts recovered:
Camden chert, yellow chert, blxie-gray/tan Fort Payne chert, fossiliferous Fort
Payne chert, ferruginous simdstone, Tallahatta quartzite, Pickwick chert, and
Bangor chert. Other types of materials that occur infrequently include
non-Tallahatta quartzite, quartz, conglonerate, hematite, limonite, and a
nutber of exotic cherts. Camden, Pickwick, and yellow cherts, quartz amd
qucurtzite pebbles, ferruginous sandstone, and conglomerate are derived from
the Tuscaloosa formation that cap the uplands in the north and east of the
study area and may have been present in the Tcetbigbee Valley as alluvial
gravel. These lithic materials were considered local raw materials in this
stu<^.
Pour geological formations (Tusceiloosa, Fort Payne, Bangor, and
Tallahatta) sipply the vast majority of raw materials. Of these the
Tusc^doosa formation (Marchen and Steams ) , capping the uplands to the
north and east of the study area, was the most extensively exploited. It was
the parent material for most of the alluvial sand, gravel, edible, and boulder
deposits contained Camden, Pickwick, and Yellow cherts, quartz, aM quartzite
pdbles, ferruginous sandstone, cind conglomerate. These lithic materials are
considered local raw materials in this stuefy.
CAMDEN CHERT: Camden chert occurs eis well-rounded cobbles with a cortex from
1-2 mm thick. It is highly variable in color, texture, luster, and
workability. The most ccnmon colors range from vdiite to yellow to olive
yellow (2.5Y8/2, 2.5Y8/4, 2.5Y8/8, 2.5Y7/8), although some pieces are light
gray to gray (2.5YN7/, 2.5YN/) . Grain size raitges from fine to coarse, and
several colors and textxures are often found in the same cobble. Luster ranges
from dull to medium, and knapping quality ranges from poor to fair. The
material is often "very heu^d cHid tough" to knap v^ien unheated (Kalin, personal
cdtimunication: ) .
YELIXW TUSCALOOSA: Yellow chert also comes in well-rounded cobbles with a
thin cortex. Color ranges from yellow to yellowish brown (10YR6/6 and
10YR5/8) and is uniform throughout the cdble. Grain veuries from fine to
ooaurse even within the same odble. Luster varies from dull to medium.
Flaking quality ranges from poor to good.
PIQBIICK: Pickwidc chert outaxpe in the Fort Payne fonwation in the western
mldtSie Tennessee Valley, particularly PidcMick Reservoir. Pidwick chert has
also been reinoorporated into the Tuscaloosa formation. Cobbles are distinct
because of their banding or mottling. Three colors are usiially present
blue-blade to dark gray (5B4/1, 5EN4/, 5EN6) at the cortex, yellow to vAutish
yellow (10YR8/3, lOYRS/S, 10^7/8), and red (2. SYR 6/8-4/8) at the center.
The material found in the Fort Payne fannation nomally has a hic^ier porosity,
medium to coarse grain size, and dull to low luster, but flaking is fair to
good. Pickwick diert incorporated in the Tuscaloosa fomatlon has nedium to
fine grain due to re-silification. Luster is dull to medivm, and flaking
quality is fadr. The material found in Pickwick Reservoir forms in flat,
angular cobbles with heavy cortex, but the redeposited Tuscaloosa cobbles are
rounded with a thin cortex.
EHWXSDPQS SMPSTOME; Ferruginous sandstone is found in the Tuscaloosa and
other l^pp^ Cretaceous formatiens in the research area. Oiartz sand grains
are cemented by silica and iron coopounds whidi give it a reddish blade to
black color (2.5YR3/4, N2.5/) . It is coarse grained, but can have a medium to
meditm-fine luster on fresh breaJes in well-oenented pieces. Althou^ flaking
quality can be fair in strongly cemented specimens, most of the tools made
from this material have been pecked or ground, either intentionally or throu^
use.
COWaCMEyATE; Conglomerate is oonposed primarily of rounded or subangular
yellow chert fra^nents greater than 2 nm in diameter cemented by a
fine-grained matrix of quartz grains, iron oxides, and silica. This is a
coarse-gredned material with medlmn to medium-fine luster on fresh surfaces.
Flaking is fair in well-oemented specimens. Conglomerate was often used for
large, heavy-duty tools.
The closest outcrops of the Fort Payne formation are in the middle
Tennessee Valley in the extreme northeastern comer of Mississippi
approximately 75 km (46.5 mi) from the sites investigated (Stadth ).
FORT PAYNE CHERT; Fort Payne chert is highly variable, ranging ficom light
gray or blue to blue-gray, dark gray, or blade (7. SYR N 7/, 5B 7/1, 5B 4/1,
7. SYR N2) , with blum translutoent mottles occurring in some of the darker
varieties. The lifter colored, more coarse grained forms could beoome yellow
or tan with weathering (10YR8/6 and 10YR6/8) . Most of the Foxrt Payne cherts
has a porous, medium to coarse grain, althou^ seme of the darker specimens
are more ccnpact and have finer grain. Luster ranges from dull to
roedium-hi^, with darloer specimens having hi^ier luster. The flaJeing quality
ranges from fedr to excellent. Nodules often have a thick cortex of
rough-grained coarse chert \duch must he removed to reach the more worlcable,
finer grained interior.
F06SILIFERCXJS FORT PAYNE; The fossiliferous fomns of Fort Payne range from
li^t gray to blue-gray, to bluie, vhite, tan, and brown (10YR7/2, 5BG6/1,
5BG5/1, 10YR8/2, 10YR7/6, 10YR6/8) with fossils normally being a sli^tly
darker opaque shade than the surrounding matrix. The grain and fossil size
seems to correlate with the chert color. The darker blue-gray and blue forms
have a medium-fine gredn size and highly frequented aiell fossils peppered
throughout the nodule, while the lifter colored cherts h2nre a mediim to
• coarse texture with leurger and more varied sizes of fossils. Finer cherts
275
have a medltin luster. Ihe coarser cterts eidiibit a dull luster or no luster
at all. Ihis material often forms in thick, blocky nodules, with square block
fractures, which have medivm to thick pitted cortex. !n)e flaking quality is
fair to gcxxi. Ihe majority of fossils are crinoid frai^nents, but the key Fort
Payne formation fossil indicators are ^leciflc brachiopods, and the absence of
Bryozoa.
11)6 closest source of Bangor chert to the midden mound sites was
appraximately 8Q km (49.6 mi) to the east (Bond ) . Fossiliferous forms of
Bangor chert derived from the thidc-bedded, dark bluish limestone usually
occur as blody slabs, itdiile non-fossiliferous types occur 2is rounded nodules.
BftNGOR CHERTS; lluree types of Bangor chert (Bl\ie-Green, Little Mountain
Bangor and fossiliferous Bangor) were described in the Phase I Interim R^ort.
Of these only Blue-Green and fossiliferous Bangor are present in the Phase III
saople. Blue-Green Bangor ranges in color frcm li^t blue-gremi through dark
blue-green, and sometimes dark gray (5BG6/1, 5BG4/1, 5y4/l) . It is usually
unlfooily colored, although sometimes there was a shift in color shade from
the exterior to the interior of a cobble, with the darker diade on the
interior. It is fine grained, and thin flakes tend to be translucent. Ihe
Itister is mediixn to hi^, and flaking quality is fair to excellent. It occurs
in irregular, flat to round nodules and has a cedcarecus cortex.
Fossiliferous Bangor has a similar color range. Fossil fragnents often appear
translucent vAiite. Key index fossils include Bryozoa. Ihe background matrix
surrounding the fossil inclusions is fine grained, but sometimes tie fossils
have been leached out, leaving voids in the chert. Luster is medium to high.
Modules are normally thick and blocky with square tiracture planes.
Ihe Tallahatta formation (Copeland ) is found in outcrops across
south-central Alabesna and central Mississippi. Ihis is the most distant
source of lithic neterials (160 )an or 99.2 mi frcai the research area) ocmnonly
used for midden mound tool manufacture.
TMUfflftJTA gUARTZIlE; Tadlahatta quartzite is a sandstone that has been
metamorphosed to quartzite by silica cementation. Talladiatta is recognized by
the ^>arkle of its crysted gredns, and is sometimes referred to as sugar
quartz (Lloyd et al. ) . This median to coarse-grained material varies in
color according to its degree of weathering - from gray and blue-gray with
vdiite mottles when fresh to opaque tan when weathered (2.5yM8/, N7/, 7/2, N6,
6/2) (White ) . Relatively coarse grain and poor cementation causes
Tallahatta quartzite to erode rabidly. When freshly broken flaking was fair
to good, but after artifact manufacture weathering obliterates flake scars.
Fort Payne, Bangor, and "Otiier" cherts as well as Tallahatta quartzite are
considered nonloced raw naterials.
Other raw materials used in small quantities for tools include iraci ores,
hematite and limcxiite, petrified wood, and greenstcxie. Ihese were described
in the Interim Reports.
Raw material type was scored for all archaeological specimens. IVpe was
based on the Ikuversity of West Florida's corparative collection as well as
the previous descriptions. If items did not match any of the samples in the
collection, they were sorted as "Other chert" or "Other Raw Material." Ihe
amount of cortex on a tool or piece of debitage was also recorded to the
nearest 5%. Ihe amount of cortex on the tool is often an indicator of
reductiCTi stage and manufacturing technique.
276
Raif material quality was scored for all chert. Evaluation of chert as
good, fair, or poor was based on texture, the presence of fossil inclusions,
and the presence of fracture planes. On the vhole, the quality of Camden
chert, by far the most oomnon raw material in the collection, is good. Ihe
quality of very small pieces of chert, or diert pieces that had been overly
daneged by heat, were scored "can't determine." At present there are no
criteria established for evaduating the quality of stone oomnanly used for
ground stone iaplements.
HEAT THEAMIBir; Ihe a^lication of heat to chert often causes visual and
structural changes in the material. The series of heating e9q)erimmit:s
conducted to document changes in color, luster, and workability in Camden,
Pickwidc, and Fort Payne cherts has been described in Section 3 of this
chapter. Prdiistorically, cherts could had been intentionally heated as part
of the manufacturing sequence or heated accidentadly, or heated both
intaitionally and unintentionally. Both intentionally heated and
unintentionally heated pieces can exhibit the same characteristics of color
and luster. Heat treatment is the term usuedly used to indicate the
intentionad heating of materials to ir^xrove their knapping quality. In the
absence of prehistoric heat-treatment facilities, it is not possible to
measure intentional heating directly, but the selection of heated pieces for
tool manufacture can be documented (Belim and Green ) . In the recording
system the variable HEAT TT^EATMENT refers to pieces with characteristic color,
luster, and possibly pot lid fractures and crazing, resulting frcm heat
application, that have been woidced, and/or used after heat application, this
selection of materials iaplies an intention to heat.
Because different cherts re^xnd in different ways to the application of
heat, sanples of unheated and heated materials must be anxailable for
comparisons, the heat-treatment experiments provided a ccnperative collection
of uzheated and heated Camden, Pickwick, Fossiliferous Fort Payne and Wilson
Dam blue-gray Fort Payne used to evaluate heating. If heated and unheat'^
samples of raw materials were not present in the type collection, and if there
was no difference in luster on heated surfaces and flake scars indicating pre-
and post-heating surfaces, archaeologic5d pieces were scored "Can't Determine"
for heat treatment. HEAT TREATMENT was scored as "present", "possible",
"absent", "can't determine", or "not sqpplicable" , for all non-chert raw
materials.
The variable HEAT TREATMENT POINr CF OCCURRENCE pertains to the stage in
the manufacturing sec[uence in vhich heat treatment occurs. It was scored for
pieces vhich had HEAT TREATMENT recorded as present or possible. Attribute
states for this variable include: "cobble/oore stage," "flake blank stage,"
"bifaoe stage," "can't determine," said "not i^licable." Pieces scored
"absent," "can't determine," or "not applicable” for HEAT TREATMENT were
scored as "not applicable" for HEAT TBEATJBMT POINT CF OOCURRaKE.
HEAT ALTERATION refers to those pieces that have been heated but not
subsequently worked or used. Heat-treated pieces can also be heat altered.
For example, a roughly fomed bifaoe can be heated as part of the
manufacturing sequence and then further reduced to produce a finished tool.
This tool can subsequently be broken due to further epplication of heat aixi
discarded. The second application of heat will be characterized by various
forms of heat fractures and by smdced and burned surfaces. The scorizig for
heat alteration includes information on the state of manufacture attained at
the time of heat edteration. HEAT ALTTRATICN was scored "absent," or
"present" at the cobble/oore, flake blank, or bifaoe stage," or "present on a
277
finished tool,” "can't determine," or "not applicable." Although ground stone
tools were scored for this variable, the criteria used were more nebulous.
Burned sandstcne eidubits color change and has a crunbly texture.
TBCttCLOGY; TBCBNOLOGICAL dASS identifies the reduction strategies used to
produce tools. This assessment was made on the basis of tool attributes such
as the initial form of the modified piece, the mode of modification, and the
amount of shaping the tool has undergone. This variable provides a way to
evaluate the investment of energy in tool manuf2K±ure as well as a way to
describe manufacturing techniques and the selection of these techniques
throu^ time. Many of the attribute states reflect the results of the
eiqperimental program, as well as preliminary visual inspection by the project
lithic qiecialists. Ground stone items vnere grouped under three attribute
states primarily to reflect energy expenditure. These include intentionally
shaped ground stone items, such as use-modified ground stone, and unidentified
ground stone fragments.
Several technological class attribute states in the recording sc^mme are
for items normally considered ddoitage of various sorts. These attributes
were included to account for items classed as utilized flakes and chunks
(during £haae I and II) , but assessed as d^itage in Riase II. Many vmre fire
spalls without any apparent retouch or use wear, vhile others were
resharpening flakes «hose ground or crushed platforms served as the basis for
their origlnad classification as tools.
Blfaoe cross-section is helpful in evaluating manufacturing techniques,
point styles, and, in seme cases, function. For example, in the experimental
program, reproduction of Kirk, Eva/Morxow Mountain, Sykes-44hite Springs, and
Benton points consistently produced items with different cnnes-sechians.
These (uxjss-sections were related to the type of blank used, and flaking
sequence. The diamond cnboss-sechion of most drills is lilcely related to their
funchion. Five named cross-section shapes are included among the attribute
states in the recsording scheme: bixxmvex, plano-convex, plano-plano,
rhcnboidal, and diamond. Irregular cross-sechiens were scored as "other.” In
most cases the cnnoss-sections of broken bifaoes were reexedad as "can't
determine." Non-bifaclal artifacts were reocedad as "not applicable."
The GBCfETfdC SHAPE in plan view of edl relatively whole pieexs was
recorded. Geometric shape is important in desenribing stages of bifacx
manufacture and point styles, measuring energy investment in tool manufacture,
and 2is8es8ing tool function. Attribute states include "ovoid", "round",
"teardrop", "triangular", "rectangular", and "square" shapes, an
"amorphous-Ehape" extego^, and an "cth^" ernttegory. Ground stone tubular
beads and drilled cores were oexied as "rectzoigular . ”
PRESSUFE FLAKING OR RESHARPENING PATTEEN is a variable which £^lied to
shaped artifacts cxily. It recsords the pattern of pressure flaking in terms of
two techniques, inward verses downward application of pressure. Inward
pressure produces relatively long, narrow flakes with father terminations,
vhile downward pressure produces relatively short, wide flakes with more
abrupt terminations. Either one or both of these techniques cxxild have been
used to produce beveled edges often assexiated with particular point styles,
tool functiexs, or resharpening techniques. Serration, the production of a
sawliJee edge, is also oonsidered a pattern of edge modification on shaped
tools. Combinations of inward pressure, downward fmessure, and serraticxi were
included in the attribute states. If a bifaoe exhibited none of the above
modificaticxis, it was scared as having "no ocxisistent pattern." Biface
fragments with only small edge sections were scared as "can't determine" fcjr
this variable.
278
OCmsiBBSS OF lASr FCXM has three attriJaute states, %Aiole, broken or
can't detandne and ms reoogdad for all Itaae. The ocsipletennss of artifacts
in an HBDfwhlaje could reflect several t^pes of bdutvlors, sudi as intensity
of tool use, methods of artifact disposal, and prdiistoric traapling. A piece
was considered ehole if the entire outline vtes represented. Staall nicks along
an edge that could have been caused prehlstorlcally or during excavation or
storage eere xxst treated as breaks.
The variable EESOfT PRESQIT records that part of an artifact which has
entered the archaeological record. This variable allows sorting of artifact
categories which contain ocnparable infarmation. For exaaple, pieces of
undetermined element, comer fragments and edge fragments will not contain as
imxdi infomiBtion or will give misleading information about the amount of tool
manufacture, use, or curation at a site, since maany of these items represent
parts of the saam tool too small to refit. If segment type can be identified,
their nmiaers and distribution within a oenponent are more meaningful. For
example, broken projectile point bases may have bemi brou^t back to camp
attached to tool shafts. These broken proximal ends could h&ve been reworked
or discarded in an area of retooling.
For shaped pieces, the distal end refers to the functioned end of the tool
and the proximal portion to the butt end or hafting end. For tools with
discernible flake morphology, the distal end is the flake termination, and the
proodmal end is the striking platform end. Tool midsectiens have neither the
praxinal or distal end, although a proximal or distal end can contain the
midsection or part of the midaection. It was often difficult to judge the
oonpletenees of ground stone tools. The edges of worked items were exandned
for abinjptness of break, and potential fractures were examined for rounding
and/or vmathering. Many had to be soared as "can't determine."
Although fracture mechanics have been important in e9g>laining why flakes
of certain sizes or shapes are detached with various types of percussors and
help explain breaks of certain oonflguratlcns occur when too much, too
little, or poorly placed force is applied, the behavioral implications for the
range of tool fracture patterns found at archaeological sites has not been
fully explored. Johnson () suggests that fracture patterns could be
related to tool life cycles. Some fractures were the result of manufacturing
erzors, sonne result from use, and seme from disposal practices. He recorded a
nutter of fracture patterns that occurred on artifacts from the Yellow Cre^
Archaeological Project in northeast Mississippi. The initial impetus for
recording fracture types for midden ncund artifacts comes from the Yellow
Cre^ study. Ary correlations between variables vhich record stages of tool
production (PRODUCTION STAGE) , tool use (FWICriON) , arxJ fracture types are of
special interest.
In addition to Johnson's work, Callahan's () description of fractures
that occur during the manufacturing process have contributed to our
understanding of the fracture process. Senples of these fracture types were
produced during the experimental program. Additional examples were loaned to
the project by the lithic replicator from his own collection of Icmpping
failures. These formed a type collecticxi of frex^ure types with vhich
archaeological specimens could be oenpared. The following definitions of
fracture types were used in the recording scheme.
1. Direct fractures - fractures that initiate at the point of force
explication.
a. Perverse fracture - The perverse fracture (Crabtree :821) is a
twisting direct fracture vhich results "when the fracture plane
279
-4
twists on an axis of rotation oomesponding with the direction of
foroe" (Johnson :46). Ihe points of fmnoe application, bulbar
surface or negative bulbar surf2K:e, are often recognisable on these
fractures.
b. Overshot fracture - [reverse fracture (Johnson :44-45) outrepasse
(Cndsttee :80)- Overshot fincture refers to biface thinning
f laloes tdiich renove the opposite bifacial edge front which the flaloe
was struck. This t^pe of fracture can occur at all stages of
reduction. Incorrect striking angle is a cmnon cause for this type
of fracture (Callahan :85). Both overshot flaloes and the
truncated bi faces that they create were found in archaeological
context.
c. Direct surface firacture - These fractures result front the application
of foroe to a flat, thin surface rather than to an edge. This force
nay or nay not be intentional. The fractures eoddbit partial
hertzian cooes or radial pattern, and the point of force application
can be distinguished.
d. lopact fracture - These are direct fractures iihich create
longitudinally oriented flalce scars originating at the distal end of
a biface (Ahler :52).
Indirect fractures - Fractures vhicdi occur away from the point of force
application. Both Johnson () and Callahan () na)ce Ihe dichotony
between indirect and direct fractures. Ahler (personal oomunication)
points out that many transverse and diagonal fracturee originate at the
point of ixpact, but that they are less "violent" than perverae fractures.
Fractures along natural flaws can also occur at point of ijipact.
a. T^ransverse - [lateral snap (Johnson :47; Purt^ :134): end
shock fracture (Crabtree :60)). Tcansverae fracture is a t^pe of
indirect fracture «hich results in a strait break relatively
perpendicular to the long axis of a tool or flake. It occurs "when
the force of the thinning blow cacoeedB the elastic properties of the
raw material" (Johnson ) . Johnaon reports that these fractures
occurred primarily on unfinished bifaoes in the Yellow Credc sasple
and, therefore, were probably production failures. One of the
by^-products of the midden mound eoqjerimental program was the creation
of mai^ of Johnson's fracture types. Three variations of transverse
fracturee were recognized: transverse fractures which are strai^t in
cross-section, transverse with lip ^ftiich has a gentle or fairly diarp
bend at one margin, and transverse with tongue %diich has a nnre
gradual curved cross-section. This last type of fracture occurred
when experimenttd pieoee were used to pry hard material, or as the
result of inpact on hard surfaces.
b. Diagonal - diagonal fracturee asevaae the sane forme as transverse
fractures, but are not perpendicular to the long axis. It is only
possible to distinguish between transverse and diagoned fractures
when the long axis of a piece can be dstemdned. Fractures which
were 70** or less to the longitudinal axis were considered diagonal.
When the longitudinal axis of a piece could not be determine,
fractures were considered to be transverse.
c. Fracture alcng natural flaws - Fractures which initiate or follow
internal fracture planes or small crystal inclusions. Many of the
raw materials used in our experimental program shomd internal
fracture planes often coated with iron staining or lined with
crystalline formations that could predispose the material to fracture
in a way unintended by kn^fjpers.
d. Heat Fractures - Heat fraictures can take severod forms; the smooth,
curved, jigsaw- like crenate break (Purdy , ; Jc^inson :49;
Rick ) ; the pot lid fracture, a shallow bcwl-shaped section
popped off of a surface ;^thout any evident point of origin (Rick
) ; and finally the (Mvansion fracture which consists of "large
pot lid fractures turned sideways so as to truncate a bifaoe," i.e.
the pot lid will occur on the fracture plane (Johnson :50) .
E}qpansion fracture for the midden mund project also includes grcxps
of overlapping small pot lid fractures occur on the fracture
surfaHce of a broken piece and fracture surfaces oovered with so many
small, overlapping pot lids that the break hais an almost a sugary
texture. This type of frakcture occurred admost exclusively on dark
blue-gray Fort Payne chert.
e. Haft snap - Haft sn^)s occur at the juncture of a tool blade and its
haft element. The fracture ocxifigurations most often associated with
haft snaps aure transverse, tramsverse with lip, and transverse with
tCHigue. Jc^mson considers this breadc to be the result of tool use.
This relatively narrow area of the tool was presumably securely
anchored to a shaft and most susoq>tible to fracture under stress.
Ccmbinations of these fracture types are also possible. Categories were
established for those ocmbinations that occurred most commonly on experimental
pieces or that seemed ocnmon in a cursory examination of archaeological
specimens. Whole pieces were scored as "not appliczdsle" for this variable.
PRGDUCnON STAGE was recorded for all artifacts. Assigrment to attribute
states were to a gireat extent baaed on a ocmbination of other variables in the
recording scheme including technological class, various attributes of shape,
and edge finishing, as well as evidence of use. E)qperimental work provided
type specimens with which to conpare archaeological ones. Ihe aim of
recording production stage was to place an item within a manufacturing
sequence vdiich includes the following:
1 . Core/oore fragments - blocks or chunks of raw material with evidence of
flake removal.
2. Stage 1 bifaoes - crude bifaoes analogous to the nuclei produced during
early bifaoe reduction or the bifstoe products of initial edging and
thinning. They are irregular in pl5m and side view and usually shew no
evidence of wear. Most of the items in this group are probably knapping
failures of cxie sort or another.
3. Stage 2 bifaoes - bifaoes ocmparable to the products of secondary or
cKivanoed thinning in the experimental program. IhQr are m^un refined
bifaoes more reguleu: in plan view and cross-section than Stage 2 bifaoes,
but their edges have not been pressure flaked into a definable pattern.
They usually show no evidenoe of uae-^wsar, but they can exhibit edge
preparation. Ohese pieces were probably broken throu^ knepping errors.
4. Stage 3 bifaoes - highly refined bifaoes, regular in both plan view and
cross-section, although pressure flaking or resharpening patterns along
edges are l2kcking. They do not ha\^ hafting elements and usually shew no
evidenoe of use wear. Edge preparaticxi is often present.
5. Finished tools - finished bifaoes and other chipped, shaped, and ground
stone. Finished bifaoes are technologically ocnplete in terms of the
producticn of hafted narrow or broad blade items. Most of the pieces shew
evidence of vise and/or resharp^ung. Uiif^ial tools Mere placed in this
category based on amount of shaping and the presence of use Mear. Other
chipped tools were placed in this category if they showed evidence of use
wear. Technology as well as wear help place stuped ground stone in this
variable. Ground stone modified through use only vies considered a
finished tool.
Broken items as well as pieces that showed a oanbination of
characteristics that made it impossible to be placed in ai^ of the above
categories were scored as "can't determine." Unaltered chipped or ground
stone was soared as "not applicable" for this variable.
USE PHASE is edso a ocnpoeite variable and adnittedly subjective. It was
designed to place an item vdthin its use and di^xised context rather than
within a manufacturing sequence. Used items need not be finished tools.
Descripticns of the attribute states for this varizdsle are as follows:
1. Tools are considered vumsed if they show no evidence of use-wear. Shaped
tools that appear finished but had no evidence of use were scored as
"unused", although tentative functions could be eissigned to them under the
EUNdFIGN variables.
2. Used and still useful items have disoemible use-wear and are whole, or
enoui^ of the functional unit is present on broken pieces, so that, in the
estimation of the recorder, the tool could still be used for its original
function.
3. Used and discarded tools are items that have use-wear but are no longer
oonplete enouc^ to perform their original function. They have not been
resharpened or recycled subsequent to breakage.
4. Resharpened and still useful items have resharpened edges and are whole or
vhole enough, so that in the estimation of the observer its original
function could be peirformed. The usefulness of these items is very
subjective, indeed, since use-wear has often been obliterated by the
resharpening, and the original function can only be inplied. Resharpened
and discarded tools have evidence of resharpening and are broken in such a
Mmy that the recorder felt that the orlgirad. function of the tool could
not be performed.
6. Recycled and still useful pieces have been reshaped and/or reused for some
function other than its original one. Ccsinon types of recycled tools are
broken bifaoe points that have been reworked into a bifaces with
transverse working edges and broken bifaces with the used fracture edges.
Others are tool fragments which have projections that can be used for
incising or graving.
7. Recycled and discarded pieces are the seae as above, except that the
functional unit with evidence of the rec^led function is heavily worn or
broken.
nycnoW: study of tool function has a special orientation. The precise use
oi a tool is of no ^mcial interest, but information on intensity of tool use,
multipurpose tools, utilized flakes, and recycling is significant. The
assignnent of a furx:tion, or functions, to tools vAien possible was based on
evidence of microscopic edge damage. Interpretations of edge damage were
based cn use-wear e^qmrlments and detailed descriptions of wear reported in
the literature (Abler ; Odell , , ; Keeley ; Brink ) .
Macroscopic attributes such as edge angle and edge configuration were used in
conjunction with microscopic evidence (lurie ) .
BecaMae it Is possible for diffsrent parts of the sme tool to be tised for
different tasks, the appropriate functional unit ("FU”) of analysis is the
esployable \aiit defined by Khudson () as
"that iMplaswnt se9aent or portion (continuous edge or projection) deemed
appropriate for use in per farming a specific taA, e.g. , cutting,
scraping, perforating, drilling, chipping. The unit ms identified by
deliberate retouch and/or apparent poet-production utilizaticxi
nxsdification, and its boundaries were defined subject to the analyst's own
concept of 'habituad use.'"
For exanple, a hafted pointed bifaoe has three FUs, the tip and two
laterad marine. If only the tip shows traces of use, the bifaoe lias only one
function unit. If the tip and both margins Imve been used, and one margin
eadiibited two kinds of vear, then che bifaoe has four functional units. The
oono^ of employable unit has been used by Odell () , Joslin-Jeslce () ,
axid Lurie () to anadyze large collections of artifacts, as viell as in
Phase III. It should be noted, however, that tool configurations also provide
information that most not be ignored in making functional interpretations.
Tbol mass, shape, and the relationship between enployable units were often
isportant clues to function in our analysis.
Not all tools in Phase III, however, eadiibited edge damage. Mhmn edge
damage was not present, a tentative function was aissigned to shaped, finished
tools on the basis of position of retouch, edge angle (Wilmsem ) , edge
oonfiguration, and, in some, cases artifact mass (Lurie ) . Tentative
function is a usef^ oonoqot for tools that may have been used on soft
materials, or for tools used very briefly, sinre they develop wear slowly or
not at all. For the purpose of analysis, it was important to separate these
tools frem those dioiidng actual traces of wear. These tentative functions
iimre labeled "potential" in the recording scheme. Althou^ the assignment of
tool function baaed on morphology alone was a questionable procedure, the
criteria for making these assignments was as explicit as possible.
It is often difficult to distinguish roiczofla)dng due to menofacture frem
use-wear especiadly on edge-retouched pieces (Keeley : Brink ) .
Shearing, a manufacturing technique used to baede tools or to prepare edges,
can also be interpreted as use-wear. Similarly it is difficult to separate
use-related microflaking on unmodified edges from unintentional "wear" caused
by prehistoric tranpling, recent trowel or shovel activity, and storage
procedures. The manufacturing and use eoperiments conducted as part of this
project provided inforroation on severed of these problems. This infoonation
and other more arbitreury conventions were used to distinguish use from other
types of edge alterations. The following oonventiens were used:
1. Flaking was considered retouch rather than use if the flate scars extend
more than 1 ran onto the surface of a tool. Asswiixag that retouch could be
more regular than use-wear, edges with flake scars of mixed size were
considered use-wear, eqracially if the majority of the flake scars were
less than 1 mm.
2. Sheared edges with flake scars greater than 1 nm in length were considered
retouched edges. If there was no further evidence of edge alteration,
these edges were considered unused unless they were opposite other used
edges. If they were located opposite used edges, the sheared retouch
edges were scored as possible tool backing. Sheared edges with flake
scars less than 1 ran in length were considered use-wear of some kind.
3. An attaipt was made to distinguish manu&Krturing edge ahcasicn £ian
use-related edge damage. During maxBifaK±uxe of bifaces edges ware often
sheared or ground to straigthen platfoces. Ihe edge produoad was similar
to one with use-wear that might had developed during cutting or sawing
abrasive material. If alaraded edges were due to edge strengthening, it
oculd be expected that the abraded edge %«ould be less localired on pieces
that were not highly refined. On refined tools they could be broken by
flake scars subsequently removed from the prepared edge, or they could be
restricted to specific areas such as haifting elements or blade shoulders.
4. Differenoes in patination and flake scar angle were used to identify
recent edge ^dteration caused by excavation techniques and bag wear. A
shovel or trowel that scr2pes or chunks into a tool edge produces
microfledcing, but these flake scars are often triangular with terminations
perpendicular to the tool edge. Fresh flake scars will be Hotter in
color than the rest of the tool if it was made from some variety of Camden
chert and darker if the tool was made fron blue-gray Fort Payne. Bag wear
was chcuracterized by small. Isolated nicks or flake sc2u:s along an edge.
Flake scars are lifter or darker than the rest of the tool depending on
the type of raw material, frcm vhich the tool was made. Neither of these
kinds of edge damage produce smoothed or rounded edges.
5. Most utilized flakes in the sample hetd ^aignentary functional units. They
were recorded as used, but function unknown.
All pieces were examined using a stereoscopic microscope (American Optical
or Bausch and Lamb Stereo Zocm) at medium power (10-70 power) . When edge
damage was encountered, a sketch of the piece vas made, and the auea of use
noted. In the recording scheme there were four variables allotted to function
(FUCnON, FlWCnCN 2, FUNCTION 3, and RBCIOH) FONCHCN) . Ihe variable
attributes were the same for each of these. The nuncber of functional \inits
having a specified function was also reccrded. In this method of recording
there was no need for oonbined use categoxdes.
For the purposes of this study low-edge angles were 0-45®, medium-edge
angles 46-75®, and steqj-edge angles 76® or greater. The angles measured were
the spine angle (Tring^ian et al. ) or preparation angle (Lurie ) .
Measurements were taken with a goniometer. Edge configuration refers to the
shape of an edge in plan view and the shape of an edge in side view. Hard
materials were stone, antler, and bone. Medium materials were wood, cane, and
other veget£±>le matter. Soft materials were meat, fre^ hide, and sinew.
Grasses and dry hide were considered s^>arately since wear produced by these
materials seem to be very distinctive.
Cutting/ sawing refers to unidirectionad and bidirectional motions,
reflectively. Generally a cutting or sawing tool will be held at an angle
approximately 90® to the material worked. Motion will be parallel to the tool
edge. Wear beyond the inmediate edge should occur on both faces, but the
extent of facial wear will depend on depth of penetration. Scraping could be
unidirecticnal or bidirectional. Scraping tools were usually held at
a^sproocimately 90® to the material worked. Motion will be perpendicular to the
tool edge. Wear onto a face or faces will depend on vdether the motion was
unidirectional or bidirectional, as well as on the hardness of the material
worked. Planing, whittling, and drawing were unidirectional motions
pezpendiculeu: to the tool edge. Tools were held about 30° to the material
worked. Wear will be greater on one face than the other. Chopping was a
unidirectional motion delivered 45-90® to the material wcnked. Adzing was a
unidirectional motion at an angle \jp to 45° to the worked surface.
284
The descaription of the Midden Mound lithic stud^ iinits has focused cn
stylistic^ technological, and functional a^sects of Ardiaic tools. Tool
mozphology, a ccnbination of all these three aspects, often obscures variation
in aisseBDblages that occurs along one of these nose specific lines of inquiry.
For the sake of ootparison the nunibers and percents of artifacts in each
morphological class are presented by site and time category.
FUNCTIC3NAL CATBQOKEES
01 No use-wear ^parent. The piece is too inocnplete to assign a tentative
function.
02 Potential cutting soft material. Low angled enployable units that are
strai^t to sinuous in side view. Ihese edges or edge segments are
tisuedly produced by bifacial retouch.
03 Potential cutting/ sawing medivsn to hard material. Employable units with
medium-edge angles, strciight edges (side view) , zund serrations or
denticulations (plan view) fall into this categpry. Since it was expected
that wear would develop quickly on hard to medium materials, few tools
appecu: in this category unless a cache of finished but unused tools is
found.
04 Potential piercing soft material. Employable units are thin points or
projections that seem to be intentionally produced. Edges of the
projections may or may not be retouched.
05 Potential piercing/drilling/graving mediixn to hard material. Employable
units are thick points or projections, including burin lype bits.
06 Potential scraping soft materied. Employable units have medium to steep
edge angles, rounded edge configurations in plan view so that tlie material
worked will not be cut or nicked and stjaight edges in side view so that
there is maximum contact with the material to be vxmrked. Edges may be
unifacially of bifacially worked.
07 Potential Scrzping/Planing hard material. Employable units are steep and
straight in both plan and side view. They are usually unifacially shaped
and edge retouch is also unifacial.
08 Potential Chopping. In this case functicxi is based on edgre angle,
oonfiguration, artifact sh^se and size. Tools of this sort are relatively
large with strai^t edges and medivsn to ste^ edge angles. They may be
hand-held or hafted.
09 Possible backed edge. See convention 2 above.
10 Used edge. Type of use can not be determined. Broken used edges often
fall into this category, since patterning of flake scars can not be
determined.
11 Cutting/ sawing soft material. Employable units have low edge angles with
stred^t to sinuous edges. Edges are li^tly anoothed or rounded.
MicroflzOces When present on an unretcuched edge would have feather
terminations, smioothing extends both faces of the artifact, both
within flake scaurs and on flake scar ridges.
12 Cutting/ sawing medium material. Bnployehle voiits will have smoothed over
or rounded edges. Step fracture due to manufacture are often smoothed
over, ffeeur will extend onto both faces of the tool, however, since the
material worked is not soft, wear on the faces is restricted to flake scar
ridges and is not as inveisive onto the face of the tool as in cutting soft
material. It is possible that striations will zdso be present. If so,
they will be paradlel to the working edge of the tool. These edges
usually have medium edge angles and strzdght edge configurations in plan
and side view.
13 Cutting/ sawing hard material. Eknployable xsiits have crushed edges wmiA* up
of SBull step fractures. These edges usually have msditxn edge angles and
strai^it edge ocxifigurations in plan and side view.
14 Cutting/sawing mediuD to hard material. Mear indicates use cn sonething
other than soft material, but the t^pe of material is in question.
J5 Perforartdng soft to medium material. Projecticns exhibit smoothing,
rounding, or polish.
16 Drilling medium material. Projections used to drill medium iwH-Ariai show
blunting, rounding, and polishing of edges. Step freKrtures seldom occur.
Striations are perpendicular to the long axis of the projection, and
uneven beveling indicates direction of motion. Drilling projections euce
usually thicker than perforators. The cross-section is often
diaroond-sh£Q)ed. Edges of the projections are bifacially worked to attain
this shs^.
17 Drilling hard material. Projections of employable units have
configurations as those in 16, but wear includes crushing and step
fractures along the sides of the projection, especially on the tip.
18 Drilling material unspecified. Wear patterns include rounding, polishing,
cnishing, and st^ fractures.
19 Scraping/planing soft material. Wear on enployable units is primarily
edge rounding, anoothing, or dull polishing. Striations occur if dirt or
grit exists on the surface of the soft material. Wear extending onto the
face of the tool will most likely be unifacial. If the unit has been
retouched, wear extending onto the face will be on the flake scar surfaces
as well as ridges.
20 Scrsping/planing medium materials. Employable units have shallow,
uniform, unifacial microflake scars. The flake scars have feather
terminations. Edges are slightly rounded or amoothed.
21 Scraping/planing hard materials. Employable units have crushed and
blunted e^s. Step terminations are frequent on edge damage flake scars.
Striations, if present, will be perpendicular to the working edge.
22 Scraping hard material - dry hide. Scraping dried hides produoes very
distinctive wear. The scraping edge or unit quickly beoonss very rounded
and polished. Wear extends onto the face in contact with the hide.
Ridges of retouch flake scars become rounded and polished as well.
Microflaking does not seem to occur.
23 Scraping material undetermined.
24 Chopping/pounding soft/medium materials. A variety of morphological forms
can exhibit chopping or pounding wear. E^qerimantal woodc (Boydstcxi and
Lurie ; Dodd ) indicates that wear is usually in the foam of
rounding, smoothing, and polishing rather than crushing, unless the
material chopped/pounded is hard. Edge angles are steep. The %hole tool
will tend to be large, since weight is an inportant attribute in
performing these ac:tivities.
25 Chopping/pounding hard material - Same tool cxxifiguraticxi as above, but
wear incluides crushing, large flake sc:ars with step termination, and
fac^eting on surfaces.
26 Digging - "These are large core tools with an elongated flattened form
manufactured by free-hand percussion to foam an acutely angled working
edge on at least one end. Use-wear (oonsists of pronounc^ed rounding and
smoothing and edge perpendicular striations. Polish is sometimes present.
Wear traces cxntinue up the edge and well onto adjoining tool faces. This
type of wear indicates penetration into a soft material, presumably soil,
with a chopping work actlcxi motion" (Abler :35) .
286
27 Slotting/graving/grooviiig - Baployable tsiits axe projecticns, often with
edges and fairly thick croes-aectlons. Edges, if present, axe short and
have madiun to steep edge angles. Itese kinds of tools axe subject to
heatvy attrition. The projections or edges do not becone rounded, but
oontinually xesdiaxpen thenselves. Tipe axe crushed and have flake scars
with st^ teminations. The sides of the projections will be rounded,
blunted, or often polished.
28 wedging - Ttools used for wedging often have triangular or lenticular
cross-sections and medium edge angles. Opposing edges will have crushing
as a result of use: one end crushed by a haniner or billet, the other fron
contact with the subject material.
29 TOol backing - Bnployable units are often sheared edges. Wear includes
anoothing or polishing as the result of contact with the hand.
30 Raw material svpply - Cobbles, cobble fragnents or multi sided pieces
without flake ventral surfaces. They eachibit at least one relatively
large negative flalce scar or more than four analler negative scars (Abler
a:506-508) .
31 Pignent source - fiiployable vinit, usually the surface of an iron ore is
abraded and sometimes striated.
32 Practice pieces - "These tools are small and thick, unpattemed objects
modified by bif2cial or unifacial retouch (visually percussion) around most
or all of the tool perimeter. These tools lack wear, but eodiibit edge
crushing and step flaking from percussion flaking. These items are
hypothesized to be nonpuxposely flaked stone made by idle flint knappers
or to be the results of learning efforts )::y ineaqperienoed knappers" (Abler
and Christenson :134) .
33 Other nonutilitarian items - Iseads, pendants, etc.
34 Abrading (a) is characterized by the presence of flat or concave surfaces.
Tbols may be hand-held or stationary. This distinction is made under
morphological t^pe, but since most of the midden mound materials axe
fragnentary, this distinction was difficult to make, and most items were
categorized as ground stone fragtents. The material abraded is not
specified here, but if traces of material, such as hematite, were detected
on surfaces, th^ were noted in the written oannents.
35 Abrading (b) is characterized by abraded grooves on tool surfaces. These
tools have been used for sheqping/diaxpening wood or bone objects.
Abraded grooves are often found in conjunction with edaraded surfaces.
Therefore, these tools were often recorded emt multipurpose ground stone
tools under morphology.
36 Anvilstones are characterized by pecked pits and small linear grooves.
They may be hand-held of stationary. A preliminary survey of the midden
mound material indicates that most were either hand-held or too
fragmentary to be labeled. Both hand-held and station£u:y anvils are
called anvils morphologically.
37 "Nutting stones” are characterized by anoothed pits on tool surfaces.
These pits originally may have been pecked, but use or intentional
modification have created the anooth pit.
38 Inpact fractures - possible projectile.
HAFTING; The presence, possibility, or absence of a hafting element has been
scored for all items. Bifaoe hafting element morphology is most consistently
used to place items into stylistic categories >^ch may represent prehistoric
"cultures" or may be sensitive time markeirs. Hafting element configurations
may also be related to function. Seme farms of notching, base shape, or edge
tzaatamt illicit allov the object to be iirbedeled and aecured to a shaft or
foiDedMift in vniys that %iould restrict slipping or twisting during use. Ihis
would be afplicable to other technological classes as wall as bifaoes.
Fbr thM itesB with hafting elanents, several qualitative and
quantitative variables were recorded. HAFT TYPE records the configuration of
the entire hafting elenent. This variable is based on the presenoe and
location of notches. Attribute states for HAFT TXFG included lanceolate or
shallow side notched, deep side notched, comer notched, sbemned, and basal
notched varieties. Unusual haft ocnfigurations not listed above were scored
as "other.” When hafting elenents were inocnplete their configuration was
usually scored as "can't determine."
HAFT BASE (XKtFIGJBKnctl was also recorded for all hafted pieces. This
variable refers to the shape of the haft element basal margin. Base
configurations %iiere characterized as convex, cxnoave, strai^t, bifurcate,
pointi^, or rounded. HAFT SIDE OR BASE TREADCMT referred to marginal
modificKttions of the haft element sides and/or base. The modifications, if
present, are in the form of beveling, thinning or fluting, grindiig, or
cxnbinations of these.
Four quantitative variables %iiere recxsrded for hafting elements:
HAFT NECK WIDTH, HAFT BASE WIDTH, HAFT ISKIIH, and NECK ANOE. The first
three of these variables were recxgded during Phase I and II. When
measurements fbr these three variables were not available for artifacts in
Phase III analysis, measurements were taken using the following Phase I
definitions.
a. Neck vddth (juncture width) - distance cxi the ocsronal plane between
junctures. The cxoronal plane was "the plane thich includes the margin”
junctures were "the paired, most distal points on the edge beginning at
the most medial vertices not on the basal plane and ending at the next
vertex on the margin moving initially tcward the prcsciroal end of the
midline."
b. Base width - "distance cxi the cxsronal plane between ends of the base."
c. Length *■ "Perpendicular distance between the basal plane and the
transverse plane \duch incliides the jizictures."
Johnson () found that haft angle was an important variable for
distinguishing ^Bong types of stenrasd bifaces. He defined haft angle "the
angle formed by the intersection of lines drawn to bisect the angles defined
hy the notches tdiich form the haft." (p. 144) (Figure 8.27). Following
Johnson's oonvenlions if only one side of the base was present, "symmetry was
aissaned and the angle at vdilch the line bisecting the one remaining notch
intersects the longitudinal axis was measured and doubled to give an
approocimation of the haft angle." (p. 144)
To record this measurement in a standardized way, hafted tools were
xeroxed and angles drawn on these outlines. The line bisecting an zuigle was
drawn using the midpoint of the line connecting the blade shoulder (tang or
beurb) and the comer of the base, and the apex of the angle. Futato (personad
odmunication ) has pointed out that notch angles divided in this menner
will vary with shoulder width and base length and, therefore, the haft angle
measurement will vary. Differerxss in the notch angle division will be most
pronounoed when variations in shoulder width and baae length are great. Notch
angles for a random sanple of ten tools in this stufy were bisected with a
{»x>tractor. In all cases the haft angle measured using these angles were
within two degrees of the original measurements. It seems that dxxilder
288
widdi/baae length differencses do not appKeclable effect haft angles in these
specdnens. Given that ihe original dnwing of notch angles axe scmekAiat
sid^jective, these sli^t differences in haft angle neasurenents are protiably
not significant, and the term HAPT ANGU! vdll be retained for this variable.
MOBPwnn'nGy; Morphological classes have long been establidied for chipped and
ground stche tools. They are usually based on a ocefcination of formal,
functional, and technologiced. attribkes. This combination of attributes
nakes the relationdtiip of morphological classes to hunen bdmvior very
difficult to interpret (Abler vol. 4) . In addition, these classes axe
seldom standardized, nor axe they easy to standardize from one project or one
gwngrwphifiai eunea to another (Lurie ) . The edm in the lithic stud^ was to
separate those variables that can be linked to technological, functional, or
social aspects of behavior and to treat e2K:h of these a^>ects as separate, yet
related, factors influencing the ciiaracter of lithic assemblages.
During Chase I and II, 205 morphological classes viere defined - 67 kinds
of projectile point/knives, 15 kinds of bifaoe blades, eight kinds of
preforms, 14 Idnds of drills or perforators, 15 kinds of cores, 28 kinds of
scrapers, 21 kinds of other unifacial zmd bifacial tools, and 37 kinds of
ground stone tools. For this study, a smaller more manageable set of
morphological categories were established to enable ocnparisons with other
sites in the Tcnblgbee Valley vM.ch have been reported using a traditional
tool typology and with Phase I and II assemblages that were not part of this
study. Definitions of the original 205 morphological categories axe found in
i^jpendix I. Definitions retained for Phase III are noted below with an
cisterisk.
Of the 67 kinds of projectile point/knives only the Archaic point types
are given their own mor^logical classes. These are Greenbariar/Dalton, Kirk,
Big Sandy, Bva/Morrow Mountain, Cypress Credc, Sykes-4(hite parings, Benton,
and Little Bear Creek/Flint Creek (Figures 54-56) . All other named projectile
point/knives were placed in a general, other group, and the type name was
recxjrded as a written comment. The assignment of projectile point/knives to
these categories was retained from Phase I and analysis unless obvious
typological or ccnputer errors were evident. In these cases. Phase I point
type definitions were used for reassignment to a new norphol^ical type.
There are two categories for finished bifaoes other titan projectile
point/knives. The category bifaoe with lateral working edge pertains to
unhafted, refined, synmetrical bifaoes that show tratoes of wear along the
lateral margins. The category bifaoe with transverse wotidng edge includes
hafted and unhafted bifaoe scrapers. Bifaoes with lateral working edges are
often analogous to the Phase I and II category 'hifaoe blades.” The
distincrtlon between biface blade shape and technology used to subdivide this
category during Phase I and II was recorded under (ZCMs'I'kIC SHAPE and
TECHNOLOGICAL CLASS. There is one category for preforms. Frcm a
technological perspective these represent items that seansd to be unfinished
bifaoes. Their degree of refinement was recorded under PRXXJCTION STAGE.
There is also only cne category for cores. Cores are sources of raw
material. They ney be cobbles, peurts of cobbles, or nultifacial chunks
without cortex, vbich etiubit at least one relatively large flake scar or four
or more small flake scars. The distinction between free-hand and bipolar was
made under TECHNDLCX3ICAL CLASS. A ccnbination category of oore-haRmer is also
included for those cores that were modified through battering as well as
flaking. This battering was distinct from any core platform preparation.
There are three classes of uni facial tools: end acrapers, side scrapers
and a ocwbination end-end-side scraper (Figure 57) . Ih^ categories contain
all tools that are intentionally modified on one face, not sj]i{>ly edong an
edge. The distinction between end and side was made in relation to the flake
bulbar axis or the long axis of the tool if ix> bulbar surface exists. One
category each wes allotted to the following morphological types:
a. Drills - Drills are long, narrow tools that were usually bifacially worked
with a diancndr-sh^ied or triangular cross-sections. They may or may not
hetve hafting elements.
b. Perforators/Gravers - These are artifacts that exhibit short projections.
These projections are often formed by unifacial retouch although the
configiiration of the piece as a vhole could be bifacial, unifacial, or
retouched flake or chunk.
c. Burins - Burins are tools vdth bits formed by burin blows, blows struck
approximately at ri^t angles to a flat, relatively narrow surface.
d. Wedges - These are tools with relatively steep, transverse woidung edges.
Opposite edges often both show battering and crushing: one end crushed by
a haniner or billet, the other from contact with the subject material .
Cross-section deperxied on the configuration of the tool as a whole.
Bifaoe cross-sections are often lenticular. Flake or chunk cross-sections
are usually triangular.
e. Denticulates - are tools, xasually flakes, whidi have repeated, contiguous
notches separated by rid^s along an edge or edges. These notches are
larger and wider than serrations.
f. Spokeshaves - These artifacts possess one or more a steeply retouched
ooncavities or notches. Multiple notches are not contiguous. Spdceshaves
are usually made on flakes.
g. Choppers - Choppers are large tools with one or more steeply flaked
working edges. Edges could be unifacial or bifacial.
All utilized and edge-retouched pieces were placed in one category in
terms of morphological class, since without careful inspection of microflaking
size and pattern, it was not possible to distinguish between use and retouch.
For artifacts analyzed during Ihase III distinctions between retouched and
utilized modification, and among retouched or used flakes, bifacial thinning
flakes, bl2kdes and chunks was made under TTOBlQiLOGICKL CIASS.
A category for other chipped stone was included in the recording scheme
for any shaped artifact that did not fit into one of the above categories. A
written oonment was mde describing any such item. The final chipped stone
category was for fragments that can not be assigned to any other categcmy.
These could be sorted into bifacled, unifacial, and other fragments using the
variable TECHNOLOGICAL CLASS.
Morphological categories for ground stcxie can be divided into two groi^:
shaped ground stcxie and use-modified only ground stone. Sheeted categories
included the fbllcwing:
a. Pestle - These artifacts are relatively large grinding tools with a
distinctive bell-she^ed plan view. The grinding surfaces are at the
Heuoxw end, and the sides are smooth.
b. Axe/adze - A tool with a broad transverse bit and grooves for hafting.
c. Celt - Greenstone, hi^ly polished tool with a lenticular cross-section
and a biconvex transverse bit. The bit is tisually apposite a tapered poll
or butt section.
d. Atlatl wig^ - These are ground and poliehad itans with a centred hole to
facilitate hefting, various geoastric ehifies occur.
e. Bsad/bsad prsfoam - These are circular, tabular or aooMDcphic-atepad
objects ti^ have been ground and poliabad. Finished Itens have drilled
perforations presuBBkdy for attaclMent.
£. Arl - Avis are slender slivers of petrified wood that have at least one
pointed end that shows rounding and snoothing.
g. Drilled stone core *- These are plug-shaped pieces removed from pieces of
stone by drilling.
h. Other ^laped ground stone - Any intentionally shaped piece of ground stone
that did not fit into one of the above categories.
Ground stone items modified by use were £mslgned to categories based on
the type of use modification. They were the following:
a. Hanmerstcne - Tbols vhich have one or more localized areas of battering or
crushing on edges or surfaces.
b. Anvil - Pieces of stone that have been pecked or battered to form
irregular depressions, anall linear grooves, or pits on a tabular surface.
c. Ahrader (hand-held and stationary) - Ahraders show localized 2u?eas of
grinding and anoothing. Abrasion can take the form of flat or concave
surfaces or grooves.
d. Ground hematite/liinonite - These are pieoes of iron ocnpounds iifhich
exhibit abraded, grooved, or polished surfaoee.
e. Other use-relat^ ground stone - Any item thought to be modified by use
that does not fit into one of the above categories.
f . Multi-use ground stone - The tools show more than one khxi of non-flaking
modification. These different modifications are recorded under tool
function.
As with chipped stone, there is a category for ground stone fragments that
can not be placed into any of the above categories.
Upon reexamination, some material sorted as tools during Phase I azxi II
were scored as debitage. These items were scored as not applicable for
MORPHOLOGICAL CIASS.
Metric measurements for length, width, thickness, and wei^t were recorded
for moet of the whole, fonnal tools during Phase I and II. During Phase III
these measurements were recorded >dien possible for all tools except utilized
flakes less than one inch in maximum dimension. Conventions established
during Phase I and II were maintained. Measurements are recorded in
millimeters.
For synmetrical tools IfMGIH was the maxinun perpendicular distanoe
between transverse planes tangent to the tool. WIDIH was the maximum
perpendicular distanoe between planes parallel to the longitudinal plane and
tangent to paired points on the tool. THICKHESS was the maximum perpendicular
distance between planes paurallel to the coronal and tangent to paired points
on the tool.
Tools %dilch were not syrnnetriced, but did show flake morphology, were
oriented with the bulbar axis. In these cases LBI3IH was the iwixiiiitm
perpendicular distanoe between two planes perallel to the bulbar axis and
tangent to the artifact. WIDTH was the maxiaun perpendicular distanoe between
two planes parallel to the bulbar axis and tangent to the tool. THICKNESS was
the waxiiam perpendicular distanoe between two planae parallel to the bulbar
axis, perpendicular to the planes defining the maximm width, and tangent to
the tool.
Tools vdthout symnetxy vAiich showed no original flake blank orientation
vnre neasured aoootding to absolute dimensicns. laioro was the ihmH—b
perpendiailar distance between two paral lei planes tangent to the tool. WIDTH
was ihe tnaximam perpendicular distanoe between two parallel planes tangent to
the tool and perpendicular to the planes defining the length. THiaaiBSS was
the perpendiciilar distanoe between two parallel planes tangent to the tool aixi
perpendicular to both the planes defining the length and those defining the
width.
Artifact wei^t is recorded for all items, including broken ones on \duch
no length, width, or thickness measurements were possible. This provided at
least an estimation of artifact size, wei^ts are recorded to the tenth of a
gram.
Length, width, and thickness were not taken for utilized flakes less than
one inch in maxinun length. These were placed into size-grade categories
recorded during Phase I and II. The two size categories were 1/2 inch and 1/4
inch.
Ccnments on various aqects of tool raw material, function, fracture,
h2ifting, and norphology were necessary for some tools and these were coded by
the l^pe of oonmant and entered into the ocaputer. TWo of these oomnent types
were of particular importanae. First wore items that were quBstionable
artifacts. These can be eliminated fian tabulation if necessary. Second,
refits vmre coded so that the broken parts can be separately or as
vhole items. If refit items were recovered from the same level, cne of the
refit pieces was coded for xae in statisticed tests. This piece has \d)atever
artifact dimensicns were taken for the refit item. If refit pieces were found
in different levels but within the same time category, the same distinction
vtas made. If refits were from different time categories, each refit piece was
considered a sepzurate artifact.
SECTION 5; DRTA ANALYSIS RESULTS
The analysis section of this chapter contains three parts which deal with
the descriptive and explanatory questions outlined in section one. The first
part addresses the question of hafted blfaoe classification, the next part
describes the lithic assemblages for zdl the sites studied. Particular
attention will be paid to similarities and differences in technology, tool
function, and tool and ddsitage diqx^sal. Finally, the e9q)ectations for
lithic assenblages under different mobility strategies will be examined.
HAFTED BIFM2E PATTERIS AND OCSBEEATIONS
Classification of hafted bifaoes has been approached in two w^s in this
stucfy. The first has been to perform a cluster an£d.y8is using the following
four continuous variables: NECK WIDTH, HAFT BASE WIDTH, HAFT IfltiGIH, and HAFT
ANG[£. The second treats discrete variable that describe the hafting elament
(HAFT TYPE, HAFT BASE GQNFIGUBATION, and HAFT SIEC OR BASE TEEATSBIT) and the
blade segment of the bifaoe vdien it is present (CROSS SECTION, PRESSURE
FLAKING OR RESHARPOmiG PATTERf, and fUNCTICN) .
A total of 401 hafted tools were inclixled in the lithic study. Of these,
172 belong to chronologically sensitive named types described in Phase I. One
hundred and twenty seven are classified "other projectile point/knives." This
gr(x:p incliides triangular bifaoes without distinct hafting elements, named
t^pes that are represented in our saople by only one or two itans, and
residual stained pieces that could not be assigned to any other named
category. Sixty-seven qaecimens had haft element fraigpnents too inoGnplete to
clasi^fy further. The remaining 35 hafted specimens included ^ following
morphological classes: bifaoe with lateral worlcing edges (6) , drills and
perforatcaos (14) , unifacial and bifacial scrapers (9) , choppers (2) , retouched
flake (1) , and "other” tools (3) .
Of the 172 named ^)ecimens, 121 had all four continuous variables
recorded. Heasuremmnts on haft neck, base width, and haft length are
available frcm Phase I and II data sets. Hzift angle was measured during this
stucfy as a oontinuous variable reflecting haft element notching position. A
cluster analysis was performed on these 121 items to see if haft element
measurements could produce time sensitive clusters and to identic any
conespondenoe between these objectively formed clusters and more intuitive,
named bifaoe types. A SAS clustering prooedore employing Ward's methods vdiich
minimizes the error son of squares was used (SAS Institute, Inc.
:423-432) . A 0.95 value fen: R-ac]uared was used to determine the most
useful nunber of clusters. The results of a nine-cluster solution are
presented in Table 11 of Appendix III.
The procedure did not produce time-smisitive clusters. Only Clusters 7,
8, and 9 reflect restricted time periods: Cluster 7 the Middle Archaic 1, and
Clusters 8 and 9 the Early Archaic. Clusters 4, 5, and 6 contain members
primarily from the Early Azchaic/Mlddle Archaic 1 and frcm the Middle
Archaic 4. Clusters 1 and 3 contain items &cm edl time categories. Cluster
1 contains members from the Middle Archaic 1 throu^ the Late Azrhaic,
althou^ most are from the Middle Archaic 1.
Conparison of the clusters and traditionally heifted bifaoe types indicates
that the greatest correspondence is between Greenbrier /Dalton (lanceolate
bifaces) and Climters 8 and 9 (Table 100) . All but one of the
Greenbrier /Dedton bifaoes occurs in these clusters, while the clusters contain
only two items vhich have not been classified as Greehbriar /Dalton. The
majority of Kirk bifaces fall into Clusters 4 and 5, but these clusters cLLso
contain 28% of the S^ces-White brings and Benton points. Eva/Morrow
Mountain, and cypress Credc points appear primarily in Clusters 1 and 2, and
in 6 and 7. While Clusters 6 and 7 have only one other member, Clusters 1 and
2 contain over 50% of the Benton points. Most of the Sykes-White Springs
bifaoes cue found in Clusters 4 and 5, along with Bentons emd Kirks. Bentons
occur in six clusters (1 throuc^ 5) , and Little Bear Creek bifaoes occur in
Clusters 2, 3, and 4.
TABLE 100
Oanparison of clusters and named bdfaoe types.
Type
Frequency
Cluster
1
2
3
4
5
6
7
8
9
Greenbrlar
9
1
4
4
Kirk
14
1
2
5
6
Big Sandy
1
1
Eva
13
6
2
1
4
Morrow Mountedn
18
10
1
1
3
3
Cypress Creek
12
6
3
2
1
Sykes-White Springs
15
1
5
5
3
1
Benton
35
6
13
9
3
3
1
Little Bear Creek
4
2
1
1
The clusters appear to differentiate between shallow side notching (or
lanceolate) and deep side notching (Cluster 8 and 9) , and basal notching or
293
oontxacting stenned bifaoes (Clusters 6 and 7) , but do not separate aoanaer-
notchsd bifaoes from other ncn-ocxttracting steawed itens. There are several
possible reaisons for this. First, of the four continuous iBaasureBents, heift
angle vAiich reflects notching position is the most susceptible to measurement
einror. Second, the standardized method of measuring heift angle is
particularly sensitive to variations in notching position by missing
barbs, shoulders, or comers from bases and blades. Conventions established
for recording heift angle may obscure variability that can be assessed hy some
other discrete, edbeit more intuitive, variables. Third, the aaymnetry of
many of the bifaoes will also effect haft angle measurement.
Since time categories were assigned to block and level units partially on
the basis of svpposed chronologically smisitive bifaoe types, it was expected
that these types in the sample would occur within specific time categories.
Cross-tabulation of named bifaoe types time category shows this to be the case
for most of the items in the earli^ time categories (TZhle 101) . However,
the other bifaoe types are present in three or more of the tinm categories.
Althouf^ Morrow Mountain bifaoes are found primarily in the early Middle
Archaic, Bentons and Little Bear Creeks in the latter part of the Middle
Arch2d.c and Late Arohedc, a few of each type are present in other time
periods. These may be mistyped or may represent the manufacture of named
bifaoe types over a broader period of time than expected, or they may reflect
mixing of these specimens after initial deposition. S^ces-¥hite Springs and
Cypress Creek bifaoes have the most even distributions over four i-im^ periods.
These are the most poorly defined, most heterogeneous bifaoe types.
niBUE 101
Qcoee-tahulatinn of and cmtemcy.
Tins Category
Early
Archaic
Archaic
Archaic
Archaic
Late
Archaic
I
II
III
IV
Archaic
Greenbriar
9
8
1
Kirk
14
14
Big Sandy
1
1
Eva
Morrow
13
6
6
Mountain
18
8
6
1
3
Cypress Creek 12
2
6
2
2
SykesA4hlte
^xrings
15
2
6
2
5
Benton
Little Bear
35
1
2
5
26
1
Creek
4
1
1
2
Discrete variables were recorded to help define the named bifaoe types.
Cross-tabulations of these veuriables vhich include haift type (Table 102) , haft
base OQnfiguraticn (Table 103) , haft base and side treatment (Table 104) ,
cnoss-section (Table 105) , eoid pressure flaking and resharpening pattern
(Table 106) have Iseen used to generate the following descripticxis of names
types.
294
Cltw lilMlaMinn of traditinnalV
and haft
Deep
Inair- C3ontra- E>q)an
Lanceo- side Oorner Strai^t vate cting ding
late Notch Notch Stem Stem Stem Stem
Basal
Notch
Green
briar 14
(nFl7)
Kirk
(n=20)
Big Sandy
(rF2)
Eva
(n=12)
Morror
Mountain
(n=20)
Cypress
Cre^
(nFl2)
Sykes ARiite
Springs
(n«21)
Benton
(n=42)
Little Bear
Creek
(n«5)
of haft type and time
for mnrplMilogioaJ
Lanceolate
Deep side
Notched
Comer
Notched
Strai^t
Steamed
Incurvate
Steamed
Contracting
Steamed
Ejqmnding
Steamed
Basal
Notched
Middle
d
Early
Archadc
Archaic
Archaic
Archaic
Late
Archaic
I
II
III
IV
Archadc
17
2
1
1
1
4
1
5
19
4
1
2
3
-J
6
6
7
16
9
2
1
2
11
11
1
4
6
4
9
6
6
28
2
1
1
9
5
1
2
1
of baft oonfigui-atiop
Base ‘Koatinent
Greenbzlar/
concave
1
12
Dalton
strai^t
1
1
1
(n=19)
not aqpplicable
can't determine
Kirk
convex
8
(n=26)
concave
4
strai^t
1
1
4
bifurcate
1
not 2^1icable
can't determine
4
1
Big Sandy
convex
1
(n=3)
conc2tve
1
strai^t
1
Eva
convex
1
(r¥=16)
strai^t
1
3
1
3
pointed
not applicable
1
1
rounded
2
1
Marrow
convex
2
1
2
Mountain
strai^t
6
1
1
(0=17)
pointed
not applicable
1
1
rounded
3
2
1
cypress
convex
1
1
Creek
straight
2
1
1
5
{nFl2)
can't determine
S^ces-White
convex
1
1
Springs
concave
1
(n=21)
straight
4
1
10
not applicable
can't determine
1
1
Benton
convex
1
1 1
(n=21)
concave
1
5
1
1
3
stredght
1
8
1
1
2
4 2
can't determine
1
Little Bear
Creek (n=9)
straight
2
1
1
Base Treatment;
I » no treatment 2
3 - bifacial bevel 4
5 S’ ground only 6
7 - ground and bifacial bevel 8
9 » ground and fluted 10
II = not applicable 12
13 - can't determine
•aOE 105
CtoesHabulatioPS of mocphologioal rlasaeo 1 throu;^ 9 and croeo-oectian.
Cross-section
Plano-
Plano-
Can't
Biconvex
convex
piano
Diamond
Other
determine
N. A. Total
Greenbriar
d
2
1
1
6
1 19
Kirk
4
9
5
4
4 26
Big Sandy
2
1
3
Eva
Morrow
7
5
1
1
2
2 17
Mountain
cypress
9
10
1
3
23
Cre^
S^ces-Mhite
7
2
2
1
12
Springs
8
5
3
1
4
21
Benton
Little Bear
3
30
1
2
46
Creek
2
2
1
4
TMU 106
rums niwiliiliiii of hi face mcsTteilogical class and pESi
More flalriiig or
OF
EF
Seriation
Seriation
OF
BF ON BF
In+
In+
OF
BF
N In
In Down Down
Down
Down
Down
Down
Greenbriar
8
4
1
1
2
Kirk
18 1
7
4
3
1
2
Big Sandy
2 1
1
Eva
Morrow
9 2
4
1
2
Mountain
Cypress
19 4
10
1
3
1
Creek
S^ces-White
9 2
3
2
1
1
Springs
10 1
6
1
2
Benton
32 3
8
10
9
3
Little Bear
Creek
4 1
1
1
1
Iff^*3faclal BP=Bifacial Inp Inward pressure DciiieF4)ciwiiwBrd pressure
1. Greenbriar /Dalton bifaoes have lanoeolate haft elements. Bases are
usually strai^t, thinned, and heavily ground. Blade elements are
bioonvex with bifacial iiMard pre^ure flaking.
== imifacial bevel
= thinned
ground and unifacial bevel
= ground and thinned
= other
= fluted
297
2. Kirk bifaoes have oQmer--notched base elanents. Baaes are thinned ai^
ground. Base configurations be straight, cxncave, or oonveoc. Blexie
cross-section is usually plano-convex, but may also be plano-convex or
plano-plano. Pressure flaking pattern is most frequently bifacial inward,
unifacinl downward, or a combination of the two. the imifacial downward
pressure pattern is usiially interpreted as a resharpening technique.
3. Only three Big Sanc^ points are included in the saople. Ihese are deep
side-notched bifaces with thinned and ground bases. Iheir blades are
plano-convex in cross-section with inward pressure flaking pattern.
4. Eva and Morrow Mountain bifaoes have contracting stesmed or basally
notched haft elements. Bases are usually strai^t, but a great deal of
variation exists in the sample. Over half of the specimens have no
^lecial base treatment. Those that do are thinned, ground, or both
thinned and ground. Blade cross-section is either biconvex of
plano-convex. Those designated "other" have irregular cross-sections.
Pressure flaJdng patterns are for the most part unifacial or bifacial
inward.
Hofnein () considers the Eva and Morrow Mountain points recovered fcan
Spring Cave in middle Tennessee as different forms of the multistage
type, rather than two distinct, chronologicedly sensitive types. Items within
a nultistage l^pe vary in form and function because of their position in the
reduction and use-life sequences of that ^lecific artifact groap. Retipping,
rebaseing, lateral resharpening, notching variability, and loss of barbs can
modify Eva points to "include forms vhich have traditionally been classified
as Morrow Mountain points . " The Eva and Marrow Mountain bifaoes examined in
this study support this interpretation. The bifaoes occur within the same
time categories, have similar haft element and configuration attributes, and
are manufactured in the same way. In addition to these, the Cypress Creek
bifaoes that are part: of the midden mound assemblage may also fit into this
nultistage type.
5. Sykes-Mhite Springs bifaoes have expanding or strad^it stems with straight
bases. The bases are usuedly thinned and ground. Cross-section is
biconvex, plano-oonvex, or plano-plano in order of frequency. Pressure
flaking pattern is most often bifacial inward.
6. Bentcxi bifaoes have expanding or strai^t stems. Bases are usually
straic^ or convex with either unifacial or bifacial beveling.
Cross-sections are most often plano-plano followed by unifacial inward or
downward and then bifacial inward. These pressure flaking patterns most
likely reflect the use phase of the biface. Blanks and newly finished
piecses have unifacial or bifacial inward pressure flaking. These are the
larger specimens. As these bifaoes were used and edges dulled, unifacial
and then bifacial downward pressure was used to rejuvenate them.
The distinctican between Benton and Sykes-Mhite Springs bifac:es is someshat
problematic:al in the literatiu?e of the mid-South. However, Futato (:124)
defines the Sykes-White Springs cluster represented in the Cedar Creek
drainage of northwestern Alabama as "a broad cxxtinuum of comer renxTved and
steraned forms. The White Springs-like specimens overlap morpholc^gically and
tanporally with Morrow Mountain and are probedsly most uoimon in the earliest
part of the cluster. The more distincrtly steamed forms, mare Sykes-like,
overlfp morphologically and temporally with Benton material and are probedsly
most ccnnian in the later part of the cliaster." The Sjdces-White parings
1
bifaoes in our sanple are distinctly stemnad, but are not usually plano-plano
in cross-section nor do they eidiibit beveling on the hafting element. Ihe
three specimens vidth plano-plano cross-section occur in time category 5, but
Sykes-White Springs bifaoes with biconvex and plano-convex cross-section occur
in that time category also. All but two of the specimens are made from Camden
chert. One is made from fbrt Payne emd one &cm quartzite.
nie predcmin^uloe of plano-plano cross-section, bifacial beveling on the
hafting elements, and unifacial and bifacial downward pressure flaking pattern
on the blades sets off the Benton and ^kes-White Springs bifaoes in the
sanple. In addition, 30 of the Benton bifaoes are made from Fort Payne chert,
13 fron Canden, and two from material that can't be determined. Before time
category 5, six Benton bifaoes are made fron Fort Payne chert, three are of
Camden. During and after time category 5, 24 Bentcm bifaoes are made from
Fort Payne chert, 12 from Camden or some other raw material. The change in
manufacturing techniques necessary to produce more Benton-like bifaoe appears
to occur gradually, over seme period of time, on primarily Icxml raw material.
Hie addition of more Fort Payne chert does not signal a change in the
technique of manufacture but a shift in raw material selecticm. The
manufacture of plano-plano Benton-liJee bifac%s 3:equires large blanks. The
switch to iitfXirted Fort Payne blanks may have been necessary, if cxhbles of a
suitable size cxxild not have been obtained.
7. The small niinber of Little Bear Creek bifaioes in the sample are veuriably
stemnned bifacres with straight bases, but without cx>nsistent base or side
treatment.
Analysis of traditionally named biface types and tcxsl functicxis were made
in an effort to determine any patterns (Tables 107 and 108) . From this it
eppears that the function or funcrtacais for vhich hafted bifac:es were used are
varied and ncTt specific to any named l^pe. Bifaces were used to cat and saw
soft, medium, and hard material. Eight of the iixxmplete bifaoes have ispaert
fractures in^ceting contact: at the distal end with seme hard material. These
fractures are often interpreted cis use as a projeertile that has exme in
centact with bone, wood, or possibly heurd grexmd. All of the inpext freutures
are large. The force required to generate them weis probably cxsnsiderable, but
the possibility that these tools were acxiidentally dropped onto a hard surface
eind broken can ncjt be discounted.
TABU 107
Nimter of different functions reexmded for traditionally named hafted bifaoe
N
No
function
1
function
2
functions
3
functions
Greenbriar/
Dedton
19
9
6
4
0
0
Kirk
26
9
10
7
0
2
Big Sandy
3
0
2
1
0
0
Eva
17
6
8
3
0
1
Morrow
Mountain
23
4
12
7
0
0
Cypress
Creek
12
3
6
3
0
1
299
Hahei of differ— L ftwctin— recceded for txaditio— lly a— ed haflad faLfsoe
typ— (oonH— il>. _
Type
N
No
function
HSiSSi
2
functions
Recycled*
Sykes-White
springs
21
13
4
4
0
1
Benton
46
17
16
13
0
3
Little Bear
Cre^
5
4
1
0
0
1
Other
155
99
36
14
1
5
TMU 108
Hrfted bifeoe tool tmction by ■ocphnlf^cal cl— s. _
One Functicn Nunber TWo Functions Number
ReoQcded of Necxarded of
Function Tools Functions _ Tools
Morphological Class 1 - Greehbriar/Dalton
Unidentified 2
Cutting medium 1
Piercing soft 2
Drilling medium 1
Pot cutting soft/Pot cutting medium
Pot cut soft/Pot piercing soft
Pot cutting/SaMing medi\m and
Pot drilling medi\m/hard
Cutting soft/iirpact hard
Morphologiced Class 2 -- Kirk
Pot cutting soft 2
Unidentified 2
Cutting mediim 2
Cutting medium/
hard 1
Piercing soft 3
Pot cutting soft/medivm
Pot cutting medixm/Pot pierc:^ soft
Pot cutting mediun/unidentified
Pot cutting soft ijqpact hard
Cutting soft/tool backing
Cutting medium/piercing soft
Morphological Class 3 - Big Sanc^
Unidentified 1 Cutting medium/piercing soft
Cutting medium/
hazd 1
1
1
1
1
1
1
1
1
1
2
1
Morphological Class 4 - Eva
Unidentified 2
Cutting medium 2
Cutting medium/
hard 1
Drilling medium 1
impact hard 1
Pot cutting soft/Pot cutting medium
Pot cutting soft/Pot drilling medium
Cutting soft/impact hard
1
1
1
300
Recorded
of
Reoarded
of
Function
Tools
Functions
Tools
Morphologiced Class 5 - Marrow Mountain
Pot cutting/
Saw medium
1
Pot cutting soft/Pot cut/Saw mediun
1
Uhidentified
3
Pot cutting soft/Pot piercing soft
2
Cutting soft
1
Unidentified/unidentified
1
Cut/Saw medium
1
Cutting soft/Pot cut/Saw medium
1
Cut/Saw hard
Cut/Saw mediixn/
1
Piercing soft/pot cut/Saw mediun
1
hard
2
Piercing soft/cutting soft
1
Drilling
1
Morphological Class 6 - Cypress Creek
Pot cutting soft
Pot cut/
1
Pot cutting soft/Pot piercing soft
1
Saw roedlvm
1
Cutting soft/Pot piercing soft
1
Unidentified
1
Cut/Saw mediun/Iiqpact hard
1
Cut/Saw mediixn
Cut/Saw mediun/
2
hard
1
Morphological Class 7 - SykesHNhite Springs
Pot cut/
Saw mediun
3
Pot cutting soft/Pot piercing soft
1
l^dentified
1
Pot cut/Saw mediun/Pot drilling medi\in/hard 2
Cut/Saw medium/
hard
1
Cutting medivm/unidentified
1
Of the 327 hafted bifaoes identified as sene kind of projectile
point/knife for the variable MCM’fKXOG^r 164 had no function assigned to then.
These are broken itens lacking sufficient blade naigins to eissign even a
tentative function. One hundred and one have a single function assigned; of
\ these, 39 are \^le. Fift^-seven have ttro functions recorded; thirty-four of
these are vAiole. From these results, there is no indication that any named
class is specif icedly designed for one peurticular function. The overall ratio
of vAiole tools with one function to whole tools with two non-recycled
functions is 1:14. There is no discernible tendency for any of the named
t^pes to have more multiple uses than other types.
I Thirteen pieces have a recycled function and were considered vhole.
Functions include piercing soft material (1) , drilling medium material (1) ,
graving (6) , screwing medium or imspecified materials (4) , and scraping dry
liide (1) . It should be noted that many of the tools in the morphological
classes of bifacial scrapers and drills are prob^ly recycled bifaoes that
were originally \ised as projectiles or knives and saws. Diagnostic base
I configurations are often modified in the process of rehafting, and traces of
previous functions are often obliterated by later functions.
301
In a further analysis of the patterning of hafting type tfaraui^ ties, a
croas-tabnlation of haft types and tiae Gategory sas pertanaad (Tfeble 103) .
On the noet general level, saee haft types appear to be uaad as an indicator
of chxcnology CM>le 103) , vhether th^ have been assigned to a gl^en naaed
marphologijcal type or not. Lanceolate or ooener-notehed points are found
pretoainately in Early Archaic or early Middle Archaic context. Oontracting
BteBBMd and basally notched bifaces axe noet proadnent in the early Middle
Archaic and Middle. Archaic 1. Straight steaned and expanding steaeaed bifaces
are found most often during the late Middle Archaic, but they are found in
fairly large nuihers in other time categories as sell. In these cases, haft
type alone is not a very reliable indicator of chronology. Of the ten de^
side-notched bifaoes, half axe found in Early Archaic and half in Late Archaic
context. Again, other charcx^eristics are necessary to distinguish which of
the time categories is represented if a deep side-notched biface is found.
mi MRilEREAL SELECTION
. Several questions about lithic raw material were posed in this study,
including questions about the amounts of different kinds of materials us^ for
tools, the quality of raw materials for tools, and the extent of chert heat
treahnent throu^ time. Data from debitage aj^ artifacts can be used to
address these aiqpects of lithic technology. An analysis of these data was
conducted, and the results are presented in Appendix III. The study indicates
that chert was the primary material used for stone tools throu^xxxt the
Archaic (Tables 109, 110, and 111) . The non-chert waterial axe primsrily
ferruginous sandstone and, during the later part of the Archaic, TSallahatta
quartzite. The highest percentage of non-chert debitage is fcx^ in units
from 22It576 (time categories 2 and 3) and 22lt539 (time category 5) . These
units ocntain 12%, 20%, and 13% non-chert lithic materials, respectively.
Overall the amount of non-chert ddoitage in the various asaaablages increases
slightly during the late Middle Archaic and Late Archaic. Far exaaple,
nonrchert debitage increases frcm 1% in the Early Archaic to 13% in the Late
Archaic at 22It539.
The percent of artifaxrts made from non-chert materials at most sites
during noet time pericxis is greater than the percent of ncn-chert debitage.
The d^ltage to tool ratio for non-chert material never exceeds two to one.
It is likely that non-chert tools, such as sandstone abraders, required little
manufacturing which would have produced fla)ce ciebitage, and that artifacrts
such as quartzite bifac^es were made elsemhere £Uid brou^t cxito these sites.
s
221t623 Tine Category 6
Block D 348 95
5 368
22It623 Tine Category 5
Blbdc C
673
94
44
6
717
22It624 Tine Category 6
Block A
1,071
94
74
6
1,145
Block B
949
93
74
7
1,025
Block C
589
95
33
5
622
Total
2,609
94
183
6
2,782
214
90
25
10
239
IMCB 110
CInrt debitage anMOKy by sits.
block, levi^.
tine <
ntegocy.
asd xm
_
Tuscaloosa
other Chert
"*<->‘*'^♦**'■8 1 1 III
2It539 Tine Category 5
Block A 1,111
82
237
18
7
1 1,355
Block B 2,103
86
346
14
1
1 2,449
Total 3,214
85
583
15
8
1 3,804
22It539 Tine Category 4
Block A 1,895
91
181
9
2
1 2,076
22It539 Tine Category 2
Block A 3,756
93
287
7
13
1 4,056
Block D 10,581
97
373
3
8
1 10,962
Total 14,337
95
660
5
21
1 15,018
22It539 Tine Category 1
Block D 4,298
94
264
6
9
1 4,571
22It576 Tine Category 3
Block D 4,322
95
182
4
30
1 4,534
22lt576 Tine Category 2
Block D 455
97
16
3
471
22It576 Tine Catecjory 1
Block D 4,359
99
20
1
4,379
22It590 Tine Category 4
Block B 1,267
95
56
4
1
1,330
304
Canden/Pickwicik/
Block
Tuscadooea
Ft. Payne
Other Chert
Total
Frequency Percent
Frequency
Percent
Frequency
Percent
22It590 Tine Cateqory 2
Block MU 2,811
94
181
6
14
1
3,006
22lt590 Tine Cateqoxy 1
Block fU 3,954
92
312
7
18
1
4,284
Block AY 6,344
92
496
7
25
1
6,875
Block E
503
93
33
6
5
1
541
Block F
950
89
117
11
5
1
1,072
Block J
1,209
96
44
4
0
0
1,253
T.S. 13
317
97
10
3
0
0
328
Total
13,277
93
1,012
7
54
1
14,343
22It621
Tine Category 2
Block A
2,614
97
49
2
26
1
2,689
Block C
2,011
96
68
3
10
1
2,089
Block E
1,269
97
50
4
4
1
1,323
Total
5,894
97
167
3
40
1
6,101
22It621
Tine Category 1
Block A
1,994
97
40
2
18
1
2,052
Block C
1,756
98
23
1
23
1
1,802
Total
3,750
97
63
2
41
1
3,854
22It623
Tine Category 6
Block D
310
89
36
10
2
1
348
22It623
Tine Category 5
Blodc C
614
91
54
8
5
1
673
22It624
Tine Category 6
Block A
895
84
170
16
6
1
1,071
Block B
694
70
254
26
1
1
949
Block C
424
72
163
28
2
1
589
Total
2,013
77
587
23
9
1
2,609
22It624
Time Category 5
Block A
194
92
18
8
1
1
214
ISBIf 111
Artifact raw aetrarial tv tine oateqcacy*.
Tine
Raw 22It539
22lt576 22It590
22It621
22It623/22It624
Cateen^zv Material N %
N
« N
%
N
%
N %
1
C/Y/P 205 82.6
369
95.8 454
77.1
173 94.5
FP/FFP 22 8.9
3
.8 75
12.7
5
2.7
0 Chert 4 1.6
4
1.0 28
4.8
4
2.2
Non-chert 17 6.9
9
2.3 32
5.4
1
.5
Total 248 100
385
99.0 589
100
183
99.9
305
Cafgoiry
Material
22It539 22It576 22It590 22lt621
N %
N
N
N %
N
2
382
88.9
62
81.6
104
82.5
397
85.9
FP/FFP
77
8.1
6
7.9
9
7.1
35
7.6
0 Chert
17
1.8
2
2.6
5
4.0
13
2.8
Non-chert
19
2.0
6
7.9
8
6.3
17
3.7
Total
945
99.9
76
100
126
99.9
462
100
3
c/y/p
652
75.4
FP/FPP
72
8.3
0 Chert
19
2.2
Non-chert
122
14.1
Total
865
100
4
c/y/p
209
67.1
9
52.9
145
70.7
FP/FPP
59
19.0
1
5.9
16
7.8
0 Chert
17
5.5
0
4
2.0
Non-chert
2€
8.4
7
41.2
40
19.5
Total
311
100
17**100
205
100
5
C/y/p
404
55.2
24
55.8
13
76.5
216
78.5
FP/FFP
238
32.5
16
37.2
0
26
9.5
0 Chert
13
1.8
0
0
10
3.6
Non-chert
77
10.5
3
7.0
4
23.5
23
8.4
Total
tST
loo
45**l0(i
17**100
ITT
166
6
C/y/p
326
54.8
FP/FFP
186
31.3
0 Chert
17
2.8
Non-chert
66
11.1
Total
55r
166
* Items recxarded as "can^t deterndne" not includeS^
** Represented by features only.
C/Y/P><:anden/Yellow Chert/PicikMicJc
FP/FFP*=Fort Ps^ne/Fossiliferous Port Payne
O ctertsOther chert
Chert materials can be groi;ped into local cherts and nonlocal cherts.
Local cherts are by far the most ocninon material present in both the detoitage
and tool assemblages from all sites at all time periods. "Other" chert
cWaitage comprises only a small percentage (1% or less) of any of the lithic
assemblages. In all assemblages the percentage of tools made frcsfi "Other"
chert is greater than the percent of debitage. These tools, usually small
bifaoes or bifaoe fragnents, apparently xeere manufactured scmswhere else. Thi
percent of tools made frcm "Other" material ranges from 1% in 22It576 (time
category 1) to 5.5% in 22It539 (time category 4) , but there is no ^tial or
temporal patterning to these small variations. For example, the second
hic^st percent (4.8%) is from 22It590 (time category 1) ; the second lowest
percentages (1.8%) are frcm 22It539 (time categories 2 and 5) .
Fort Payne chert comprises a relatively anall part of the lithic ddbitage
at all of the site in time categories 1, 2, and 3 (e^prcKimately 4%) , and the
percentage of tools is also relatively low (less than 8% overall) . Most of
the Fort Pz^ne chert reocfvered fran the earlier ocnfxxients is fbssilifercus
Fort P^ne rather than the blxie-gre^ Fort Payne fou^ in the later oonponents.
The aoDunt of Fort Payne ddbitage increases substantially in tine
categories 4 and 5, 22It539 (9% and 15%, re^pec±ively) and in tine category 6,
22It623/22lt624 oonbined (21%) . FOrt Payne is 26% and 28% of the chert
ddoitage in Blocks B and C at 22It623. As with the "other” nonlocal raw
naterials, Fort Payne contributes nore sudt»tantially to the tool inventory
than to the debitage. For exanple, while 15% of the debitage in tine category
5, 22lt539 is of Fort Payne chert, 32.5% of the tools are nade frcn\ this
naterial. Likewise, while 21% of the ddoitage in tine category 6,
22It623/22It624 is Fort Payne, 31.3% of the tools are.
At 22It623/22It624 the anount of Fort Payne nay be greatly exaggerated,
though# because nany of the FOrt Payne bifaoe fracpoents probably were parts of
the sane tool. This particular unit contains over a hundb^ed snail fracpnents
of Fort Payne that are very similar in color, texture, and mottling. An
estimated three or four broken bifaoes could have produced these hundred or
nore pieces. The high percentage of Fort Payne chert in the feature material
in tim category 4, 22It576 may also represent fragments of only a few tools.
These two cases not withstanding, the smedl amount of debitage in relation to
the amount of tools at these sites supports the contention that Fort Payne
bifaoe blanks or finished tools were brought into the midden mound sites and
that minimal amounts of knapping were necessary to finish, resharpen, or
rework them.
It should be noted that Fort Payne debitage and tools do not increase
during time period 4, 22It576 and 22It590 or in tine category 5,
22It623/22It624 . The introduction of this material into the Tcmbi^see Valley
appears to vary from site to site in these later time periods.
BM4 MftTEKLAL QUALITy
Most of the chert used by the Archaic inhabitants of these middei mound
sites was of good quality. While a formal assesanent of cdiert quality was not
made during Ihase I and II, an examination of the sasple debitage units shows
that the quality of raw material is generally good, except at Site 22It621.
In the sample units from this site, flakes often have areas of spongy cortex
or rou^ texture, even on small flakes. All artifacts included in this study
were scored for raw material quality. A frequency analysis of these data
revealed that, except for tliose units represented by feature material only,
(Table 112) less than 7% of the chert is poor quality. Even fair quality
chert is less than 10% at all sites, except at 22lt621 and 22It623/22It624.
TMU 112
Tbol quality by cataegocy*.
^me 22It539 2Mt576 22It590 22lt621 22It623/22It624
Category Quality N % N % _ N % _ N % _ N %
1
Poor
Fedr
Good
Total
9 3.8
13 5.5
214 90.7
236 100
1 .3
12 3.2
362 96.5
375 100
8 1.4
50 8.8
508 89.9
566 100
1 .8
44 36.1
77 63.1
122 100
307
2
Poor
7
.8
0
3
2.4
10
3.0
Fair
91
9.8
2
2.8
12
9.6
106
31.5
Good
828
89.4
69
97.2
110
88.0
221
65.5
Total
9?6
100
71
100
125
100
337
100
3
Poor
5
.7
Ffdx
42
5.7
Good
686
93.6
Total
733
100
4
Poor
1
.4
4
40.0
0
Fair
22
8.5
2
20.0
10
6.0
Good
236
91.1
4
40.0
157
94.0
Total
259
100
10**100
167
100
5
Poor
6
1.0
0
1
7.1
Fair
30
4.8
2
4.9
2
14.3
Good
590
94.2
39
95.1
11
78.6
Total 626 100 41**100 14**100
6 Poor
Fair
Good
_ Total _ ^ _ _
* Items scored "can't determine" not included.
** Represented by features only.
2 .5
55 12.5
384 87.0
441 IbO
In order to ocnpare possible changes in the quality of Fort Payne and
local cherts through time, time categories 1, 2, and 3 haive been designated
Early Archaic, time categories 4 and 5 Middle Archedc, and time category 6
Late Aichadc. Cross-tabulations of cdiert types and these collapsed time
categories show tliat the amount of good quality Port Payite chert varies little
through time and variation that occurs is not directional (Table 113) . The
amcunt of good quality local chert decreases through time (Table 114) .
TMU 113
Qroea-'lahnlat.lnn quality of Fort Payne chert by oon*piimi Archaic period.
Frequency
Percent
Rcw % Can't
Column % _ Poor _ Fair Good Determine Total
Early 5
0.59
1.65
71.43
14 269 15
1.65 31.72 1.77
4.62 88.78 4.95
58.33 35.77 23.08
303
35.73
308
Frequency
Percent
Bant %
Can't
Column %
Poor
Fair
Good
Determine
Total
Middle
2
2
324
28
356
0.24
0.24
38.21
3.30
41.98
0.56
0.56
91.01
7.87
28.57
8.33
43.09
43.08
Late
0
8
159
22
189
0.00
0.94
18.75
2.59
22.29
0.00
4.23
84.13
11.64
0.00
33.33
21.14
33.85
Total
7
24
752
65
848
0.83
2.83
88.68
7.67
100.00
Qii-Sguare 23.062 with 6
!»• Prob =
.001
TRK2 114
CYnen laNiIntirjii qoality of local — iterlal by ooUapaed Ardiaic period.
Frequency
Percent
Rcw %
Can't
Colixm %
Poor
Fair
Good
Determine
Total
Early
22
339
2,682
200
3,243
0.48
7.38
58.37
4.35
70.71
0.68
10.43
82.55
6.16
64.71
71.22
72.68
51.68
Middle
11
99
798
110
1,018
0.24
2.15
17.37
2.39
22.20
1.08
9.71
78.24
10.78
32.35
20.80
21.63
28.42
Late
1
38
210
77
326
0.02
0.83
4.57
1.68
7.09
0.31
11.66
64.42
23.62
2.94
7.98
5.69
19.90
Total
34
476
3,690
387
4,587
0.74
10.36
80.30
8.42
100.00
Chi-Square 132.964 with 8
II
.000
Although chi-square statistics show that non-randcm changes in chert
quality have probably occurred in the use of both these materials, the changes
are due in large measure to the amount of chert for vhich quality could not be
determined. Quality of smaller pieces of chert and those that were heat
309
altaeacad are difficult to assess, and in the Late Archaic period the mean
artifact size decreases and the eraount of heat alteration increases. Since
mean artifact sizes decreases and the amount of heat alteration increases in
the Late Arcdiaic, a hi^ier percent of specimens were scored "can't determine
for this variable.
HEAT
AMD HEAT ALTERATION
During Chase I and II the presence or absence of heat treatment was
recorded for Camden and IXiscaloosa or Yellow chert. In the sanple units
selected for this studY, Tuscaloosa chert debitage was found primarily in
22It576. Snail amounts are present in 22It539 and 22It621, and to an even
lesser extent in 22lt590 and 22It623/22It624 (Table 115) . In most cases the
ratio of heated to non-heated Tuscaloosa is less than 1:1 (Table 116) . Heat
treatment of this material seems to have been of moderate iafgrtanoe. The
ratios of heated to unheated Camden chert shows considerable variation among
the size categories and with one exception (22It576, time category 1) , the
ratios increase as the size of the debitage category decreases. This aeemo to
indicate that heat treatment occurred after artifacts had been rou^y shaped
and larger flakes and chunks of material removed. It is possible, however,
that evidence of heat treatment at the earliest stages of manufacture can be
found in the "fire-cracked chert" Introduced Rode category ^diich was not
included in this stud^. Until this emttegory of material is included, only a
tentative assesanent of the stage at vbich heating took place can Ise made.
Given the information available, there ^ipears to be no clear pattern in the
ratio of heated and xubeated Cenden throu^ time, Ixit the highest ratios ajre
in the earlier time categories.
ctaect fay sine, site, and
u
H
U
H
U
H
22It539
T. C. 1
3
9
440
201
2,956
652
0
0
0
0
7
1
T. C. 2
T. C. 4
34
1
25
3
1,649
198
605
87
9,492
1,378
2,328
280
0
0
1
1
3
1
5
6
35
7
18
37
T. C. 5
10
17
381
178
1,868
646
0
0
3
13
36
65
22It576
T. C. 1
14
3
479
124
2,636
423
0
5 •
16
95
96
465
T. C. 2
0
0
27
18
300
89
0
0
2
3
17
19
T. C. 3
9
9
520
252
2,493
971
0
2
14
34
126
184
22lt590
T. C. 1
29
40
1,371
767
8,235
2,914
0
0
0
0
6
4
T. C. 2
5
20
366
215
1,589
601
0
0
0
1
0
7
T. C. 4
2
6
107
67
826
241
0
0
0
0
0
0
22It621
T. C. 1
9
4
408
146
2,478
659
0
0
0
2
0
4
T. C. 2
3
20
533
459
3,345
1,417
0
1
0
2
5
24
22It623
tTcTT 0 1
71 22
410 105
0 0
0
0
1
0
T. C. 6 0 1
12 7
241 47
0 0
0
0
0
0
22lt624
T. C. 5 0 1
28 14
98 46
0 0
0
0
0
0
T. C. 6 4 6
203 77
1,342 352
0 0
0
0
7
7
IMieated IMftiheated T.C.=Tiine category
TMM 116
IMdo of t—twa to onbei
itad Itsoaloosa and Ctiwn debitage lag aits,
tJae catmorv. and aiae.
•
Time
Size
Material Type Site
Category
Tusc2d.oc>sa 22It539
1
7.00
2
.60
1.90
4
.16
.19
5
.23
.53
22It576
1
.16
.21
2
.66
.89
3
.41
.68
Camden 22lt539
1
.33
2.18
4.53
2
1.52
2.72
4.07
4
.33
2.27
4.92
5
.58
2.15
2.89
22It576
1
4.60
3.86
6.23
2
1.50
3.37
3
1.00
2.06
2.56
22It590
1
.73
1.79
2.83
2
.25
1.70
2.60
4
.33
1.59
3.42
22It621
1
2.25
2.79
3.76
2
.16
1.38
2.31
22It623
5
3.22
3.90
6
1.70
5.12
22It624
5
2.00
2.13
6
.66
?.63
3.80
311
Haat traatxMit was recxsrded for all tools examined during Fhaee III (Table
117) . The amthods for scoring this variable pravides a coneervatiwe astimate
of intentional heating of chert raw material (see Section IV) . It tfpBars
that heat tzeatnent decreases considerably during time categories 5 6 when
the use of Fort Fayne chert for tools increases. In the esgmrimantal work,
blue mottled, dark gray Fort P^ne did not react favonAily to heat, and an
estimated 50% of the artifacts in these tine categories are made from this raw
material. It is unlikely that this material was intentionally heated. Luster
contrast was used to score heat treatment on lighter varieties of blue-gray
Fort Fayne. Few of these pieces appear to be heat treated. This lighter
material may not require heating to infaxfve worJcability (Kadin personal
ocmnunication ) or may have been heated and shaped to a stage in which
luster oontreist would not be evident before being brought to the sites.
117
Artifact
Heat
25lt539
22It576
22It590
22It621
22It623/22It624
Category
Treatment N
%
N %
N
%
N
%
M
%
1
Present
142
61.7
246 65.2
349
63.2
139
76.3
Absent
88
38.3
131 34.7
203
36.8
43
23.6
Total
230
100
377 99.9
552
100
182
100
2
Present
540
58.3
42 60^9
78
65.5
287
63.6
Absent
387
41.7
27 39.1
41
34.5
164
36.4
Ttotal
927
100
69 100
119
100
451
100
3
Present
526 71.0
Absent
214 29.0
Total
740 100
4
Present
198
69.7
6 60.0
92
55.4
Absent
86
30.3
4 40.0
74
44.6
Total
284
100
10**100
166
100
5
Present
322
50.5
15 37.5
7
50.0
124
48.5
Absent
326
49.5
25 62.5
7
50.0
132
51.5
Total
658
100
40**100
14**l0b
25r
~jm
6
Present
236
45.6
Absent
292
54.4
Total
tt 1 a-
sir
100
** Represented by features only.
Although an attempt was made to determine the stage of production in v^ch
heat treatment was acconplished, most items had to be scored as "can't
determine" for this variable. C3nly 190 heat-treated pieces oculd be scored
for time of oocurrenoe. Almost adl of the pieces are from the eeurly time
categories (Eeurly Archedc, Middle Archedc 1, and Middle Archedc 2) . Of these
12.6% were heated in the oobble stage, 68.9% were heated in the flake blank
stage, and 18.4% were heated in the unfinished biface stage. These data
indicate that heating took place early, but in gener^d not at the eeurliest
stages of oobble reduction.
312
Heat alteration of chert applies to those pieces with indications of
heating, but vAiich have not been selected for further reduction. The
inplication is that these pieces were not intentionedly heated. Analysis has
indicated that the percent cf heat alteration is fairly ocnsistent through
tine, but it varies from 11-30% at different sites (Table 118) . Heat
cdteration is lowest in frequency (10-13%) during the earlier tine categories
in 22It590 and 22It621, and in 22lt576 during tine category 3. The greatest
enount of heat alteration occurs in 22It623/22It624 during the Late Archaic
(32%) . There is a shift in the nanufacturing stage at which this alteration
takes place. During the later tine categories (4, 5, and 6) nore heat
alteration appears on finished tools and is most likely the result of
unintenticxial heating. The sli^t increase in heat alteration coincides with
the increase in Fort Payne chert at 22It623/22It624.
In sunnary, the great najorit^ of debitage and tools in this stuc^ cure
varieties of chert. Ferruginous sandstone tools required little manufacture,
they are generailly tools be virtue of being used. Quartzite tools were
probably made elsewhere and tirouf^t to the midden mound sites. The chert is
predoninantly local material derived fecn Tuscaloosa gravels. "Other" chert
artifacts, liJce quartzite tools, were prc^>ably made off site. The presence of
Fort Payne chert increases at sene sites during the late Middle Archaic and
Late Archaic. This increase probably represents the inportatic^n of bifac^e
blanks or finished tools into the midden mound area.
While the quality of all chert is generally high, the quality of local
cherts decreases during the Late Archaic. This decrease in quality of Icx^ed
cherts corresponds to the increase in the amount of Fort Payne chert Ircsught
into the area. Heat treatment, the intenticmal heating of raw material,
decreases through tine. Heat treatment ranges frcni a high of over 76% in time
category 1, 22It621 to a low of 37.5% in tine category 5, 22It576. Although
most cherts are good quality, ejqierinental work has indicted that heat
treatment probably makes local cherts easier to knap and pejhaps more
esthetically pleasing. Heat treatment seems to cxxnr after bifat^es or cores
have been rou^ily shaped. Heat alteration, the unintenticmal heating of chert
increases in the Late Archaic. Both the decrease in heat treatment and the
increase in heat alteration may reflect the pc»r tolerance of Port Payne chert
to heat. Slnc3e heat alteration in the later time pericxis cxxxirs primarily cm
finished tools, it may reflect longer ocxnipaticjn, and therefore, greater
chanems for unintentional heating, at csrtedn sites.
TME£ 118
Artifact heat alteration by tine cabeqaty*. _ _ _
Tine
Heat
22It539
22It576
22lt590
22lt621
22It623/22It624
Alteration N %
N
%
N
%
N
%
N %
1
Absent
176 74.3
292
75.4
517
87.2
164
89.6
Present
Cob-cxDre
1 .4
3
.8
3
.5
0
F.B.
0
8
2.1
10
1.7
0
Biface
1 .4
3
.8
11
1.9
3
1.6
Finished
2 .8
6
1.5
12
2.0
14
7.7
C.D.
57 24.0
75
19.4
40
6.7
2
1.1
Total
237 99.9
387
100
593
100
183
100
118
jaetlAct h—tt alteration by tiae catagory* (ooBtijanadO .
Tine
Heat
22It539
22lt576
22It590
22It621
22It623/22It624
Cateqory Alteration N
%
N
%
N %
N
%
N
%
2
Absent
Present
757
72.7
57
76.0
112 86.2
405
87.7
Cob-cxjre
8
.8
0
0
0
F.B.
7
.8
1
1.3
3 2.3
2
.4
Biface
11
1.0
3
4.0
2 1.5
4
.9
Finished
19
1.8
1
1.3
2 1.5
39
8.4
C.D.
239
23.9
13
17.3
11 8.5
12
2.6
Total 1,
,041
100
75
99.9
130 100
462
100
3
Absent
Present
751
87.5
Cc4>-oore
4
.5
F.B.
3
.3
Blfaoe
15
1.7
Finished
33
3.8
C.D.
52
3.8
Total
858
99.9
4
Absent
Present
229
74.1
13
76.5
162 78.2
Cob-cxire
0
0
0
F.B.
2
.6
0
0
Bifaoe
10
3.2
1
5,9
5 2.4
Finished
34
11.0
1
5.9
11 5.3
C.D.
34
11.0
2
3.8
9 14.0
Total
309
99.9
17**100
207 9§.9
5
Absent
Present
556
77.0
24
60.0
14 82.3
201
73.6
Cob-core
1
.1
0
0
1
.4
F.B.
5
.7
0
0
0
Bifaoe
14
1.9
2
5.0
0
5
1.8
Finished
73
10.1
4
10.0
0
3
1.0
C.D.
74
10.2
10
25.0
3 17.6
63
23.1
Total
723
100
40**100
17**99.9
273
99.9
6
Absent
Present
403
68.3
Ccb-core
3
.5
F.B.
2
.3
Bifaoe
36
6.1
Finished
65
11.0
C.D.
81
13.7
Total
r ' Ta
590 d9.9
* Chert only; items soared "can't detennine" not incliided.
** Represented by features only.
F.B.=Flake blank C.D.=Can't detennine
314
moKaflGicavL oass
The leoarding of technological variables was designed to asperate
manufacturing use and stylistic aspects of stone tool technology. Attributes
of the varied TBOHOLOGICAL CLASS identify basic lines of tool manufacture
and provide a measure for the ancunt of energy invested in tool production.
For example, the manufacture of a bifaoe requires more work than the
manufacture of a Tinifacial tool, which in turn requires more viork than
retouched or utilized flakes. Shaped ground stone requires more time and
energy to make than use-modified ground stone. Althouch the fom of hafted
bifa^ changes throughout the Archaic stage, it is not certain that the
underlying methods for tool manufacture have changed. In the midden mound
asseshlages as a %hole, cores and tools representing both bipolar, to a mall
extent, and free hard and soft haoner reduction methods are present, as are
tools requiring all levels of time and energy investment.
One way to oonpare the use of the methods of manufacture is to construct a
measure which will take into account both the nunher of different
technological catecfories of tools present in an aseenblage (diversity) , and
the distributfon of these categories %ri.thin an assemblage (equitabilify) . Ihe
Shannon-Wiener index assesses both of these characteristics (Anick ;
Pielou ) . Diversify indices have been calculated for each analytical unit
in the study (Table 119) and scores for diversify (H) and equitabilify or
evenness (J) plotted (Figure 48) In ocnstruc^ing these indices both chipped
stone and ground stone attribute categories have been included, but categories
for debitage have been eliminated. Both diversify and evenness of assemblages
varies for units in our stufy, but there is no consistent pattern either for
one site at different time periods or for different time c^ategories. This is
one indication that site fypes are not consistent either throu^ time or
across the landscape at any one particular time.
•StME 119
DLwanify indioes for
Site
of fyP6S
Present
clams by mite
Diversify
(H max)
Evenness
(J’)
Diversify
(H')
5
1.
1.0
^ 590
+
576 F*at.
0.9
0.8
0.7
^ 590 Feat.
^ 576 Feat.
576 ^ ^ 590
• 624
539
^623
576 ^ . 539
A 624
'
621 ■ ^ 539
#
590
576
.621
A
623
0.6 H
0.5
0.7
0.8
0.9
1.0
• s Time Category 1
■ = Time Category 2
#s Time Category 3
^ B Time Category 4
• s Time Category 5
A B Time Category 6
1.1
1.2
Figure 48 Diversity and evenness gr^* for lithic technolxjgical class
and time category.
Site
Ntiriaer Maximra
Site
Time
Category
of Types
Present
Diversity
(H max)
Evenness
(J')
Diversity
(H')
22It621
1
12
1.
.
.
2
18
1.
.
.
22It623
5
17
1.
.
1.
6
15
1.
.
.
22It624
5
14
1.
.
1.
6
18
1.
.
.
* These time catecrories are repzesented by ^eituxes cnly.
H" = p. log p. , p^(i*l, ...s)
H max = iBg s
J' - H'/B max
vihere, s » muter of i^pes present
H' - Index of type diversity (information content)
H max =: Maxiimin possible diversity with given s.
J' - Evenness or equitability of specimens among types
Althou^ there is no clear dividing line among the units in the center of
the graph (Figure 48) , for oonparative purposes the ^>aoe was partitioned into
cells of low, medixxn, and high diversity and equitability. Units with high
diversity and high e^tability can be thou^ of as all-^xirpose, intensive
occupations (22It590 time category 1, 22It590 time category 2) , Units with
lew diversity and equitability can be thought of as leqparadic, ipecialized
oocipations (22It621 time category 1 , 22It623 time category 6) . The ixnits in
the low diversity, high equitability cell, in this case represented by
features at 22It590 time category 5 and 22It576 time category 5, may represent
intensive procuirement or roaintenanoe activities. These features may represent
clearap activity from a limited set of activities. Those units in the middle
cells represent oocqpation with intermediate characteristics and more
generalized oocupatim of medium intensity.
Tb get an estimate of the amount of energy invested in tool manufacture as
well as evaluate manufacturing techniques during different time periods and at
different sites, several categories of chipped stone were collapsed, and
ground stone items were dichotomized into use-modified and shaped. An
analysis of the information from thrae collapsed categories (Tables 120 and
121) shows that ground stone tools, usually ahraders and hanmers, make up a
relatively snail part of the assemblages. Most of the ground stone recovered
are fragnents with areas of grinding or smoothing. It is usually not possible
to tell if these foagnants were once part of a shaped item or a fraepent of
stone modified through use only. When this distinction can be trade,
use-modified ground stone is the most oemmon. Shsped ground stone items are
rare in edl assenhlages. Shaped ground stone tools and ornaments can require
cxxisiderable time to manufacture when oempared to chipped stone tools, ai^ in
some areas of the Midwest these tools seen to proliferate vhen a sedentary
lifestyle inen^eases (Lurie ) . Four of the stu^ units contain 60 or more
pieces of ground stone, and five units contain five or more shaped pieces.
None of these umits are from the Early Archaic. A total of 12 shaped tools
WBce reooweoned frcm the earlier tine categesries (1, 2, and 3) , %diile 20 %iieze
reooveted fran the later ones (4, 5, and 6) . Baaed on this slim evidence, it
does not gqpear that a great deal of energy went into ground stone artifact
manufacture, althouc^ more ground stone %«bs used in the later time periods.
The collapsed chipped stone categories included:
1. Utilized flakes/chunks - this category contains all pieces, flakes,
blades, or chunks that are not retouched but used.
2. Retouched tools - this category contains all flakes, blades, or chunks
that have been retouched on an edge of edges but not on the faces of
the tool.
3. Bifacial reduction flakes this category contains all flakes or blades
with bifacial platforms or conplex dorsal flake scar patterns that have
been retouched on an edge or used as a tool.
4. Bifaoes - this category contains all bifaoes.
TMU 120
rhimwd sboae artifact form by tine oaLegory*
22It539
22It576 22lt590
22It621
2^It623/22It624
T. C.
Form
N
%
N
%
N
%
N
%
N %
1
Utilized
flake/chunk
Edge retouch
56
26.2
97
27.3
89
17.9
63
17.9
only
Bifaoe
27
12.6
56
15.8
88
17.7
5
3.8
reduction
flake
61
28.5
92
25.9
58
11.7
17
12.9
Uniface
4
1.9
6
1.7
14
2.8
4
3.0
Bifaoe
Non-bipolar
56
26.2
74
20.9
165
33.2
36
27.1
ocre
3
1.4
4
1.1
7
1.4
2
1.5
Bipolar
oore/tool
Ikiidentified
0
1
.3
3
.6
0
fracpnents
7
3.3
25
7.0
73
14.7
6
4.5
Total
214
100
355
100
497
"350
133
100
2
Utilized
flake/chunk 206
22o9
11
17.7
14
12.4
125
33.7
Edge retouch
only
Bif£K3e
151
16.7
5
8.1
21
18.6
46
12.4
reduction
flake
198
21.9
9
14.5
19
16.8
21
5.7
Uniface
23
2.5
0
7
6.2
14
3.8
Bifaoe
Non*-bipolar
267
29.5
21
33.9
37
32.7
126
34.0
core
15
1.7
1
1.6
3
2.7
11
3.0
Bipolar
oore/tool
Unidentified
1
.6
0
2
2.7
0
fragments
43
4.8
15
24.2
10
8.9
28
7.5
Total
904
100
62
100
133
100
371
"T5o
318
Focm
N %
flake/chunk
Ed^e retouch
only
Biface
reduction
flake
Unifaoe
Bifaoe
Non-blpolar
core
Bipolar
oore/tool
Unidentified
fra^nents
Total
N %
156 22.2
89 12.7
92 13.1
5 .7
262 37.3
15 2.1
81 1.5
702 99.9
M %
N %
4 Utilized
flake/chunk
42
15.1
1 9.1
Edge retouch
only
Bifaoe
11
3.9
4 36.4
reduction
flake
35
12.5
1 9.1
Utiifaoe
5
1.8
0
Bifaoe
Nc3n-bipolar
160
57.3
4 36.4
core
5
1.8
0
Bipolar
oOTe/tool
Unidentified
0
0
fragnnents
21
7.5
1 9.1
Total
279
100
11**100
Utilized
flake/chunk
Edge retouch
81
12.5
5 12.5
only
Bifaoe
39
6.0
2 5.0
reduction
flake
63
9.8
4 10.0
Uniface
7
1.1
0
Bifaoe
Non-bipolar
389
60.2
18 45.0
core
15
2.3
1 2.5
Bipolar
oore/tool
unidentified
3
.5
0
fragnnents
49
7.6
10 25.0
Total
646
100
40**100
25 16.0
17 10.9
17 10.9
0
74 47.4
0
0
22 14.1
156 100
0
61
28.4
1 7.7
15
7.0
0
23
10.7
0
9
4.2
4 30.8
81
37.7
2 15.4
5
2.3
0
6 46.2
21
9.8
13**100
215
100
319
22It539
22It576
22It590 22It621
22It623/22It624
T. C. Form
N %
N %
N % N %
N %
flake/<±unk
81 16.3
Edge retouch
only
27 5.4
Bifaoe
reduction
flake
84 16.9
Uniface
3 .6
Bifaoe
248 50.0
Non-bipolar
core
6 1.1
Bipolar
oore/tool
2 .4
Unidentified
frct^nents
45 9.1
Itotal
496 99.9
are not included.
•
IMUS 121
QcDund atnne artifacts by tiae calaqDgy.
Ground
22It539
22It576
22It590 22It621
22It623/22It624
T. C. Stone
N %
N %
N % N %
N %
1
Use modified
3
33.3
4
44.4
27
84.4
0
Shaped
Unidentified
0
0
0
0
fra^nents
6
66.7
5
55.6
5
15.6
0
Total
9
ibo
9
100
32
100
0
2
Use modified
37
42.5
3
60.0
6
75.0
8
57.1
Shiped
Unidentified
7
8.1
0
1
12.5
1
7.1
fra^nents
43
49.4
2
40.0
1
12.5
5
35.7
Total
87
100
5
100
8
100
14
99.9
3
Use modified
72
66.1
Shaped
Unidentified
3
2.8
fragments
34
31.2
Total
109
100
4
Use modified
8
32.0
2
33.3
15
36.6
Sh^)ed
Unidentified
1
4.0
2
33.3
0
fragments
16
64.0
2
33.3
26
63.4
Total
25
100
6*99.9
41
100
320
T
Gtomd
I2it57S 25lt5M
IKtS3nStitS2T
. c.
Stone
M
%
N %
N %
N %
N
%
5
Use modified
32
43.2
1 33.3
3 75.0
15
65.2
Shaped
Iftiidentified
7
9.5
0
0
5
21.7
fraignents
35
47.3
2 66.7
1 25.0
3
13.1
Total
tT
100
3* 100
4* 100
23
ibo
6
Use modified
43
71.7
Sheqped
unidentified
5
8.3
fragments
12
2.0
Total
60
99.9
Represented by features only.
An examinatlcn of these categories indicates that, cn the vhole,
technology was geared toward bifaoe reductiGn. Bipolw: manufacture was never
an isportant method of producing at the sites under oonsideration. Formal
unifacial tools are also relatively rare, about 3% of the assenblages when
they are present at all. Often even utilized or retouched flakes are the
by-products of bifacial manufacture. In general, a lot of energy is devoted
to the manufacture of chipped stone tools.
Althou^ the production of bifaoes can be a time-oonsutning oocipaticn, and
eilthcuc^ there are ocnoentrations of lithic debitage at these sites, the
variable FRODUCTICN STMZ shows that most (over 80%) of the tools in the studty
units are fizuL^ied (TSedole 122) .
Although 201 stages of manufacture are present in study units exoept for
those rcfiresented lay feature material only, early stages of manufacture may
have occurred off site. Ihe predominanoe of small size ddOitage (1/4 inch or
less) lends credenoe to this explanation. Alternatively, knappers may have
been less skillful, and the mxiber of knapping failures small. Ihe best
evidence for on-site tool manufacture ocmes from Early Archaic and Middle
Archedc I ooccpations at 22It576, 22lt590, and 22lt621. In these units
unfinished tools meke ip> between 15% and 19% of the aseenblages. Knapping
oonoentraticns of debitage and unfinished broken bifaoes have been recovered
from Early Archaic 1, 22It576 (see below).
In general, lithic manufacturing methods, as opposed to stylistic elements
remains consistent during most of the Archaic. Bifaoes are the primary type
of shaped chipped tool, and the manufacture of these bifaoes could easily have
provided flake blakes for the few unifacial tools, retouched flakes and
utilized flakes in the asseahlages. Ground stone tools are primarily
unchaped, utilized pieces. Tools representing most attributes states of
lUCUNQLOGICAL CLASS are represented at all sites in edl time periods.
Differences among oonponents are more likely due to difference in site
ftnction, and therefore, the selection of specific tool from the available
repertoiire than to any basic change in the ways tools are manufactured.
Other analyses of lithic materials from regions adjacent to the stuc^
area, have shown that there is an increase in manufacture of narrcsHsladed
hifaoes during the Late Arc^iaic (Ensor , Futato ) . An informal
evaluation of hafted bifaoes in the Phase I and II lithic type collection
321
Indicates that a similar shift occurs here, but that the inaiiaer of
nartcw>4>laded hifaoes included in Fhzise III was very limited. Ihese bifaoes
%/oate
4
.6
Biface 1
16
2.4
Biface 2
22
3.4
Biface 3
14
2.2
Finished
590
91.3
Total
6ir
99.9
4
Cc3b/cx)re
5
2.2
0
3
2.1
Bifaoe 1
3
1.3
0
4
2.8
Bifaoe 2
8
3.5
1
7.7
6
4.3
Biface 3
5
2.2
1
7.7
5
3.5
Finished
105
90.7
11
84.6
123
87.2
Total
226
99.9
13**100
205
99.9
5
Cob/oore
16
3.2
1
3.3
0
3
2.1
Biface 1
4
.8
0
0
3
2.1
Biface 2
13
2.6
1
3.3
1
10.0
1
.7
Biface 3
18
3.6
4
13.3
0
4
2.8
Finished
452
89.9
24
80.0
9
90.0
130
92.3
Total
503
100
30**99.9
10**100
141
100
6 Cob/core
Bifaice 1
Bifaoe 2
Bifac% 3
Finished
_ Tbtal _ _
* Items S(3ored as "can't determine"
** Represented by features only.
5 1.1
6 1.3
6 1.3
13 2.9
417 93.3
^ _ 447 99.9
not included.
TOOL FDNCnai
IhB asaesaonant of tool function was undertaken to evaluate intensity and
variety of tool uae. Tool function here la uaed synoncBcusly with tool use
which la Identified moat often by traoea of use-wear. Thaee qualltites are
reflected both in the nvnber of different functlona that are perfcatmed with
the sane tool and the total mxnber of different functions perfooned with tools
at any site during a particular tijne period. The mnher of tool functions
repceaented on any one tool presents a mixed bag of information (Table 123) .
Sites with midden as viell as feature material in units from later time
categories (4, 5, and 6) have a fairly homogenous range of function types per
tool. PeroentagM of tools with no functions assigned range from 36% at
22It623/22It624 to 46% at 22lt590. Tools vdth one function range from 39-45%.
Ihoee with two functions range from 7-10%. Those the three functions range
from 1-4%. Multipurpose tools, therefore, ^^jpear to be scarce. Recycling is
4-5% in edl these units except at 22lt539 during time category 5. Eight
percent of the tools here were recycled.
nK2 123
Ifcatoer of tool ftaictlona by site and tiam oateaocy,
22It559
22It576
22It590
22It621
22It6i3/22It624
T. C.
Functions
N
%
N
%
N
%
N
%
N
%
1
no function
75
30.1
126
31.9
255
41.2
80
44.0
1 function
90
36.1
167
42.3
254
41.0
73
40.1
2 functions
64
25.7
68
17.2
79
12.8
22
12.1
3 functions
recycled
10
4.0
17
4.3
8
1.3
4
2.2
functions
10
4.0
17
4.3
23
3.7
3
1.6
Total
249 9^.^
m
15o
619
"TO’
TST
100
2
no functicxi
326
28.4
39
52.0
55
39.3
169
36.5
1 function
391
34.1
23
30.7
56
40.0
222
47.9
2 functions
295
25.7
7
9.3
15
10.7
51
11.0
3 functions
recycled
52
4.5
4
5.3
4
2.9
11
2.4
functions
84
7.3
2
2.7
10
7.1
10
2.2
Total
1,148
100
75
100
140
100
463
100
3
no function
313
34.4
1 function
477
52.4
2 functions
70
7.7
3 functions
recycled
6
.7
functions
44
4.8
Total
9lF
“TO
4
no function
136
40.0
4
18.2
102
46.4
1 function
132
38.8
13
59.1
88
40.0
2 functions
34
10.0
5
22.7
15
6.0
3 functions
recycled
8
2.4
0
4
1.8
functions
44
4.8
0
11
5.0
Total
340
100
22
100
220
100
323
no function
325 40.2
24
53.3
12 57.1
92
36.4
1 foBction
341
42.2
19
42.2
7 33.3
114
45.1
2 functions
69
8.5
2
4.4
0
25
9.9
3 functions
recycled
9
1.1
0
0
9
3.6
functions
65
8.0
0
2
13
5.1
Total
809
100
45
99.9
21 100
253
100
no futx:tion
268
43.2
1 function
253
40.7
2 functions
63
10.1
3 functions
recycled
13
2.1
functions
24
3.9
Total
.621
“Too
During the earlier time categories (1, 2, and 3) there was a wide range of
different functions for tools. The percentage of tools with no functions
assigned ranges from 28-52%; for tools with one function fiDcm 31-48%; for
tools with two functions fr^ 9-26%; and for tools with three functions from
1-5%. Ttools %id.th recycled functions range from 2-7%. 22It576 time categonT ^
is the most anomalous unit. It has the hi^iest percent of tools with no
function assigned and the lowest with one and two functions assigned. These
extremes my be due to the small sanple size - only 75 tools were examined -
or as suggested elesewhere, an important activity at the site may have been
tool roanufaKTture. Uhfinished tcx>ls or tools broken during production would be
less likely to exhibit use-wear. 22It576 time category 3 and 22It621 time
category 2 Boppeax to have the most eaqmdiently used tools. They have the
hic^iest percentages of tools with only one function. 22lt539 time categories
1 and 2 have the most intensively used tools. Ihey have the fewest tools with
no finction recorded and the most with two and three functions. 22It576 time
category 1 is similar to these last two imits vhile units from 22It590 and
22It621 time category 1 are intermediate to extremes of eoqpedlent or more
intensive tool use.
Table 124 presents the potential uses based on assessment of wear patterns
for each provniienoe unit in this stuc^. It reveeds that the ramhers of tools
with potential functions are very few except for 22lt590 time category 1,
22It576 time category 3, and 22It539 time categories 2 and 5. The absolute
nunber of tools with potmntied functions depends primarily on the nunober of
whole or almost %ihole tools in the unit. The most ocmicn potential function
is cutting or sawing medixxn to hard material. These tools often have another
potential function: drilling or graving a medium to hard stddstanoe. Cutting
and piercing soft material are next in frequency for potential use. Seventeen
tools have edges appropriate for use as tool backing. A few have a scraping
or chopping potentied function.
324
TliDe Category 1:
22It539
22It576
22It590
22It621
Functicn
N
%
N
%
N
%
N
%
“T
5
1.7
2
3.3
Potential cut/eaw madiun/hard
2
1.2
1
.4
7
2.4
0
Potential piercing soft
0
1
.4
5
1.7
1
1.5
Potential drill/grave madiun/hard
0
0
3
1.0
0
Potential earthing soft
1
.6
0
0
0
Potential 8cng>ing/planing hard
0
0
2
.7
0
Potential chopping
0
0
1
.3
0
Poesible backed
0
1
.4
0
4
6.6
Cutting/eaMing soft
62
38.3
88
35.2
54
18.3
10
16.4
Cutting/saMing mediun
27
16.7
34
13.6
28
9.5
7
11.5
Cutting/sahdng hard
4
2.5
1
.4
1
.3
4
6.6
Cutting/saftdng madiun/hard
5
3.1
6
2.4
14
4.7
4
6.6
Perforating soft/madiun
3
1.9
12
4.8
15
5.1
8
13.1
Drilling madiun
0
1
.4
1
.3
1
1.6
Drilling hard
0
0
0
0
Drilling
0
0
4
1.4
1
1.6
Scraping/planing soft
0
0
1
.3
0
Scraping/planing mediun
5
3.1
13
5.2
7
2.4
3
4.9
Scrig>ing/planing hard
1
.6
4
1.6
2
.7
2
3.3
Scraping dry hide
1
.6
3
1.2
1
.3
0
Scraping
24
14.8
36
14.4
62
21.0
9
14.8
Chopping/pounding soft/mediun
0
1
.4
1
.3
1
1.6
Chapping/pounding hard
2
1.2
0
13
4.4
0
Slotting/grooving/ ingraving
7
4.3
15
6.0
17
5.8
1
1.6
Wedging
0
2
.8
3
1.0
0
Tool bacdcing (hard wear)
12
7.4
23
9.2
19
6.4
0
Ran^ material supply
3
1.9
3
1.2
8
2.7
2
3.3
Pigment source
0
1
.4
5
1.7
0
Practice piece
0
0
0
0
Other ncn-^utilitarian
0
0
0
0
Abrading (a)
1
.6
1
.4
11
3.7
0
Abrading (b)
1
.6
0
0
0
Anvil
0
0
0
0
NUtting stone
1
.6
0
2
.7
0
Possible in?ojectile - inpact
0
3
1.2
3
1.0
1
1.6
Total 0 162
100.0
250
100.0
295
100.0
61
100.0
Time Category 2:
Site
22It539
22It576
22It590
22It621
Function
N
%
N
%
N
%
N
%
Potential cutting so^
9
1.2
0
1
1.5
3
1.5
Potenti2d cut/ saw mediun/hard
10
1.4
0
1
1.5
2
.9
Potential piercing soft
2
.3
0
0
1
.4
Potential drill/grave medivin/hard
6
.8
0
1
1.5
3
1.3
Potential scraping soft
1
.1
0
0
0
Potential scraping/planing hard
0
0
0
1
.4
4
124
Bpol fttrUcw by tiae category and site (ooptinaed) . _
Tine Cateqorv 2 : site
22It539 22It576 22It590 22It621
Function N % N % N % N %
0
0
0
0
PoBsible backed edge
3
.4
0
0
2
.9
Cutting/sakdng soft
180
24.4
10
27.0
11
16.4
22
9.7
Cutting/MMing mediun
85
11.5
4
10.8
8
11.9
20
8.8
Cutting/saMing hard
14
1.9
0
0
3
1.3
Cutting/sawing medivm/haxd
22
3.0
5
13.5
3
4.5
10
4.4
Perforating soft/medium
38
5.2
0
7
10.4
40
17.7
Drilling medium
8
1.1
0
0
4
1.8
Drilling hard
0
0
0
1
.4
Drilling
1
1.1
0
0
9
4.0
Scraping/planing soft
3
.4
0
3
4.5
5
2.2
Scraping/planing meditan
39
5.3
3
8.1
0
8
3.6
Scraping/planing hard
21
2.8
0
1
1.5
3
1.3
Scraping dry hide
9
1.2
0
0
6
2.7
Scraping
122
16.6
6
16.2
10
14.8
39
17.3
Giopping/pounding soft/mediun
0
0
1
1.5
0
Chopping/pounding hard
7
1.0
2
5.4
3
4.5
5
2.2
Slotting/grooving/ ingratving
73
9.9
2
5.4
4
6.0
3
1.3
Wedging
0
0
1
1.5
2
.9
Tool backing (hard wear)
22
3.0
2
5.4
4
6.0
3
1.3
Paw naterial sipply
19
2.6
0
3
4.5
12
5.4
Pigment source
3
.4
0
0
0
Practice piece
1
.1
1
2.7
1
1.5
0
Other ncn-utilitarian
4
.5
2
5.4
1
1.5
0
Abrading (a)
27
3.7
0
5
7.5
6
2.7
Abrading (b)
2
.3
0
0
1
.4
Anvil
5
.7
0
0
1
.4
Nutting stone
0
0
0
1
.4
Possible projectile - iitpact
1
.1
0
1
1.5
4
1.8
Total
727
100.0
37
99.9
67
100.0
226
99.9
Time Category 4;
Site
221t539
22It576
22It590
Function
N
%
N
%
N
%
Potential cutting soft
1
.8
Potential cut/ saw mediun/hard
3
2.3
1
7.1
2
2.3
Potential piercing soft
0
0
0
Potenticd drill/grave raediunn/hard
. 3
2.3
1
7.1
1
1.2
Potential scraping soft
0
0
0
Potential scraping/planing hau?d
0
0
0
Potenticil chopping
0
0
1
1.1
Possible backed edge
0
0
0
Cutting/sawing soft
10
7.8
0
19
21.6
Cutting/ sawing medixm
15
11.6
1
7.1
11
12.5
Cutting/ sawing hard
4
3.1
0
2
2.3
Cutting/ sawing mediun/hard
10
7.8
1
7.1
2
2.3
Tune Category 4:
Site
22It539 22It576 22It590
Function
N
%
N
%
N
%
Perforating soft/medium
19
14.7
2
i4.3
3
3.4
Drilling medium
3
2.3
0
1
1.1
Drilling hard
0
0
0
Drilling
1
.8
0
2
2.3
Scraping/planing soft
0
0
0
Scraping/planing mediiim
6
4.7
0
0
Scraping/planing hard
7
5.4
0
0
Scraping dry hide
0
1
7.1
0
Scraping
11
8.5
0
14
15.9
Chopping/pounding soft/mediuttn
1
.8
0
0
Chofjping/pounding hard
0
0
5
5.7
Slotting/grooving/ingraving
11
8.5
0
7
8.0
Wedging
0
0
0
Tool backing (hard viear)
8
6.2
2
14.3
3
3.4
Raw material supply
4
3.1
0
2
2.3
Pigment source
4
3.1
0
3
3.4
Practice piece
0
0
0
Other non-utilitarian
0
2
14.3
0
Abrading (a)
4
3.1
2
14.3
8
9.1
Abrading (b)
1
.8
0
0
Anvil
0
1
7.1
1
1.1
Nutting stone
0
0
0
Possible projectile - impact
3
2.3
0
1
1.1
Total
129
14**
99.9
88
100.0
Time Category 5:
Site
22It539
22It576
22It590 22It621
Function
N
%
N
%
N
% N %
Potenti2d. cutting soft
3
1.0
1
6.3
0
2
1.5
Potential cut/ saw mediixn/hard
17
5.8
1
6.3
0
2
1.5
Potential piercing soft
1
.3
1
6.3
0
1
.8
Potential drill/grave mediun/hard
6
2.0
0
0
1
.8
Potenti2il scraping soft
0
0
0
0
Potential scraping/planing heird
0
0
0
0
Potenticil chopping
1
1.3
0
0
2
1.5
Possible backed edge
4
1.4
Cutting/sawing soft
25
8.5
4
25.0
1
12.5 13
9.9
Cutting/ sawing medium
15
5.1
0
0
4
3.1
Cutting/ sawing hai^i
5
1.7
0
0
2
1.5
Cutting/ sawing median/hard
22
7.5
1
6.3
0
13
9.9
Perforating soft/mediim
36
12.2
2
12.5
2
25.0 8
6.1
Drilling medium
3
1.0
0
0
6
4.6
Drilling hard
1
.3
0
0
0
Drilling
0
0
0
6
4.6
Cstagory 5: _ Site
Function
22It539
N %
22It5>6
N %
i2It590
N %
22It621
N %
HK3
0
Kl
2
1.5
Scraping/planing mediixa
4
1.4
0
0
3
2.3
Scraping/planing hard
3
1.0
0
0
1
.8
Scraping dry hide
1
.3
0
0
3
2.3
Scraping
64
21.8
1
6.3
0
27
20.6
Chopping/pounding soft/medium
1
.3
0
0
0
Chopping/pounding hard
3
1.0
2
12.5
0
5
3.8
Slotting/grooving/ ingraving
16
5.4
0
0
10
7.6
Medging
4
1.4
1
6.3
0
0
Tool backing (hard wear)
3
1.0
0
0
3
2.3
Raw material scpply
18
6.1
1
6.3
2
25.0
3
2.3
Pigment source
7
2.4
0
0
3
2.3
Practice piece
3
1.0
0
0
0
Other non-utilitarian
3
1.0
0
0
1
.8
Abrading (a)
20
6.8
0
3
37.5
6
4.6
Abrading (b)
2
.7
0
0
3
2.3
Anvil
3
1.0
0
0
0
Nutting stone
1
.3
0
0
1
.8
Possible projectile - irpact
0
1
6.3
0
0
Total
294
100.0
16**100.0
8**100.0
131
100.1
Time Category 6; _ Site
22It623/22It624
EXmctlon N %
Potentieil cutting soft
5
1.8
Potential cut/ saw medivn/hard
6
2.2
Potential piercing soft
2
.7
Potential drill/grave mediun/hard
1
.4
Potential scraping soft
0
Potential scraping/planing hard
1
.4
Potential chopping
0
Possible backed edge
0
Cutting/ sawing soft
37
13.4
Cutting/ sawing medium
15
5.4
Cutting/ sawing hard
6
2.2
Cutting/ sawing medixxn/haird
21
7.6
Perforating soft/mediun
22
7.9
Drilling medium
6
2.2
Drilling hard
0
Drilling
7
2.5
Scraping/planing soft
0
Scraping/planing medium
17
6.3
Scraping/planing hard
5
1.8
Scraping dry hide
4
1.4
Scr2^ing
38
13.8
328
that show use wear, but to vMch no s^iecific function oould be assigned.
** Represented by features only
@ Totals may exceed the total inxnber of tools because tools often have more
than one function.
l^denty-seven different functions have been recorded for tools based on the
presence of use-wear. Ihese functions need to be examined with caution for
several reasons 1) the e}q)eriinental use-wear studies were limited, 2) use-wear
will usually reflect only the last or the most destructive activity for whicdi
the tool was used, and 3) it is by no means certain that the tool was used at
the site from yihijdx it was recovered. It is possible at best to get a genered
idea of the range of activities for which chipped stone and ground stone tools
were used at a site. Althou^ diversity statistics were not generated for
tool function, assenblage diversity and evenness eure apparent from the nuttber
and percentages of different use-wear types present. If the nuDober of
functions present is arbitrarily partitioned into low (10-17 different
functions) , msdiun (18-22 different functions) and hi^ (23-27 functions)
categories, sannple units diversity can be assessed sis follows:
hicb divers!
22It539 T.C.
22It539 T.C.
22It621 T.C.
22It576 T.C.
medium diversi
22It539 T.C. 4
22It576 T.C. 1
22lt590 T.C. 1
22It621 T.C. 2
22lt623/22lt624 T.C.
22It623/22lt624 T.C.
low diversi
22lt539 T.C.
22It576 T.C.
22It590 T.C.
22It590 T.C.
22It621 T.C.
T.C. = time category
Diversity does not show a consistent pattern by site or by time category.
Nor does diversity in tool function necessarily oorreqxxid to diversity in
technological class. For exai(>le, 22lt590 time category 2 has the hi^test
diversity score in terms of technology but low functions diversity. Tools
made in different ways are \iaed for sane of the same functions. On the other
other hand, 22It621 time category 2 is hi^ in functional diversity and low in
technological diversity. Tools made in the same manner are used for a variety
of functions. With these caveats it could be posited that units vdth more
diverse funcjtions represent longer term ocsctpation. If this is so, then base
canp ocxupations viere present in the early Middle Archaic.
Die percentages of tool furnction are very uneven in all units. The most
frequent cxxurring uses are partly a function of the specdficity with vhich a
viao-woar pattern could be applied during analysis. Cutting soft, mediixa, and
hard materials and scraping are inportant tool functions in all units. These
uses can be cxmsidered a base line functional assemblage. Perforating soft
materials and drilling inedivxa or hard materials and soraping dry hide,
graving, wedging, abrading, providing raw materia] are usually less toequent
functions and are far more inportant in seme units than others. All of these
aictivities can be associated with maintenancse tasks, and their inportanoe at a
site way reflecTt its status as a base caip habitation. A scale can be devised
to measure additions to the baseline functional assemblage. Sample units are
given a point for each additional function oentributing 5% or more to the tool
assenhlage, and the results are presented below. The units with the most
point are the most likely to be multipurpose cenps.
ttiit
1
2
3
4
5
22It539
time category 1
time category 2
time category 4
X
X
X
time category 5
X
22lt576
time category 1
time category 2
X
X
time category 3
X
22It590
time category 1
time category 2
time category 4
X
X
X
22lt621
time category 1
time category 2
X
X
22It623/
22It624
time category 5
time category 6
X
X
Using this scale it also appears that at least t»ro base camps are present in
the l^per Tcmbigbee Valley during the Middle Archiac 1 period.
TDOL DISPOSAL AND INTENSITSf CS" SITE USE
Other aspeerts of assemblage anadysis addressed are the intensity of tool
use and tool disposal. The percent of vhole and broken tools is an indication
of intensity of tool use and site use. The longer a tool is used, the greater
its chance of being broken. The longer the site is used, or if more people
live at a site, the more likely tools will be broken. Data on vhole and
broken tools (Table 125) shew that the hi^iest percentage of vhole tools
(52-62%) are in the earlier time categories, especially those units
330
represented by features csnly (22It576 time categcories 4 and 5; 22It590 time
category 5) . Since most of the features are pits, the hig^ percentage of
brdcen tools probably represents the final use of these facilities as dupe or
garbage pits. Althou^ the density of tools and dri>itage are greatest at
22It539 and 22It590 during time category 2, the percentage of broken tools is
lour (37.3%) and (53.0%), re^rectively) . At 22It590 tools may be used less
even if the site is more intensely us^. The percentage of tools with more
than one recorded use is low, but this is not so for 22lt539. 22It539 does
have a relatively high percentage of recycled tools, tools presumably broken
and irewcadced. At 22It576 tiiae categories 2 and 3, unlike most units from
earlier parts of the Archaic, have very high peraentages of broken tools.
Artifact density is low to migrate and the percentages of tools with more
than one recorded functicxi is low. During time category 2 at 22It576, the
hic^ percentage of broken tools may rqxcesent manufacturing activities. Early
stages of bifaoe reduc±icxi are well represented, and over half the tc»ls had
no functions recorded at all. The percentage of recycled tools is also Icm.
nau 125
JIrtifact ^^^^i^Nweso by time oataaoty*
Category
Complete
22lt539
N %
22It590
N %
22It621
N %
22It623/22It624
N %
1
Whole
130
52.2
208 52.7
254
42.1
76
41.5
Broken
111
44.6
184 46.6
341
56.5
101
55.2
C.D.
8
3.2
3 .8
9
1.5
6
3.3
Total
249
100
395 100
604
100
183
100
2
Whole
662
62.2
24 31.2
63
47.0
195
42.0
Broken
397
37.3
53 68.8
71
53.0
266
57.3
C.D.
5
.5
0
0
3
.6
Total 1,
,064
100
77 100
134
100
464
100
3
Whole
313 35.9
Broken
556 68.8
C.D.
3 .3
Total
872 100
4
Whole
106
33.3
5 29.4
83
39.2
Broken
211
66.4
12 70.6
128
60.4
C.D.
1
.3
0
1
.5
Total
318
100
17**100
212
100
5
Whole
263
35.3
14 31.1
6
31.6
119
42.8
Broken
478
64.2
31 68.9
12
63.2
158
56.8
C.D.
3
.4
0
1
.5
1
.4
Total
744
99.9
45**100
19**100
278
100
6
Whole
295
34.6
Broken
384
64.8
C.D.
3
.5
Total
592
99.9
* chert only
** Represented by features only
C.D.=can't determine
331
Another variable iimch tracks the intesity of tools use and disposal is
USE STAGE (Table 126) . Resharpening, as %iell as recycling, is monitored.
Resharpening signals intensity of tool tiae. Both the reshiurpening and
recycling stages Mere inportant at 22It539 during time categories 4 and 5 and
is probably asscx;iated vdth more intensive vise of Fort Payne chert.
nu 126
Artifact uee etaoe bv time cetaaocv*.
22It539 rStsii 22It590 22It621 22It623/22It624
TC Use Stage _ N% N% N% N% _ N %
unused
37
17.2
78
21.9
165
31.6
77
45.6
Used/Useful
107 49.8
185
51.9
218
41.8
56
33.1
Used/Discard
55
25.5
78
21.9
116
22.2
31
18.3
Reshrp/Ueeful
1
.5
5
1.4
8
1.5
2
1.2
Reshxp/Disc^rd
3
1.4
2
.6
9
1.7
2
1.2
Recyc/Useful
12
5.6
8
2.3
5
1.0
1
.6
Recyc/Discard
Total
0
215
"100
0
356
100
1
522
.2
100
0
169
100
3
4
lAwsed
163
17.1
25 38.5
28 24.6
135
32.2
Used/Useful
526
55.1
17 26.1
55 48.2
150
35.8
Used/Disc:ard
182
19.1
22 33.9
29 25.4
112
26.7
Reshrp/Useful
7
.7
1 1.5
1 .9
9
2.1
Reship/Discard
9
.9
0
0
3
.7
Recyc/Useful
65
6.8
0
1 .9
8
2.0
Recyc/Disc3rd
2
.2
0
0
2
.5
Total
954 99.9
65 100
114 100
419
100
Ifnused
159 21.6
Used/Useful
246 33.4
Used/Discard
307 41.7
Reshrp/Useful
10 1.4
Reship/Discard
6 .8
Recyc/Useful
8 1.1
Recyc/Discard
0
Total
756 100
Unused
46
19.1
3 20.0
58 33.1
Used/Useful
70
29.1
5 33.3
60 34.3
Used/Dlsc2u:d
97
40.2
5 33.3
46 26.3
Reshrp/Useful
4
1.7
1 6.7
1 .6
Reshxp/Disc^curd
10
4.1
0
5 2.9
Recyc/Useful
13
5.4
0
5 2.9
Recyc/Discard
1
.4
0 6.7
0 2.9
Total
241
100
15**100
175 100
Unused
95
17.4
6 22.2
3 21.4
59
25.0
Used/Useful
176
32.2
10 37.0
2 14.3
87
36.9
Used/Disc:ard
198
36.3
8 29.6
6 42.9
78
33.1
Reshrp/Useful
17
3.1
0
0
3
1.3
Reship/Disc:ard
28
5.1
1 7.4
1 7.1
3
1.3
Recyc/Useful
26
4.8
2 3.7
2 14.3
6
2.5
Recyc/Disc:ard
5
.9
0
0
0
Total
545 99.9
27**99.9
14**100
236
100
6
Unused
92 20.9
Used/Useful
154 34.7
Used/Discard
165 37.4
Reshrp/Useful
5 1.1
Beship/Discard
13 2.9
Recyc/Useful
9 2.0
RBcyc/Disc2u:d
4 .9
Total
441 99.9
^ Chert only; itenis sooted "can't detennine" and "not applicedsle" not
included.
** Represented by features only.
Hie recording of fracture types was undertoken to investigate the
relationsdiip between breakage and life cycles. One thousand seven hundred and
seventy-seven items di^lay one fracture type. In a few cases this fracture
type is represented more than once on the item (Table 127) . Seven hundred and
twenty-four have a ocnbination of break types. Of these, 477 have two break
types present, and 157 have ocmbinations of three break types. By far the
most oamnon type of fracture found in the archaeological ^lechnens in the
saeple is some form of transverse fr2tcture - transverse (tf=291) , transverse
with lip (nF374) , and transverse with tongue (ufIIS) . Hie diagcxvd. fracture,
similar in fonn to the transverse fracture, but with a diffej^tt orientation
to the long axis of the tool, is also a prcminent type (nF270) . Ccmbinations
of transverse and diagonal fractures are most ocninon combination of fracture
types. Singly, or in ocnbinatian, transverse and diagonal fractures make up
61% of the fractures in the sample. Fractures caused by heating are next in
frequency 0 crenated (nF273) and eipansion and pot lid (n=189) . Hiere are 70
examples of ocnbinations of heat fractures, and 54 ocmbinations of heat
fractures and other fracture types. Singly, or in combination, heat fractures
make up 23.2% of the fractures in the sample.
nms 127
ftegBency of fracture types for the entire nnnilr*. _
Fracture Type _ Frequency _ Peroentage
Perverse
23
1.0
Overshot
16
.7
Diagonal
270
12.2
Transverse
297
13.4
Transverse with lip
374
16.8
Transverse with tongue
115
5.2
Direct surface
77
3.5
Crenated
273
12.3
Expansion
182
8.2
Pot lid
7
.3
Impact
11
.5
Heift snap
79
3.6
Natural flaw
21
.9
127
F*~t*— ^ of fractare types for the entire
(oontinaed) .
Peroentaqe
Conbination of heat fractuires
3.2
Cenbination of heat and other fracture type
Conbination of material flaw and other
54
2.4
&acture type
37
1.7
Conbination transverse
58
2.6
Transverse and diagoial
240
10.8
Transverse and inpact
1
.8
2,222
* total does not inclvde the follcwing:
Other 32
Other ocnnblJiaticai 157
Not £^licable 3,173
Can't detennine 700
FrcKTtures involving natured flaws in raw material are few. Only 21
exanples of single breaks resulting fron a material flaw are present in the
sanple. Thirty-seven specimens have a fracture generated along a material
flaw in oonbination with sane other fracture type. Ihe lew percent (2.6%) of
material flaw fracture attests to the good quality of the raw materials
avzdlable to midden mound inhabitants. Perverse (i^23; 1%) , overshot
(i¥=16; 7%), iafact (i^29; 1.3% [includes oonbinations of impact and
transverse]) , and haft snaps (r^79s 3.6%) account for the rest of the
identified firacture types. Inpact fractures and haft snaps indicate the cause
or position of the fracture. Morphologicedly these fractures are transverse
or diagonal.
As can be seen in Table 128, the frequency of fracture types on Fbrt Payne
chert differs from that of all cherts. Heat-related firactures are far more
compn on Fort Peyne chert; single oocurrence &equency is 21.2% greater in
Fort Peyne chert. The single occurrence of diagonal or transverse fractures
is correspondingly 18.6% less than for all chert. Difference between Fort
Payne and €dl cherts for other fracture types and for oenbinations of fracture
types are slight.
niBEE 128
Frequency of fracture types - 1.
Fracture Type
All
Nmber
Chert
Percentage
Fort Payne Chert
Nixtber Percentage
Perverse
23
1.0
2
.4
Overshot
16
.7
1
.2
Diagonal or any transverse-
single oocuranoe
1,056
47.5
151
28.9
Direct surface
77
3.5
Heat related - single occuranoe
362
16.3
195
37.4
Impact and Inpact & Transverse
29
1.3
5
1.0
Haft sn^
79
3.6
27
5.2
Natural flaw and oonbination
natural flaw and other
fracture types
58
2.6
1
.2
128
PtcquBPcy of ftactare types - 1
All Chert
Fort Payne Chert
Fracture Type
Percentage
Oonbinaticn of heat fractures
and heat and other
fracture types
224
10.1
50
9.6
Gembination Transverse and
Transverse & Diagonal
298
13.4
69
13.2
Total
2,222
100.0
522
100.1
Frequency of combined fracture types - 2.
Fort Payne Chert
Fracture Type
Number
Percentage
Nixrber
Percentage
Perverse
23
1.0
2
.4
Overshot
16
.7
1
.2
Direct surface
77
3.5
Inpact and Inpact & Transverse
29
1.3
5
1.0
Haft snap
Natural flaw and combination
79
3.6
27
5.2
natural flaw and other
fracture types
58
2.6
1
.2
All heat related
586
26.4
245
46.9
All Diagonal and Transverse
1,354
60.9
220
42.1
Total
2,222
100.0
522
100.1
Cross-tabulation of fracture type and producticn stage (Table 129) shows
that most fracture types are associated with finished tools. Since the
hafting element of a tool is usually the l£ist step in manufacture, it is not
surprising that all haft snaps occur on finished tools. Ihe fractures aure
probably due to use, but the specific use is xsiknown, since the blade element
is missing. No refit blades and haft elements were found. Impact freKJtures
occur only in the later stages of prodiK:tion and are also likely to be
use-related. Althou^ diagonal and transverse fractures are present singly,
or in ocmbination, at all stages of reduction, they are far more oatiuon in
finished tools (86%) . Eighty-two percent of the heat-related fractures are
associated with finished tools and are likely unintentioncd. results. Only
8-10% of heat fratctures axe cissociated with early stages of biface reduction
(biface 1 and 2) . Hiese may 2dso be unintentioned results of heat treatment
fcdlure. If this is the case for at least seme &actures, then the rate of
heat-treatment fedlure is generally lew. It is edso possible that most heat
treatment took place at other sites, and failures were not brou^t to the
sites in our study. The relatively few genuinely fire-cr£8ced chunks in the
"fire-cracked chert" cateepory favors the second interpretation.
Cemparisons of fracture types by site and by time category show two trends
(Table 130) . The occurrence of single or ocmbination heat-fractures increases
in later time categories, and the percentage of cdl cembinations of fracture
types also increases through time. Ihe increase in heat fractures is in part
related to the increaised amount of Fort Payne (22It539 and 22It623/22It624)
but may be related to more intense site use as well (22lt590) . Except in
feature context, material flaws are few in later time categories. The
relatively high percentage of material flaws at 22lt576 in time categories 1
335
22It576
ITirri liM I M I I I M I I Ml I M I I I
N %
Perverse/Overshot
Diag/Ttsv-single
Direct surface
Heat-single
Inpact
Haft sn2qp
Matericd. flaw
Heat-ocnbined
Diag/Trsv^-oanbined
Total
Perverse/Overshot
Diag/Trsv-single
Direct surface
Heat-single
lipact
Heift snap
Material flaw
Heat-ocxnbined
Diag/Trsv-caiibined
Total
Perverse/Overshot
Diag/Trsv-single
Direct surface
Heat-single
lirpact
Haft snap
MatericLL flaw
Heat-ocndt}ined
Diag/Trsv-corbined
Total
Perverse/Overshot
Diag/Trsv-single
Direct surface
Heat-single
iv
■- 1 1 •
■ ,
» • < I
I - •
Use Stage
22It539
N %
22It576
N %
22lt590
M %
22lt621 22It623/22It624
M % N %
T.C
. 4
mpact
0
3.5
0
0
Haft snap
5
3.5
0
4 5.4
Material flaw
1
.7
2
33.3
3 4.0
Heat-ocmbined
11
7.7
1
16.7
4 5.4
Diag/Trsv-ocmbined
17
12.0
0
12 16.2
Itotal
142
99.9
6**100
72 100
Time category 5
Perverse/Overshot
1
.3
0
0
1
1.0
Diag/Trs^^single
116
37.8
6
27.3
3 75.0
47
49.5
Direct surface
12
3.9
2
9.0
0
6
6.3
Heat-single
96
31.3
8
36.4
1 25.0
7
7.4
Inpetct
1
.3
0
0
3
3.2
Haft snap
16
5.2
0
0
1
1.0
Material flaw
3
1.0
0
0
0
Heat-ocmbined
26
8.5
2
9.0
0
10
10.5
Diag/Trsv-ocnbined
36
11.7
4
18.2
0
20
21.1
Ttatal
307
100
22
99.9**
4 100**
95
100
Tine category 6
Perverse/Ovmrshot
1
.4
Diag/Trsv-single
102
36.0
Direct surface
19
6.7
Heat-single
80
28.3
Inpact
7
2.5
Haft snap
10
3.5
Material flaw
0
Heat-combined
27
9.5
Diag/Trsv-ccmbined
37
13.1
Total
238
100
"can't determine”, emd "not s^licable".
** Represented by features only.
2 and at 22lt590 in time category 2 may indicated restricted access to good
raw materials or increased manufacturing activity.
Density of cultural materials is often used as an indicator of site use
intensity. Intensity of site use increases as more individuals occupy an area
for a given period of time, or as the length of time a given group of people
^3ent at a site increases, or both. Natural phenomena, such as rate of
sedimentation or deflation, also effect density. Site sanpling may present
problems as well. For this reason severed, blocks and features from each site
and time category were included in the units for Phase III lithic anedysis
when possible. Usually the block at the center of the site and one or more
peripheral blocks were included in the S2nple for each time period. The
docuDentation of the relatively low rate of sedimentation and energy
investnent in site structures leads to an expectation that tool and dehitage
density dxjuld be greater during time categories 5, and 6, the later part
of the Middle Archaic and Late Archaic. Actually, the densities were hi^iest
during tlie early Middle Archaic at 22It539 and 22It590.
Ihe density of tools and debitage ISor midden and feature units in the
sanple studied has been calculated for each time category (Table 130) . The
densities are based on the total nunhers of tools and debitage recorded for
each site, block, and level (see Tables 12 and 13 of ^^pendix III) . These
data, rather than those for tools and d^itage examined in the lithic study,
have (seen used to calculate densities because of retrieval problans and the
small sanple of debitage examined during this stuy^.
TAKE 130
Tool aad debitage denaities by site and time category (rasdier of items per m*)
_ Time Category _
Site
1
2
3
4
5 6
22It539
Tools
Debitage
11-9
352.1
109.2
1,586.1
51.6
341.6
52.0
165.8
22lt576
Tools
Debitage
6.0
72.7
3.0
27.9
30.6
197.3
22lt590
Tools
Debitage
29.0
596.5
88.1
1,933.1
35.2
277.8
22lt621
Tools
Debitage
10.2
173.3
33.6
294.6
22It623/22It624
Tools
23.9 34.7
Debitage 104.4 229.1
Differences in artifact and debitage densities axe evident both among
sites within the same time category and among different time categories at the
same site, but the reasons for these differences are difficult to determine.
In fact, artifact densities are lowest during the Early Archaic, are moderate
to hi^ during the earlier Middle Archaic, and are moderate during the later
part of the Middle Archaic and Late Archaic. Little patterning exists in the
ddbitage densities. Debitctge densities ecre lew to moderate at all site and
all time categories except 22It539 and 22It590 during the early Middle
Archaic.
22lt576 has low densities for both tools and d^itage during the Early
Archaic and early Middle Archaic periods. The presence of several knapping
concentrations. Features 116 and 118 in particular, and another oonc^tration
of over 600 flidees in the northwest comer of the large excavation block, the
presence of crude-to-medium broken bifaces, and the presence of spent tools
indicates that the site was used as a manufacturing and retooling station
during at least one early oocipation. Time category 1 (Kirk) , Levels 15 and
16 proved an unusued opportunity to describe one activity prominent at the
site. Four blocks at 22It576 v^ere excavated to levels that might contain
Early Archaic deposits. Only the western two-thirds of Block D (12x8 m or
39.6x26.4 ft)) contained ten artifacts and pieces of debitage or more. Ihe
occi^tion associated with these levels appears to be restricted to the center
high area of the site. Ihe density of artifacts within these two levels of
Blodc D is 14.6 per cubic meter, the da:«ity of d^itage is 182.0 per cubic
meter, 'niese figures are in line with the other Eaurly Archadc cissenblages .
IXiring the lithic analysis, 270 tools were examined from Levels 15 and 16
midden. 'I\4enty-one tools and 487 pieces of debitage were examined from
Features 116 and 118 at 22lt576. These features were lithic concentrations
within two meters (6.6 ft) of each other. Both features were subjected to a
refitting analysis v^ch revealed new information on their composition.
Lithic material fran Feature 118 is almost exclusively non-heated yellow or
Camden chert. Based on evzduation of raw material color, texture, and cortex
debitage, four different cobbles are r^resented. IVto artifacts in the
inmediate vicinity of the feature appear to be made of material present in
Feature 118 (Figure 49) . Lithic material from five or more ccU)les of heated
Camden chert predominates in Feature 116. The distinct mottling and veining
characteristic of heated Camden chert makes it possible to recognize
individueil cobbles. No attenpt was made to reconstruct cobbles from the piles
of debitage. Nine tools or pieces of debitage arranged in an arc
^:praximately 1.5 m (5 ft) to the northwest from both Levels 15 and 16 a^spear
to be made of naterial from Feature 116 (Figure 49) . Flake size and amount of
cortex on the flakes indicates that earlier or less refined stages of
reduction are represented by the heated materials in Feature 116. The
concentration of debitage in the northwest comer of the unit is predominately
small (1/4 inch) flakes of heated chert without cortex. No heat-treatment
facility was identified in the excavated units. It is not possible to
determine if any of the unheated material was subsequently heated and then
further reduced. There are cores or core fragments in the assemblage. It is
possible that l2uge bifaces, both heated and unheated, were brought to the
site and reduced.
Nineteen artifact fragments were refit, both within and between Levels 15
and 16. Several of these were crude-to-medium bifaces that appear to have
been broken during manufacture. They are all made of materials similar to
those found in the features and lithic oonoentraticsi. None of the whole,
finished, sh^>ed tools ^peaur to match the concentrations of debitage. Three
of these vbole finished tools are small, Kirk bi faces with steeply beveled
edges. It is tempting to posit that these stone elements have been removed
from their shafts, discarded, and replaced with new, larger pieces. These
refurbished tools may than have been remcved from the site for use else^ld1ere
(Jeske, personal ccmmunication ; Stevenson ) . Tools broken during
manufacture and left at the site contributed to the moderate debitage-to-tool
ratios (Table 131) for these levels. Alternatively, the tools may have been
related to other activities that also occurred during that occupation.
Recorded functions for tools in these levels show a wide veuriety of uses as
Joiives, saws, drills, and scraper.
Refitting flakes from cobbles in
Feature I 18
Refitted flakes from cobbles
Feature 1 16
«
V
V
\
N
•
•
N. ^ N
■9
N. H
Ir \
;
TT ■ ■■
\ '
\ '
* 4
• 4 m- — -
Refitting pieces from general midden
Levels 15 and 16
Figure 49 Distributiai of flates from cxbbles cuid refit pieces.
nss 131
Ttaol-tD-deiiitaqe ratioe. _ _ _ _
Block Level T.C. Tools Debitaqe Debitaqe-to-Ttool Ratio
22It539
A
5
5
77
340
4.4
A
6
5
44
237
5.3
A
7
5
25
149
6.0
A
8
5
42
135
3.2
A
9
5
72
282
3.9
A
10
5
89
353
4.0
Total
349
1,496
4.3
B
6
5
175
608
3.5
B
7
5
209
808
3.9
B
8
5
214
802
3.7
B
9
5
76
316
4.2
B
10
5
95
282
2.9
Total
829
2,816
3.4
Total
T. C. 5
1,178
4,312
3.7
A
11
4
97
370
3.8
A
12
4
73
493
6.8
A
13
4
35
362
10.3
A
14
4
101
961
9.5
Total T.C. 4
306
2,186
7.1
A
15
2
125
1,341
10.7
A
16
2
99
1,397
14.1
A
17
2
117
1,342
11.5
Total
341
4,080
12.0
D
16
2
504
6,529
13.0
D
17
2
182
4,618
25.4
Total
686
11,147
16.2
Total
T.C. 2
1,027
15,227
14.8
D
18
1
116
3,456
29.8
D
19
1
31
1,006
32.5
D
20
1
4
160
40.0
D
21
1
0
27
Total T.C. 1
151
4,649
30.8
TRHS 131
Ttaol-to-debitage ratios (oontimed) .
Ibols
to^ltool Ratio
22It576
D
9
3
328
2,129
D
10
3
316
2,454
D
11
3
119
1,098
Total T.C. 3
763
5,681
D
12
2
33
304
D
13
2
20
228
Total T.C. 2
53
532
D
14
1
64
432
D
15
1
141
2,163
D
16
1
132
1,332
D
17
1
21
491
D
18
1
3
66 :
D
19
1
3
26
D
20
1
1
25 :
D
21
1
1
17 ;
Tatal
T.C. 1
367
4,552 ]
22It590
B
6
4
22
182
B
7
4
15
99
B
8
4
70
613
B
9
4
9
128 ]
B
10
4
20
367 ]
Total
T.C. 4
136
1,389 1
AW
8
2
130
3,093 2
Total
T.C. 2
130
3,093 2
AW
9
1
90
2,461 2
AW
10
1
42
1,206 2
131
Ttaol-to-dibifqe ratios (oa»»*-4i»i««^) . _
Blcxdc Level T.C, Tools b^itage Ddjitaqe-to-Ttool Ratio
22lt590
E
13
1
30
494
16.5
E
14
1
1
56
56.0
E
15
1
0
4
Total
31
554
17.9
F
10
1
40
747
18.7
F
11
1
21
362
17.2
F
12
1
4
83
20.8
F
13
1
1
1
1.0
F
14
1
0
1
Total
66
1,194
TO
J
7
1
49
522
10.7
J
8
1
25
165
6.6
J
9
1
6
262
43.6
J
10
1
6
186
31.0
J
11
1
4
53
13.1
J
12
1
0
13
Total
90
1,201
TOT
13
7
1
9
262
29.1
13
8
1
0
52
13
9
1
1
14
14.0
13
10
1
0
3
13
11
1
0
0
13
12
1
0
0
13
13
1
0
0
Toted
10
331
30
Total T,
.C. 1
675
14,659
21.7
22It621
A
6
2
22
554
10.1
A
7
2
59
642
10.9
A
8
2
88
1,230
14.0
A
9
2
44
532
12.1
Total
246
2,958
12.0
C
6
2
2
C
7
2
4
C
8
2
208
1,518
7.3
C
9
2
62
571
9.2
C
10.1
2
20
138
6.9
Total
296
2,227
7.5
343
22lt£21
E
8
2
1
E
9
2
74
1,225
16.6
E
10
2
33
189
5.7
Tot2d
108
T74l?
13.1
Total
T.C. 2
650
6,599
10.2
A
10
1
26
513
19.7
A
11
1
12
628
52.3
A
12
1
13
443
34.1
A
13
1
8
219
27.4
A
14
1
1
148
148.0
A
15
1
4
116
29.0
A
16
1
1
20
20.0
Total
65
2,087
~2T7l
C
10.2
1
21
159
7.6
C
11
1
49
356
7.3
C
12
1
35
279
8.0
C
13
1
10
214
21.4
C
14
1
19
185
8.7
C
15
1
9
141
15.6
C
16
1
10
132
13.2
C
17
1
14
139
9.9
C
18
1
12
99
8.3
C
19
1
8
75
9.4
C
20
1
0
7
C
21
1
0
2
c
22
1
0
4
c
23
1
0
1
Total
187
1,793
9.5
Total T.C. 1
252
3,880
15.3
%
22It623
C
6
5
72
398
5.2
C
7
5
41
160
3.9
C
8
5*
37
159
4.3
C
9
5
13
46
3.5
Total T.C. 5
163
596
4.7
D
6
6
47
365
7.8
D
7
6
18
231
12.8
Total
T.C. 6
65
596
9.2
344
22It624
A
6
6
58
770
13.3
A
7
6
46
374
8.1
Total
104
1,144
11.0
B
6
6
115
665
5.8
B
7
6
51
360
7.1
B
8
6
79
194
2.5
Total
245
T75T9
5.0
C
6
6
52
472
9.0
C
7
6
18
150
8.3
C
8
6
8
84
10.5
Total
78
706
9.0
Total
T.C. 6
427
3,496
8.2
A
8
5
20
149
7.5
A
9
5
26
90
3.5
Total T.C. 5
46
239
5.2
The hic^st densities of both tools and ddaitage during tine category 2 at
22It539 and 22It590 were not anticipated. At 22It590 the debitage to tool
ratio is the hic^iest (23.8) of all the sites at all tine periods. Since
22It590 is represented by a single 2x2 m (6. 6x6. 6 ft) block during this tine
category, our sanple unit nay not be representative of the site as a whole.
It nay be a loiapping area from vAiich finished tools han/e been removed to sene
other portion of the site or to seme other site. Ihe function identified for
the remaining tools do not represent a wide range of activities. In contrast,
22It539 has a noderate d^itage-to-tool ratio (14:8) , but the functions
recorded for the tools in the units examined frem this tine period are varied.
Ihe presence of burials at the site edso attests to a more lengthly
occcqpation.
Although the density of aurtifacts is relatively hi^ at 22It539 during
tine periods 4 and 5, the debitage densities are low. Ihe ddt>itage to tool
ratios indicate that far less tool manufacture took place than during earlier
tines. Ihe ddDitage/tool ratios are similarly low at 22It623/22lt624. During
the latter part of the Archaic the increased use of Fort Pa^ne chert probably
accounts for these low ratios. Fort Fa^ne is brou^t into the sites as
unfinished bifaces or as finished tools. Ihe finishing, or refurbishing, of
these tools would produce primarily atiall debitage, oft^ smaller than 1/4
inch.
Althou^ the dd3ita:ge~to~tool ratios differs from site to site and through
tine, the percentage of debitage in the three size categories (Table 13,
i^jperxlix III) is very consistent from site to site and throu^ tine. One-inch
pieces oenprise 1% or less of the d^itage; 1/2 inch debitage ranges from
11-19% (if 22It623 and 22It624 are considered as on a site) ; and 1/4 inch
debitape ranges from 81-89% of the total. Ihe sijiple percentages of debitage
345
categories either ineisk variation in manufacturing activity at these sites / or
the manufacturing process is a very vddespread and stable one.
SlMRRy
From examination of individual cissemblages Archaic lithic technology
appears to be essentiedly conservative, variations in the cissemblages eu?e due
to selection frcm a set of techniques known fron at least the Early Archaic
on, rather than to any innovation. The proposed increase in manufacture of
naixoiiHbladed bifaces during the Late Archaic may be an exception. Ihis
conservatism can be seen in the forgoing locdc at raw material selection and
treatment, technological classification, tool function, and tool di^)osed.
Traditional tool morphology (Table 15, i^ipendix III) also reflects the
conservatism. Several a^)ects of morphology are worth noting: 1) all of the
assemblaiges are dcminated by retouched and utilized flakes; 2) the percentage
of hafted bifaoes increases slightly throu^ time; 3) and all other tool
types, representing either technological or functioned Veuriation are usually
found in small ncnbers in all units except those represmted by feature
material only.
It is not surprising that most of the lithic assemblage is made &om
locally available chert. This generally good quality chert was often heat
treated after the initial shading of cores or bifaoes. Manufacture of tools
fron this loc2d material took place at all of the sites in this study although
evidence for the earliest stages of manufacture are slim. Non-local cherts or
other chippable stone were brou^t into the site in the form of blanks or
finished tools. Some of these exotic materials are present in all time
periods, but the use of Fort Peyne chert increase markedly during the late
Middle Archaic and the Late Archaic at some sites. This increase in Fort
Payne chert oorrespcxids to a drop in the quality of local raw material, a
decrease in the amount of heat treatment and an increase in heat alteration.
The density of debitage and tools in various units was to some extent
unexpected, ^jecifically, the high density of materied in the Early Archaic
and early Middle Archaic indicate a level of activity of longevity of
occupation uneopected at this early time period. The variation in dmisity of
tools and ddsitage among components argues for differences in site type or
function. This notation is further investigated below.
TEST EXPECTATIONS FOR C3»NGES IN MOBILITY STRATEGIES
The research design developed for this study iiK:luded the possible causes
for the changes in lithic eissemblages that occur, or seem to occur, in the
Upper Tcmbigbee River valley during the Archedc period. It has been argued
that changes in mobility strategies placed different constraints on choices
prehistoric people made during the manufacture and use of stone tools and that
those constraints in many ways cdter the conposition of lithic assemblages.
Changes are expected to oocxir in the selection and \3se of raw materials, in
the amount of energy invested into tool manufacture, and in patterns of tool
use. These issues will now be addressed utilizing data frcm this lithic
stucfy. The description of individual assemblages given above shews that there
are differences in these variables amemg sites and among time periods, but
that the patterning of these differences is semetimes difficult to discern.
To identify patterns in lithic technology the original units of analysis
have been collapsed. Since transitiOTi frcm a more mobile to a more sedentary
w^-of-life is posited to begin with oocipations containing S^^s-White
springs and Benton bifaces, the provenience units were plaoed into one of
three new tine categories (NIC) 1) Early Archaic - tine categories 1, 2, and 3
(pre-SylGes-Miite ^arings/Benton) ; 2) Middle Archaic - tine categories 4 artd 5
(S:^ces-4(hite Springs/Benton) ; and 3) Late Archaic - tine category 6
(poBt-S^ces-White ^prings/Benton. Data is presented in the form of
cross-tabulations of variables and these new tine categories. In nost cases
Chi-square statistics are not reported, because the eqpected values in a high
preportion of cells is under five, and the Chi-square statistics are
unreliable. In addition, when natrioes are larger than 2x2 m (6. 6x6. 6 ft) the
reasons for a random or non-random distributions (even if statistically
significant) is not straight forward. All tables for this section of the
chapter appear in ;^:pendix III.
It is ijiportant to note that the transition from residential canps to base
canps proposed in the model was based on assessments of site structure, the
density of all cultural material recovered, the presence of anthropedons, and
only a general knowledge of lithic £issemblages morphological characteristics
from sites without regard to time period.
Rm MATERIAL SELBCTICN
Ihe model for rest/ material selection predicts that during the Early
Archaic the primary material used for chipped stone tools would be local, good
quality chert (Camden, Tuscaloosa, and Yellow) . Ihe supply would have been
plentiful and poor quadity raw material could be rejected without penalty.
Conversely, as occepations at sites became more intense during the Middle and
Late Archaic, local raw material exhaustion would have become more likely,
therefore, more fair to poor quzdi^ rm materials would have been used for
tool manufacture. Tables 15-17 in Appendix III shews that the quality of
local raw materials does decline from Early throu^ Middle to Late Archaic.
This may be due in part to an increase in the amount of heat alteration in
these periods vhich nay obscure raw material quality.
Ihe model's esqjectation for quality of raw material is that good quality
locad raw' material was used for all types of tools during the Early Archaic.
\ Later, when good quality material was at a premium, it would have been used
selec^vely for those tools that require greater refinement. This stuc^
indicated that the use of local raw materials does not follow this pattern
(Tables 18-20, i^pendix III) . Data show a shift &om using local materials
for "retouched only" pieces (15.8% to 8.2%) to bifaoes (26.4% to 41.9%). The
percentage of good quedity loced chert used for unifaoes and utilized flakes
' edso decreases. For the late Archaic the pattern is somewhat different. Good
! qu2dJ.ty raw materials were still used in bifaoe n»nuf2x:ture but in the fcnmi of
I used or retouched bifacial reduction flakes as well as the bifaoes themselves.
I The use of good quality local materials remains lew for edge-retouched pieces
but rises sli^tly for utilized flakes/chunks.
The model also predicts that plentiful, hi^ quality material for tool
manufacture would obviate the need for heat treatment during the Early
Archadc. Heat treatment should be more pervasive during the Middle and Late
Archaic as the sipply of good quality raw material dwindles. Tebles 21-32
(Appendix III) show that this was not the case. Heat treatment of eill raw
materials declined through time. Heat treatment of local cheziis declined
slightly, and heat treatmmit of nonlocal Fort Payne chert declines during the
Late Archaic. According to the replicator (Kalin personal ccmiunication )
local good quality cherts are much easier to work once they are heated.
Although the quality of this chert in terms of gradn, flaws, of incliasions is
347
L
good/ war]cability is much iji{xroved with heating. iMs would be true during
all time periods. During the Early and Middle Archaic, seme of the Fort Payne
chert is foesiliferous Fort Payne rather than blue-gray Fort Payne.
Fossiliferous Fort P^ne responds well to heating, while blue-gray Fort Payne
does not. The decre6ise in heat treatment for FOrt Pztyne chert is related to
this shift &om the fossiliferous to blue-gray Fort Payne.
It is expected that during the Eeurly Archaic small quantities of a wide
variety of nonlocal cherts mi^t be found in assemblages probably in the form
of finished tools. In the Middle and Late Archedc, if the quality of loced
material declined, iiipcrted materials would have been used mo3:e heavily and
selectively. Relatively small amounts of Bangor, fossiliferous Fort Payne,
other cherts, and quaxrtzite ar& found at sites during the Early Archaic
primarily in the form of finished tools (Tables 24-26, Appendix III) . These
tools vhen v4x}le are small and heavily resharpened. Often only fragments were
recovered. The amounts of these exotic materials in the sample units
decreases throu^ time. The use of blue-gray Fort Payne exhibits a differmit
pattern. Although blxie-gray Fort Payne cenprises only 5% of the Early Archaic
cherts used for artifacts, its use increases through time to 19% during the
Middle Archaic and to 25% during the Late Archaic. Although the amount of
blue-^ay Fort Payne may be inflated by breakage due to heat alteration, this
is still a considerable increase. During the Early Archaic blue-gray FOrt
Payne was present cts utilized flakes, edge retouched tools, and unifaciad
tools, ais well eis bifaoes. During the Middle and Late Archaic it is present
as finished bifex^ial tools.
TOOL MftNUIACTURE
The model for change in lithic eissemblages predicts that tool production
would have been expedient during the Early Archaic. As sites were more
intensively used during the Middle and Late Archaic, efficiency in resource
procurement and in their tools would have beocme more iitportant. Tool
production would have required increcising amounts of time and energy to
improve tool efficiency. During the Early Archedc it was e}q)ected that more
edge-retouched tools and utilized flakes than formal unifacial tools or
bifacial tools eu:e e^qiected in the Eeurly Archadc data. During the Middle and
Late Archaic more extensively sheped tools - bifacial, unifaical, and shaped
ground stone, and less utilized or siitply retouched pieces are ejpected in the
assemblciges.
These expectations are met (Table 27, appendix III) . There is a
considerable shift to the use of bifaoes during the Middle and Late Archaic.
Although bifacial reduction flakes are the products of bifacial reducticxi,
their use as tools is essentied expedient. Even the use of these flakes
decrecises in the Middle Archaic. The difference between the Early Arch2dc and
other time categories would be even greater if all of the utilized flakes from
22It621 had been included in the sanple (see Section 2) . Shaped ground stone
items also increeuse during the Middle and Late Arch^dc. Shaped ground stone
made tp 5% of all the ground stone in the Early Archaic, 12.5% in the Middle
Archaic, and 8.3% in the Late Archaic.
Another way to increcise tool efficiency is to make tools more ccnplex.
Although most tool parts are perishable, the hafting element on steme can be
used as a measure of cenplexity. The model predicts that the percentage of
hafted tools vjould increeise during the Middle and Late Archaic (Table 28,
^jpendix III) . Tools with hafting elements increeise fron 10.2% of those tools
on vhich hafting could be assessed in the Early Archaic to 18.5% in the Middle
Archaic. During the Late Archaic 16.3% of the tools have hafting elements.
I
Althouc^ there are more tools vdth hafting elements in the Middle and Late
Archaic, the hafting elements are not necessarily more standardized as it has
been argued th^ should be. The cluster anedysis perfomed on named point
types diCRfs that haft len^, neck width, and base width are not helpful in
defining either marphologiced. types, nor are they particularly sensitive to
chronology. The very early Arc^iaic points, however, form tic^ter clusters.
It is possible that heifting elements become less standardized through time.
Length, width, and thickness measurements (Table 29, Appendix III) show little
evidence for standardization of tools in collapsed technological class
categories. The standard deviations do not become significantly analler in
later time periods.
As the need for efficiency increased during the Middle and Late Archadc,
and the need for a highly portable nultipurpose tool kit would have decreased.
The tools are eiqpected to have beocme more q)ecialized. The nutber of
different functions cissigned to vhole tools during this analysis can be used
to assess the specialized or multipurpose nature of tool kits. There is a
statistically significant difference in the mitbers of different functions per
tool (Tables 30-32, Appendix III) . Tools with only one function increase
during the Middle Archaic and ^d.so, to a lesser extent, during the Late
Archadc. This is particularly ^parent for FOrt Payne chert.
TOOL USE STBATE9GIES
When sites are occupied for relatively short periods of time during the
Early Archaic, the model for lithic tool use strategies predicts that tools
will be eipediently used ^ls well as be expediently manufactured. Expediently
used tools would have received less use, would have been abandoned before they
were broken and would not have been recbarpened or reworked. . Since these
sites were not oco^pied for long periods of time, tools also would have been
less likely to break through trampling ation with minimal earlier or later visits. Ihis
single-ocnponent assemblage was radiocarbon dated at A.D. 460±50 and is
consistent with those previously obtained. This is the only known isolated
assemblage recovered in the waterway. All others have been either mixed with
later Woodland material or limited to a few isolated features. The uniqueness
of this aissemblage and its potential for answering research questions
suggested a detailed analysis.
In addition, the ceramic assemblage frcm 22It606 was further studied to
provide information on the Late Woodland development into Mississippian stage.
It is readily ocmparable to the better known contemporaneous cultures in the
Central Tonbic^Dee Vall^. The late dates on this assemblage &cm this site
add to the growing theories that the Upper Tcmbigbee Valley was either "empty”
or continued the Late Woodland tradition until the 12th - 13th century A.D.
and was not "Mississippianized" as was the central valley aivi the nearby
Tupelo Hills.
As the cer2Ru.c assemblage from the Dogvxxxi Mound was the only in situ
materials recovered in this project from the Miller II period this, too, was
included in the study. These data can be compared to other oontemporaneous
sites with larger assemblages frcm midden mound sites.
RESULTS
One of the major results of this study was a detailed t^pe and variety
description for the ceramics. These data are contained in fippendix IV and
illustrate the following information for eadi type/variety; count, temper,
interior and exterior color, average sherd thickness, surface treatment,
manufacturing technique, vessel form breakdown, including diameter and lip and
beise characteristics.
METHODS AND PROCEXXJRES
All recovered ceramics, except the eroded specimens from 22It563, from the
three sites were included in the sample for the Phase III detailed analysis.
To develop the modal analysis system to be used, ceramics frcm the three sites
were retrieved frcm storage, and all decorated sherds were spread out on large
tables. The sherds were first separated into the initial Phase I and II
classification types established by Jenkins () beised primarily on surface
finish, decoration, and temper. These groups were then further segregated
into groups of discrete decorative patterns, including technique, design or
motif, and placament of design. These categories were further broken down
351
into design elements. Line she^, vddth, design element eissociation, and
porticxis of the vessel \ilnsre the design occurred were all included in the
design elements.
Classification of sh^e, decoration or surface finish, and pciste was also
included. Attribute variables included rim, base, appendages, and overall
vessel shape. Peiste variables included tenper and temper particle size. The
ceramic analysis also recorded informaticxi on technology and function. The
ccnplete code list of ceramic variables and values recorded in this stucfy are
presented in Appendix IV, Table 1.
Uiis data set is particularly conducive to descriptive and statistical
analysis. The data were entered in a format oarpatible with the Statistical
Analy^Ccil System (SftS) software pacdcage, vhioh is available and on hand at
many institutions and corporations. The data set is permanently stored cn
tapes vAiioh are avedlable from the University of Vfest Florida. 'Rie data can
then be put on disk for statisticed manipulation. SAS User's Guides available
in most bookstores are readily available and updated yearly from the SAS
Institute, Inc. in Cary, N.C. By using the ceramic attribute and code list
presented in ^jpendix IV Table 1 and SAS software, anyone can query the data
set at any time. Seme of the descriptive analysis which is particularly
useful and informative, for example, is frequency anadysis. Tables 132-142
present some of the to^s for several attributes such as treatment,
decoration, and technology. Appendix IV, Tables 2-5 presents the complete
results of the modal analysis organized by ceramic types. A study of some of
the relationships of the attributes in a few of the ceramic types follows.
THE HEMSeXJ SPRINGS CERAMIC ASSEMBLAg!
THE STUDY AMD RESULTS
The modal analysis performed on the ceramic fragnents from the Aralia
site produced much detciiled information which can be used to address many
present and future research questions. The integrity and isolation of this
assemblage in space and time, provide a base line for future collections.
The study included edl sherds identified as decorated in the preliminary
anedysis. Then all specimens in the residual pledn and eroded categories were
inspected, and those which had a discernible decoration, and those which could
truly be classified as residual plain, were also included in the stucty. No
eroded sherds were included. Although a complete analysis of the data is
beyond the scope of this report, some patterning will be described cis an
example of the kind of studies that can be done with this data set.
From the frequency of the total number of types present in this ceramic
assemblage (Table 132) , the most abundant types are Alexander Pinched (45%) ,
Alexander Incised (22%) , and Residual Plain (29%) . A few minority types are
present but make up only 5% of the assemblage. Figure 50 grafdiically
illustrates these frequencies. Coiroborating the doroinanoe of the Alexander
Pinched type. These pr inary modes of decaraticxi tended to corroborate the
dominance of the Alexander Pinched type. Punertation is the primary mode of
deexjratiem, with 67% of the sample vdiile incising acxxxmts for only 23% (Table
133) . These differences may be acxxxmted for by classification definitions,
sinca the type Alexander Pinched could not include incising, vhile the type
Alexander Incised could contedn sexm pinching elements.
352
OTHER CERAMIC TYPES
1M>le 133
I,
I
Type
Frequency
Punctation
893
67.2
Incising
301
22.6
CoRplex
129
9.7
Stanping
4
0.3
Indeterminate
1
0.2
Nodes
1
0.2
Total
1,374
Hie design element analysis of the ceramics revealed several interesting
trends. The patterns of decoration in Alexander Incised are far more vari'id
than the patterns of Alexander Pinched (Table 134 and Figures 51-52) .
Alexander Incised has 36 different patterns, vhile Alexander Pinched had only
eight. Both types had numerous sherds with indeterminate and unnamed
("other") patterns, however, most Alexander Pinched (96.1%) had indeterminate
or no pattern. Alexander Incised had only (11.7%) with no pattern. This is
probably related to the ease of pattern determination in the incised versus
pinched sherds. In addition, of the 1,043 sherds with a motif that was
discernible, the rectilinecu: design was dominant (93.7%), while only 3.1% were
curvilinear, and 3.2% were a conbination of rectilinear and curvdlinear
designs.
TstHa 134
Freouencv of Alexander Pindied and Incised decorative
Alexander Pinched
Pattern Code Fri
Patterned punctations in two directicais
5 Zone of punctaticais outlined by triangles
8 Zoned punctated
9 Punctations in triangles over horizontal lines
3 small punctaticns belcw horizcntal line
6 Parallel lines
3 Fingemadl pinches over incised lines
4 Fingemcdl pinches over crossed incised lines
Percentage
FREQUENCY
1 iaB?1E 1
0 1 967
PftTTERN
Patterns;
17
Zoned Incised
23
Punctations Over Incisions
16
Parallel Lines
13
small Punctations Below
99
Indeterminate
Horizontal Lines
8
Zoned Punctated
7
Line Filled Triangles
31
Only One Line Shewing
6
Nested Squares
3
Nested Triangles
24
Nested Diamonds
1
Cross-Hatched
27
Zoned Cross-Hatched/ Zoned
21
Maze Design Formed Between
Nested Squares
'I\gD Paraillel Lines or a
28
Zoned Maze
Single Line
30
Peurallel Incisions From Lip
34
Fingemedl Pinches Where
With Punctations On Top
Incised Lines Cross
32
Zoned Punctated (Bar- like)
25
Nested Diamonds or Triangles
20
Zoned Cross-Hatching: Vs and
14
Dentate Stamped Zoned by
Double Vs Adjacent
Horizontal Lines
19
Nested Triangles Over Parallel
98
Other
Incised Lines
2
Vertical Parallel Incised
18
One Incised Line Below Rim
E'ron Lip
12
Punctations Over Horizontal
29
Herrir^xme
Incised Lines
22
Zoned Punctated/lftTpunctated Areas
36
Rows of Pinches and Incised
33
Fingernail Pinches Over
Lines Below
Incised Lines
37
Rows of Pinches With Crossed
35
Fingernail Pinches In Acute
Lines Below
Angles Formed When Incised
9
Punctations in Triangles Over
Lines Cross
Horizontal Lines
11
Nested Triangles With Zoned
40
Crossed Incised Lines
Punctations
39 Crossed Incised Lines
26 Zoned Cross-Hatched In Panels
Figure 51 Bar chart of patterns of Alexander Incised.
FREQUEnCY
158
34 33 16 13 8 15 10 99
PfiTTERN
Patterns;
Indetenninate
Patterned Punctations in Two
Directions
Zone of Punctations Outlines
by Pinches
Zoned Punctated
Small Purjctations Below
Horizontal lanes
Parallel Lines
Fingernail Pinches Over
Incised Lines
Fingernail Pinches Where Incised
Lines Cross
Punctations In Triangles Over
Horizontal Lines
Figure 52 Beu: chart of patterns of Alexander Pindied
UbAe 134
niwiMiiry of J^beaMPCler Pinched and Inclaed deoacative patfi— rwa (oontinaed) .
Alexandi^ Pindied
Pattexn Code
Peroentaqe
98
149
17.8
No pattern
653
78.3
Total
835
Alexander Incised
Pattern Code
Peroentaqe
17
Zoned Incised
100
24.4
16
Parallel lines
84
20.5
99
Indeterminate Pattern
48
11.7
8
Zoned punctated
46
11.2
31
Only one line showing
26
6.3
1
Cross hatched
15
3.7
3
Nested triangles
15
3.7
21
Maze design
10
2.4
4
Fingernail pinching v^iere incised lines cross
6
1.5
2
Veirtical parallel incised from lip
4
1.0
14
Dentate statnped zoned by horizontal lines
4
1.0
25
Nested triangles or diamonds with zczied punctated
and unpunctated area
4
1.0
98
Other
4
1.0
29
Herrin^x)ne
3
1.0
11
Nested rectangles with zoned punctations
3
1.0
33
Fingernail pinching over incised lines
3
0.7
35
Fingernail pinching at the intersection of
acute lines
3
0.7
39
Crossed lines
3
0.7
22
Zoned punctated and ui^)unctated areas
3
0.7
23
Punctaticxis over incisions
2
0.5
13
Smedl punctations below horizcntzd lines
2
1.0
6
Nested Squares
2
0.5
7
Line filled triangles
2
0.5
26
Zoned cross-hatched with zoned maize pauiels
alternating one above the other
2
0.5
99
NO Pattern
2
30
Parallel incised from lip with fingemadl
punctations
1
0.2
12
Punctations over horizontal lines
1
0.2
2
Zoned punctated with unpunctated zones
1
0.2
9
Puctations in triangles over horizontal lines
1
0.2
18
One iiKrised line below lip
1
0.2
19
Nested triangles over pauradlel lines
1
0.2
20
Zoned cross hatching
1
0.2
24
Nested diamonds with zones punctated and
vD^unctated areas
1
0.2
27
Zoned cross-hatched with nested squauas
1
0.2
28
Zoned maze decorations with punctations over
pauradlel incised lines (Crunp Punctate)
1
0.2
32
Zoned bar^like punctations
1
0.2
357
Bride 134
gjoBg—yar of y€Bi^—lwr Pindied and TnciiiBd deoogatiwB ptbegns (oc3Btj»»iw^)
Alexander Incised
Pattern Code
Peroentaqe
36 Rows of pinching around lip bordered by incisions
38 Rows of pinching around vessel with
1
0.2
crossed lines below
1
0.2
40 Round punctations bordering incised line
1
0.2
TOtol
409
Although the ncnber or patterns identified in the Alexander Pinched t^pe
is lew (ei^t) , the most frequent pattern is "Punctations in 1\«o Directions."
Ihis pattern oont£dns 23 specimens and makes up 66.7% of the patterned
specimens. The other seven patterns contain three or fewer specimens. The
large notber of indeterminate and no pattern identified above could well be
related to small sherd size and the tendency for decorative patterns of
punctating to be large (Jenkins :119) . In small riierds, these patterns
would ix>t be identified. In the larger sherds recovered and/or refitted in
this assemblage, the decorative pattern included either the entire vessel, or
the upper half. Tiiis problem is reflected in the results of the analysis of
the arrangement of punctations and pinches of Alexander Pinched type (Table
135) , which revealed that 95.4% are patterned in seme manner as opposed to
rand^ (1.8%) , or a oembination of random and patterned (2.5%) .
Briile 135
of pmetations and pinches in
OBT—ir t9Pe: Alexander Pinched.
Type
Patterned
682
95.4
Other
18
2.5
Random
13
1.8
Oembination
2
0.3
Tot2d
751
The method of making the punctations was edLso studied, emd the results are
presented in the bar chart in Figure 53 and Table 136. The most dominant
modes are Fingernail Punctated (45.9%) and Fingernail Pinched (43.8%), with
Hemicoidal , Indeterminate, emd Other making up the remaining 11.2%.
Tri>le 136
Deoiqn analvsis. _
Design Type _ Frequency Percentage _
Incising
U-shaped
202
67.3
V-riiaped
49
16.3
Indeterminate
33
11.0
Square
10
3.3
CoRbination
2
0.7
Other
1
0.3
Subtotal
300
358
FREQUENCY
Modes of Punctaticn:
11 Fingernail Punctated
12 Fingernail Pinched
19 Indeterniinate
15 Pinched and Punctated
18 Other
31 Incised and Punctated/Pinched
13 Hendcoidal
21 Dentate
Figure 53 Bar chart of mode of punctation of ceramic type Alexander Pinched
Uie frequeiKy analysis of the decx>rative patterns identified vathin the
Aletander Incised type is presented in Table 133 and graphically illustrated
in Figure 51. Ihis analysis reveals that the dominant Alexander Incised
patterns are Zoned Incised (24.4%), Parzdlel Lines (20.5%), and Zoned
Punctated (11.7%) . Indeterminable patterns ed.so make up 11.2% of the sherds.
Ihe remaining 32.3% of the Alexander Incised sherds are ^lit between 32
different patterns, the highest of which had only 6.3% of the total, with a
frequency of 26, whereas the others axe represented by fewer than ten
specimens each. Two factors which probably effect this wide range are the
relative small size of the specimens and the ccnplexity of partial designs on
a sherd being classified. In a detsdled modal analysis of design elements
such as this, there may well be sq>arate portions of the same design
classified as two patterns. IkjWBver, this is a necessary first st^ in
unravelling the ccnf>lex decorations of the Alexander Incised ceramics. Ihe
complexity of the decorations of this Henson Springs assemblage is also
reflected in the 125 specimens that have a ccmbination of decorative modes
(Table 136) .
Also included in this study was a classification of the method of incising
(Table 136) . Ihis revealed that 67.3% of the 300 specimens with incised lines
were U-shaped, 16.3% were V-shaped, and 3.3% were squeu^-shaped in
cxoss-sectiOT .
Ihe stucty of the methods of lip modification revealed that most (65.3%)
are undecorated (Table 137) . Of the decorated specimens, the most frequently
encountered is nicking (17.8%) , hcwever, almost as many specimens with an
indeterminate decoration (15.9%) were studied. Rims au^e predaninately folded
(92.3%) and round in cross-section (72%). Pointed and flat cross-sections
were present, but low in frequency (28%) .
360
nide 137
Idin anci ly Bodification.
Nicked
47
17.8
Cord Marked
1
0.4
None
172
65.3
Other
2
0.8
Indeterminate
42
15.9
Total
264
LIP CRDS.S-SBCTION
Unfolded
Flat
28
Pointed
38
Round
175
Subtotal
244
92.3
Folded
Flat
1
Pointed
6
Round
13
Subtotal
20
7.7
Total
261
This analysis also addressed overall vessel shape, profile, and base
shape, in addition to the lip and rim traits noted above. Unfortunately, only
24 specimens are large enough to determine the overall shape (Table 138) .
Most (58.3%) of the vessels were bowls; 10 sixnple, one each of restricted,
oonoidal, flared, and restricted oonoidal bowl. Also present are eight
bealoers, all of the Alexander Pincdied type.
Ihese shapes include six regular
beaker-shaped, one
each with a globular base and a barrel shaped. Also
included in the ceramic eisseniblage is one boat-sh^)ed vessel and one recurvate
gldbulcu: jar.
Table 138
Overall aliape. _
Shape Type _ Frequency
Siirple Bcwl
Beedoers
10
Regular
6
Barrel Shaped
1
Globular Base
1
ConoidcLL Bowl
1
Flared Rim Oonoidal Bowl
1
Restricted Sinple Bcwl
1
Boat-shaped
1
Recurvate Globular Jar
1
Restricted Bcwl
1
Total
24
In an effort to determine the kind of vessel types in this ceramic
assenblage the vessel profiles of all possible sherds were recorded. This
I
revealed that most vessels had excurvate (61.9%) or vertical (25.6%) profiles
and are probably bcKirls (Table 139) . Incurvate profiles were identified fron
10.2% of the specimens, and these vere probably jars. Pea}ced profiles were
edso identified on 3.3% of the specimens. Following Halley's interpretation
of vessel function and site function (Halley ) , it appears that most of
the activities conducted at this site involving ceramic containers was with
coojdng and eating arxi that storage ceramic contairiers (jars) were not
frequently utilized. This, in turn, st?)ports the interpretation of site use
as a short-term cannpsite.
This study also investigated the possibility of patterned association of
ceramic types with vessel profiles (Table 139) . This revealed that the bowls
(excurvate and vertical profiles) were primarily dcminated by Alexander
Pinched. Hcwever, the jars were decorated fadrly evenly with Alexander
Incised and Alexander Pinched. Pealced rims were dcminated by Alexander
Incised. Althou^ the interpretive function of this profile is unJovcwn,
probably reflects eating or drin)cing containers rather than storage.
Tatle 139
it
Ceramic Type
Excurvate
Verticed
Incurvate
Pealced
Total
Alexander Incised
42
12
10
5
69
(28.3%)
Alexander Pinched
82
28
9
2
121
(49.6%)
Ool\aibus Punctate
2
1
1
0
4
(1.6%)
Crtnp Punctate
5
0
0
0
5
(2.0%)
Residual Plain
19
18
5
1
43
(17.6%)
Other
1
0
0
0
1
(0.4%i
Indeterminate
0
1
0
0
1
(0.4%)
Total
151
(61.9%)
60
(25.6%)
25
(10.2%)
8
(3.3%)
244
One of the dieignostic vessel attributes of Eeurly Wbodland and Late Gulf
Foxmational ceramics is base shape. In this eissemblage, base shape was
determinable on 111 specimens (Table 140) . Podal supports cu:e present on 104
bases, one has an annular ring, and five have a flat base. The niirber of
podal si:^3poi±s is determinable on only two vessels. One has four, and one has
three. In Table 140 shows the 45 determinable shapes: 83.2% are teat-shaped
and three (2.9%) specimens are wedge and annular shaped.
liable 140
Wi Baaal attribates.
General Shape
Percentage
Flat
5
4.0
Podal Supports
104
83.2
Annular Ring
1
0.8
Indeterminate
15
12.0
Total
125
Shape of Podal Supports
Percentage
Teat
34
32.7
Wedge
3
2.9
Annular
3
2.9
Indeterminate
64
61.7
Total
104
362
In suomary, modal attribute study of this Alexander assenablage has
resulted in nu^ new information, sane of which has cilready been described.
Hie decoration, in genercd, is characterized by the dominanoe of pinched over
incised types and rectilinear over curvilinear motifs. The modal pattern
analysis revealed that little patterning is discernible in the
pinched/punctated specimens, and only five patterns dominate the incised
specimens. Hie incised patterns also cover a much wider range than the
pinched. Rim attribute analysis shews that most were undecorated, folded, and
round in cross-section. Althouc^ few overall vessel shapes are discernible in
this particular assemblage, all but one are bowls. Hiis characteristic agrees
well with the dominance of excurvate and vertical rim profiles over incurvate.
Podal supports are the most oonmon identified form of base treatment.
Hiere are many other aspects of this isolated Alexander assemblage that
can be studied by future researchers. Hiese include the manufacturing and
firing teedmiques used to make the vessels, refitting studies to increase the
number of identifiable vessel shapes, and the location and association of
different decorative elements cxi the vessel surface. However, the frequency
an£Q.yses presented here is oan|>atible with that of other reseeuxhers, and it
is antiej^ted that it hopefully can be used directly in caiparative studies.
PHEGRAnCM OF THE RESULTS
Recently, several large Henson Springs coiponents have been excavated,
and, together with the Aredia site data, they can shed sane li^t on the
development of this culture. A seriation of the major ceramic types from
these sites has revealed aever2Ll developmental trends of the Alexander series
ceramics (Table 142) . First, larger percentages of fingernail punctated or
pinched pottery classified as "Alexander Pinched" occurred early. Second,
althou^ podal supports span the entire period, rim bosses appeared fairly
late. Finally, the Alexander Incised was domii^t during the later portion of
the phase.
H«£ 142
£as ma
Ar.
Alexander
Ke
Only decorated ceramics were included in this seriation, since the plain
bo<^ sherds associated with the Alexander series are not sufficiently distinct
fran those of the later Miller I and II ceramic oonplexes. At the Aralia
363
site, hcMever^ most plain sherds were probably included in the eroded
categpry. Therefore, only the diagnostic Alexander categories viere inclmted
in the seriaticn. Other studies have shcMn that the percentage of Plain in
Alexander assenblages may increase through time (Thcnas et al. , Table 2) .
The Aralia site appears to be the largest and earliest known site of the
Henson Springs horizon in the Tcnbic^see VeiUey. Ihe oerandc asseni>lage
contains the second hig^iest amount of Alexander Pinched (45%) of any other
site in the seriaticn, second cnly to the proximate Turtle Pond site.
Although a greater nvnher of Alexander sherds were collected from Aralia than
any Henson Springs site yet excavated, the Yarborough site heis produced a
greater midser of Alexander Incised sherds. There is, however, greater
morphological variability at Aralia than at Yarborou^. Only three varieties
of Alexander Incised (each defined by distinct decorative modes) were
recognized in the 560 ^jecimens of that type found at the Yarborough site
(Solis and Walling :84-86) , vhile at Aralia 36 distinct decorative modes
were observed (Table 143) . Ihis may indicate an inverse relationship between
the increasing percentages of incised decorative treatment and decrease in the
variability.
One distinctive feature of Alexander pottery - the rim boss - is
oonplebely absent at the Aralia site. The absence of rim bosses seems to
coincide tnixsrally with larger percentages of pinching during early Alexander
manufacture, vhich tends to sv^iport an early time period for the Aralia site
(Figure S3) . Another minority variety present at Aralia is Oolmbus Punctate
(3%) . Although this type comprises 21% of the decorated Henson Springs
assemblage at the Kello^ site, it a^ipears late in the series beranise of the
presence of a large enount of Alexander Incised, which outncnbers AlKcander
Pinched at that site four to one. However, a feature (136) at that site dated
760±70 B.C. contained primarily pinched pothery (Atkinson et al. :233) .
This date is not consistent vrith seriatim, and it seons that there was a
small earlier Henson firings oon|xxient present at Kellogg from vhich the
Colvnhus Punctate vms derived. Columbus Punctate is sunilar to the earlier
type Wheeler Punctate and is likely a lineal development.
1MU 143
n— TV OBfriCB at 2aiD531. _
Taai _ _ _
Sand Tbn()ered
Saltillo Fabric-Mar)ced, variety unspecified
Sedtillo Fabric-Maiiced, variety Tcmbiqbee
Saltillo Fabric-Marked, variety China Bluff
Subtotal
Furrs Cord-Marked, veuriety unspecified
Subtotal
Total, Sand Tendered
Limestone Tatpered
Mulberry Creek Plain, variety unspecified
Subtotal
Frequency
13
16
15
44
_8
8
52
6
6
Toted Ceramics
58
Ihe Henson Springs ceramic assemblage at the Aredla site has primary value
because it is a single caifxaient site vAiich allows reliable calculaticns of
ceramic types, frequencies, varieties, and attributes of the Alexander
ceramics ccmplex. Ihe recovered assemblage is also large which, adds
credibility to the range of types present. Ibe radiocarbon date of 460 B.P.
docixnents this oocif>ation within the early portion of the Henson Springs
period and confirms the sequence of the ceramic assemblages.
Ihe Turtle Pond site, also on the edge of the floo(^l£Lin, is a Henson
Springs site only 2 km (1.3 mi) south of Aralia with stratified Alexaixier
ceramics. Ihere is some vexticcd. mixing of ocmponents, but apparentJy, the
ceramic eissemblage at the Turtle Pond probably dates toward the earlier
portion of the ph£ise, since pinching outnurbers incising by 56% to 44%. It is
difficult to determine from the published cerannic analysis and discussion if
rim bosses or podcil sifsports were present (Ihonas et al. ) . It is evident
that the relative frequency of Baldwin Plain (Plain sand-tempered pottery)
incresises through time. Most levels, however, had significant amounts of
diagnostic Miller I pottery, indicating that an unknown portion of the Plain
sand-tempered pottery is Miller I (IhcRias , Table 2, Figure 2.1) .
The Yarborough site is 65 km (40 mi) downstream fron the Aralia near
Golvnbus, Ms and is the second largest Henson Springs site excavated in the
Tcnbic^see Valley. Unfortunately, most of the Henson Springs ceramic
assemblage was mixed with later oonponents. However, pieces are probably from
the middle of the Henson Springs phase, since the Alexander series oeramdcs
consisted of approxiRately 51% incised and 44% pincbed pottery. Rim bosses do
not seem to be present in this assemblage.
Of special interest at the Yarborou^ site was the recovery of one sherd
of St. Johns Incised. Bullen (:41) assigns this type to the Transitional
Period (1,000-300 B.C.) of Florida (Solis and walling ). Both Alexander
Incised and Tchefuncte Incised appear to be copies of St. Johns Incised
decoration (Walthall and Jenkins ) .
The Kellogg site, also near Golmbus, Ms is another example of a distuj±>ed
Henson brings ocmpcment. This site, however, contained several features with
diagnostic series Alexander ceramics, features each yielded a restorable
vessel piece. One was a strcd^t-sided vessel of Alexander Incised with six
podal supports which lack rim bosses (Atkinson et al. , Figure 13) . This
feature also yielded a radiocarbon date of 760±70 B.C. Considering the
ceramdc sample from this feature (seven Alexander Pinched, one Alexander
Incised, and two Pladn sherds) , the date appears to be at least 100 years too
early.
The majority of the Henson Springs assemblage from this site seriates late
in the Henson Springs ceramic sequerxs. Alexander Incised ocmprised 62% of
the assemblage, vhile Alexander Pinched comprised only 14%. Rim bosses and
podal supports were both present (Table 142) .
Site lGr2 is the smallest of the Henson Springs components considered in
this discussion. While the midden had been badly disturbed by later
occupaticHis, there were a few features recovered which could be definitely
attributed to this occupation. This occupation seriates late as indicated by
the fact that 69% of the eissemblage is incised and cxily 31% is pinched. Rim
bosses and podal supports are both present (Jenkins :55) (Table 142).
The finely executed and diagnostic Alexander series ceramdcs in the
Western Tennessee Valley and the Central and Upper Tcmbi^see V2d.ley has been
an enigma to most archaeologists in the Southeast. Adding to the problem has
been the iitpression of an abrupt esErgence and decline. The chrcxiologiced.
placemtent of tlie ceramics was initially 500 B.C. (Jenkins ; Jenkins et al.
365
) was initially met with resistance/ since the proposed origin of this
oerzniic series even inclvtded the Oiio Hopewellian florescence (ca. 100 B.C. -
A.D. 200) , and similarities do exist between Alexander and classic Hopewellian
ceramic styles. In addition/ similarities exist between Weeden Island (ca.
A.D. 400-1/000) and the Alexamder ceramic decorations.
Information retrieved from excavation and research aissociated with the
Tennessee-Totbi^DOO Waterway documented that the Alexander ceramic series
first developed ca. 500 B.C. and is a part of the developmmit of the Gulf
Ceramic Tradition, which existed from 2,500 B.C. until the Historic stage.
This tradition developed in the Gulf Co^tail Plain physiogr^hic province,
south of the fall-line. This tradition includes the consistent recurrence of
several decorative ceramic treatments, including incising, punctating,
fingernail pinching, and rocker and dentate shell stanping. Coastal Plain
vessel modes, with a more limited temporcLl duration, include podal supports
and rim bosses.
The Henson Springs Alexander series ceramics can be understood within the
oontSKt of a very long ceramic tradition, thich existed in the Coastal Plain
for the previous 2,000 years. This tradition persisted until historic times,
althou^ it was influenced with later ceramic transitions vhich intruded into
the Gulf Coastal Plain. The Alexander ceramic series is a conbination of
modes from previously existing and contemporary Coastal Plain ceramic
conplexes. The following discussion briefly outlines this development of
Alexander ceramics, the development and evolution of earlier Coastal Plain
ceramic complexes, and the relationship of those developments to the Alexander
series.
Gulf Tradition ceramics first appeared in the Gulf Coastal Plain region,
cilong the south Atlantic Coast, as early as 2,500 B.C., in the form of the
fiber-tempered Stallings Island and Orange series. Although these series
generally are distinct, both are characterized by incising and punctating.
Bet^imen 1,200-500 B.C. the decorative treatments and motifs found in these
early series began to spread westward across the Gulf Coastal Plain. They
occurred in differing ocnbinations, both spatially and tannporally. Between
700-500 B.C. shell stamping, rocker-stanping and dentate stanping, occurred
first in the Gulf Coasted Plain. Plain rocker-stanping was centered in the
Tchefuncte series of the lower Mississippi Vedl^, while dentate stamping was
located the the Bayou la Batre oonplex of the Lower Tcmbi^see Valley and
Mobile Bay regions. Dentate stanping also occurred in the Late Wheeler
complex along the Central and l^per Tcrabigbee Vall^ and in tlie western
Middle Tennessee Valley.
Althou^ Wheeler Dentate Stamped and Bayou la Batre Stamped cure similar
visually, their methods of explication are different. Bayou la Batre Stamped
was applied with a cut scedlop shell with five or six Granulations (2-4 cm or
.8-1.6 in wide) , vhile Wheeler Dentate Stamped was epplied with a stred^t
serrated ccmb-liJce device 3-4 an (1.2-1. 6 in) long. Wheeler Dentate Stamped
and Bayou la Batre Stamped possibly are historically related, i.e., one may
have developed from the other. There are no direct cissociations to support
such a development, but there is sane conjectural evidence, i.e. 1) the types
are morphologically similar, 2) they are centered along the Ttmbi^jee Vedley,
3) Wheeler groups derived much of their raw lithic material, Tallahatta
quartzite, from the Bayou la Batre area, and 4) both types seemi to be
contemporary, thou^ one may be e2u:lier than the other.
West of the Tombigbee Valley a Late Wheeler ocmplex did not develcp. The
lower Mississippi Valley area seemis to have been a melting pot for ceramic
development between 1,200-700 B.C. The Poverty Point trade network was
probably an inportant factor in this ferment, ^^rcscimately 700 or 800 B.C.,
an early Tchefuncte oonplex, kncMn as TchiUa, developed out of the Wheeler
complex in southern Mississippi. iMs conplex was ocnprised of rectilinear
incised decorative elements fron the previous St. Jc^s complex in Florida and
the dirag-and-jab incised decorative treatment of the Stallings Island conplex
in Georgia. Sinple stamping and punctation were carried over from the earlier
Wheeler complex. Rocker stamping was introduced at this tine. The source of
rocker stamping is unclear, but a likely candidate is the Bayou la Batre
conplex of the Mobile Bay area.
At approximately the sane time that the Tchefuncte pt>ase was developing or
slightly later ca. 500-600 B.C., the Alexander ceramic conplex also made its
first appearance. Alexander ceramics are both simileu: to and different from
Tchefuncte. Tchefuncte ceramics probably first occurred ca. 700-800 B.C. The
Alexander series appear to have developed out of a late Wheeler conplex (ca.
500-600 B.C.) . These chronological differences cause sane discrepancies in
the appearance of certain ceramic traits. Both Alexander and Tchefuncte
ceramics are characterized by rectilinear incised decoraticms throughout their
existence. Those that are characterized by hemicoidal punctation derived frm
the ecurlier Wheeler conplex. Ttiis decorative treatment is most prevedent in
both early Tchefuncte and early Alexander ceramics. The Alexander and
Tchefuncte ceramics are also both characterized by fingemedl pinching.
However, fingernail pinching is most prevalent early in Alexander and late in
Tchefuncte. Both are characterized by rim bosses, but they occur early in the
Tdiefuncte (ca. 700-800 B.C.) , but late in Alexander (ca. 300-100 B.C.) .
Podal s\pports also occur in both Alexander and Tchefuncte ceramics, but they
are late in Tchefuncte and early Alexander (ca. 500 B.C.) . Rocker stamping,
most prev2d.ent in late Tchefuncte, never was cKiopted by Alexander potters.
The Aralia site probably dates fairly early in the Alexander ceramic
sequence as Colvnbus Punctate comprises approximately 3% of the assemblage,
c»nd it is very similar to its predecessor, Wheeler Punctate. Pinching
outnunbers incising at Arailia by two to one, podal sipports are present, and
rim bosses are absent. Radiocarbon dates of B.C. and B.C.
indicate an approximate pl2u:en«ent of euxund 500 B.C. for this complex. These
dates are sanevbat complemented by the B.C. date for a similar ceramic
assemblage at the Kellogg site, but this date is probably at least 100 yeeurs
too old.
The Alexander oeramic series disappeared from the Toibigbee and adjacent
Tennessee Valley by approximately 100 B.C. In its place the Miller ceramic
series, part of the Woodland tradition appeared. These ceramics developed
south of the fall-line, within the Coastal Pleiin prior to 100 B.C.
Participation in the Gulf Cer2ndc Tradition terminated at this time in the
Upper Tcmbicbee Valley. Cord-marking and fabric-marking, hallmarks of the
woodland Tradition^ originated inland from the fall-line in an area bounded on
the north by the State of New York (Ritchie ) and eastern CaneKia and to
the south by the Upper Tennessee drainage (Lafferl^ ) .
The ceramics from the Dogvxxxi Mound were included in the modal analysis
stucfy conducted in Phase III. Although the sample is small (58 sherds) , this
wets the only in situ Middle Woodland assemblage recovered in this project. In
addition, it wsis the only sample from a burisd mound context recovered in all
the waterway investigations.
The small sample size, limited the informaticxi vbich could be obtcdned
from the ceramic anailysis. However, the results indicate that the fill of the
mound contedned ceramics typical of this cultural period cuid reflect the
replacement of the Henson Springs ceramic tradition with those of the north
(Table 142) .
Sand tempering is still doninant, however a anall amount of limestone
tendering is now present. Surface treatment is mich different and is either
fabric or cord marked. This tradition of surface treatment continues
throughout the Woodland stage in tlie U^jper Tcmbic^iee Valley.
nieae Middle Woodland ceramics have been recovered from cill sites
investigated in this project and many other sites investigated by others in
the waterway (Sense a; Thomas et al. ; O'Hear and Oonn ) , and they
reveal a strong hixnan presence in the valley during this period. Severed
ceremonial sites, such as the Pharr Mounds, of this period 2dso have been
doctnentad in the valley. It is unfortunate that the deposits of this period
have been mixed in the investigated project sites.
The last assemblage of ceramics included in the Phaise III stuc^ were from
the Late Woodland component at Site 22It606. These ceramics were from feature
context, and like those from the Dogwood Mound, represent the only in situ
^issenblage from this time period recovered in this project. The results of
this analysis are presented in Table 144 and 145 and reveal that the tempering
has changed from a dominance of sand to that of grog (59.9%) , with minor
amounts of shell tempering (19.2%), bone tenf>ering (15.4%) , and a mirture of
the three (5.0%) . Surface treatment is usually plain (43.0%) , or ooj.d marked
(42.3%), with small amounts of Incising and brushing (0.7%). These craits
reflect both the persistence of the Woodland tradition 2md the new influences
of the Mississippian tradition. The dates on this assemblage range &om A.D.
1, to 1,. This indicates that these ceramics were made at the
site and deposited throughout the Late Woodland and Mississippia^i periods.
One of the more interesting areas for future research with the data available
from this ceramic analysis is frequency anedysis of the contents of each pit.
A pit seriation could then be done and a more refined chronology of the
ceramic assemblage of these periods could be produced.
TSOUE 144
Suwsry qrocT- and bcue Lcppered oerasiics at 22It606,
Baytown Pledn, variety Roper
77
13.29
Baytown Plain, V2u:ietv Tishdninqo
154
26.59
Turkey Paw Plain, vcudety Turlcey Paw
18
3.10
Subtotal, Plain
249
42.98
Mulberry Creek Cord-Marked, variety unroecified
255
38.86
Turkey Paw Cord-Marlsed, variety Moon Lake
101
17.44
SubtotcLL, Cord-Marked
326
56.30
Alligator Incised, veuriety unspecified
3
0.51
Solcmon Brushed, variety Solomon
1
0.17
Subtotal, Other
4
0.68
Total
576
368
\
najK 145
OegaBic totals at 221t606.
Type Frequency
Percentage
by temper
Percentage
of Total
Shell Tempered
Mississiepi Plain, variety Warrior
118
79.1
15.3
Bell Plain, variety Hale
17
11.4
2.2
McKee Island Cord-Marked,
variety unspecified
1
0.6
0.1
Eroded
13
8.7
1.6
Subtotal
149
100.0
19.2
Miiced Shell and Grog Tampered
Mississippi Plain, variety Hull Lake
24
61.5
3.1
Bell Plain, variety Big Sandy
15
38.4
1.9
Subtotal
39
100.0
5.0
Grog Tempered
Baytown Plain, variety Roper
77
16.6
10.0
Baytown Plain, variety Tishcmdngo
154
33.3
20.0
Mulberry Cre^ Cord-Marked,
variety unspecified
225
48.7
29.2
Alliqator Incised, variety unspecified
3
1.0
0.6
Salonon Brushed, variety Sedonon
1
0.2
0.1
Subtotal
462
100.0
59.9
Bone Tempered
Turkey Paw Plain, variety Turkey Paw
18
15.1
2.3
Turkey Paw Cord-Marked,
variety Moot Lake
101
84.8
13.1
Subtotal
119
100.0
15.4
Total
769
100.0
SUtlABY
Hie ceramic attrihute axialysis performed in Phase III of this project heis
produced much nev and useful inforinaticm on the nature of the ceramic
2isse9nnblages in the Upper Tcmbi^see Valley during the Henson Springs, Miller
II, and Miller III horizons. Although a ociqplete analysis of the information
is beyond the scope of this project, a sanple nodal analysis was done with the
Henson Springs assenhlage data. This has provided base line information on
design elements, vessel shape, function, and patterns in the combination of
these traits. Hiis infometicm can new be used as a basis of oonparison for
other assemblages, both in the Tombi^see drainage and else»bere.
Hie information produced in the analysis of the Middle Woodland ceramic
eissemblage is limited by small sanple size, but it does docunent the cultural
affiliation of the mound. Hiis informatioi can new be used in the ceramic
studies of this period.
Hie information produced in the analysis of the Late Wxxiland/
Mississippian cissemblage was not analyzed in this project beyond a basic
frequency aneilysis of temper and types. Much more information can and should
be gained from the data.
369
CHAPIER X SlMftRSr OF RESULTS
This chapter is designed to sumarize the archaeological infamation
piroduoed in this stucty and to synthesize the results of various stuc^ elements
as they pertain to the investigatory t^potheses. These results provide the
essential ingredients to develop a second-stage model of the ade^>tive systems
practiced during the Archaic stage of oocupanoe in the tapper Tombic^see Valley.
In addition, reocDnendaticms are advanced concerning future directicns.
The research design vies refined at each phase of the seven-year project,
with the addition of new infomation and ideas. Throu^xxit, however, the
originad goal of understanding the processes and causes of cultural
adaptations during the Archaic in the Upper Tcmbi^see Valley remained. The
acquisition of new data and the intm:isity of anedysis led to goal modification
and changes in hypotheses to be tested.
The investigations centered on two primary areas of study:
paleoenvironnent and culture. The cultural studty focused on preserved
material assemblages, site settlement pattern, site use, and subsistence. The
paleoenvirormental study eu^sessed climatic chemge, biotic developments, and
landscape evolution.
StmARX OT ABCHAEDLOGICAL INFOBMAlTaM
Eleven sites in the Upper Toibigbee Valley were investigated in this
project. Three sites (edl midden mounds) were intensively excavated, and
large sanples of £dl cultural deposits were obtained (Poplar: 22It576; Walnut:
22lt539; and Ilex: 22lt590) . Four sites were less intensively excavated
(Hickory: 22It621; Beech/Oak: 22It632/22It624; Aralia: 22It563; and Site
22lt606) , and satiples of selected ocnponents were obtained, but all ocnponents
present were not sanpled. Three additional sites were only tested (Stadlax:
22Mo675; Mud Creek: 22It622; and the EXsgwood Mound: 22Mo531) and further work
on these sites was reocmnended.
A total of 530,787 ocxmted specimens and 2,425.7 kg of wei^ied material
was recovered from the IX sites investigated (Table 146) . The nunber of
recovered features totalled 359. Forty-five radiocarbon and 18
archaecmagnetic samples were submitted for radicmetric determination.
The three intensively excavated sites produced most of the recovered
material and features: 73.8% of all counted material, 40.2% of the introduced
rock, and 70.8% of all features. These acquisitions reflect both the higher
level of excavation effort and the hi^«r density of material at these sites.
The sites were on isolated, well-drained elevations in the wet floo<%>lain with
similar ccxifigurations and physiographic positions. The Ilex site was only
recently separated from the valley wall, vbereas the walnut and Poplar sites
were well within the floodplain of the Tcnbi^bee Ved.ley.
The Poplar site produced the most material amounting to 15% more than
recovered from the Walnut site and 34% more than that from the Ilex site.
The ceramic assenblages from these three large excenrations eure generally
similar in the proportion of tenper types, but do esdiibit seme differences.
The largest single aissenblage is from the Poplar site, vbich hcis three times
more ceramics than the other two sites. The main differences between the
assenblages are more sand- and fiber-tenpered ceramics at Ilac and more grog-,
shell-, and bone-tempered ceramics at the Walnut site. This probably reflects
more use of the Poplar site during the late Gulf Formational and Middle
WoocUand periods and more use of the Walnut site during the Late Woodland
period.
371
o o o o o
O gr-
ro O
m OD ^ O
• pH •
iH m
pH W f** pH 00 CH
• lA • ^
pH m O
W fN. w
« in in ph i
Ol m pH n pH (
ro
scoi inc^ mm mo
10 • • fH • m •
pH pH eo pH
fH «-p
Drills
, nrr? n ■'ft V^L^i ■■}’> tvti^M’»H3:.i>:-^»t ; q-i:>
^ o
O W
n w w
fO ^
•
sOO
n m
•H 00
«H ^
n «-•
O »H
r# fH
HO 00
n* •
HT •
o •
^ •
fO •
fH •
n •
^ Oh
m #H
« fH
•
rH •
m «
tn •
a •
fH •
« •
«n O'
^ n#
r* fH
r« O
o
r** fH
HO
«a^
•kW
m
m
r«»^ 00
fH
^OH
n
r*
Of ^
Ht
•H
m
t m
m
o \o
r> ♦
^ yc
yo •
cs o
>0 «
r**
m <
w p-l dw CJ-'
o
•
fo iM
g-
00 OH
O' _?
fH m
0)
u
I
rH Oh
S n?
I
^ O
1*1 •
'S n
51
r«» fH
HO r*
•
OH «n
^tn
m
HO OH
• •
f-l o
n
to ^
• »
^ i-i
in
• •
00 00
s:.*"
O r-l
to m
o# ^
HO OH
m -l o X'
o o o o o o
o o o o o
O O O r~ tN m O
O o m 00 o
f- T-t
• •
M O fO
•o rn in
r* w O'
a
5
«
g
■o |4
&a
p> «
s
s
6^ I
U
o
m
43
&
Z JS
I "5
& S
I 5 £ I
The chipped stone assenblages recovered frcnt the major excavated sites
aian eidiibited similarities and dif&rences. The Walnut site (22lt539) had
almoet 1,100 more specimens than the others, vihile the Poplar and Ilex sites
pcodooed almost equsd nunbers. These latter sites had similar proportions of
projectile point/l^ves (23% and 21%) , and Walnut had twice as many broken and
unidentifisdole fragments of chipped stone tools as the other two sites,
walnut also had more cores, preforms, and debitage than the other sites.
The nutber of ground stone tools varied widely between these three sites.
Ilex had much less that the others (t:p to 72% less) , and Walnut had the
greatest nisfcer.
The nuober of features is almost ec^ud at the Walnut and Poplar sites,
^dmreas the Ilex site has 29% less than the others. The proporiian of feature
types was similar, however, with pits being the most frequent feature type
(Table 147) . No rock clusters were present at the Wadnut site (22It539) ,
vddle the Il» site (22It590) had few hearths and fired aggregates. There
were only two prepared areas at the Poplar site (22It576) , five at Ilex, and
ei^t at walnut.
TABS 147
of featiTrwn with rmionown cultural affrllaticn.
Site
Fk
Mk
lA
GF
Woodland
Middle Late
MS
Total
22lt576
5 CSC
8 p
1 p
2 bu
1 blc
26 fa
2 pa
32 p
8 bu
2 ere
1 rc
1 fa
loc
Subtotal (13)
(1)
2 bu
Ten
8 bu
2 ere
(1)
(0)
(1)
(1)
^6 F
10 bu
2 ere
22It539
0
1 fa
4 pa
1 ere
1 CSC
1 blc
1 exc
12 fa
4 pa
3 be
22 p
17 bu
2 ere
1 P
1 cc
3 P
1 oc
1 P
Subtotal
(0)
(5)
1 ere
(19)
(25)
17 bu
2 ere
(0)
(1)
(1)
(3)
56 F
17 bu
3 ere
22It590
2 p
0
2 pa
3 rc
2 p
2 P
1 rc 1 rc
1 cxc
Subtotal
(2)
(5)
(2)
(2)
(2)
(1)
14 F
22It621
1 P
3 P
4 F
22It623/
22lt624
1 P
1 P
4 p 2 p
1 cxc
3 P
Subtotal (1)
Middle Late MS Total
Subtotal.
(1)
(1)
(13)
15 F
22M
1
St
1 F
Feature
Totals
5
CSC
1 p
5
blc
1 rc
6
P
1 fa
1
cc
1 cc
1
oc
12
p
1 fa
38
fa
3 be
1
cxc 1 p
1
rc
18 p
1
p
4 pa
8
pa
29 p
1 rc
1
St
1 h
2 bu
1
CSC
1 cxc
1 ere 1
be
7 cc
1
cxc
2 st
3
rc
Subtotal
17
p
6 F
85
F
33 F
7
F
16 F
3
F
20 F
2
F
189 F
2 bu
8 bu
17 bu
27 b
1 ere 2
ere
2 ere
HHHH
5 c
t^feature
bup*burial
st^tain
p=pit
cre^iccenaation
faFfixed aggregate
rc=Bzock cluster
ocBoerandc clvister csonchipped stone cluster
blc^lade cache cxoaoco|>lex cluster bHaotanical cluster
EAFEarly Archaic E/M^^va/Morrow Mountain SMS-fr'S^Ges-itiite ^nrings/Benton
MAFMiddle Archaic lAFLate Archaic GEXSulf Pamational M&41ississippian
At the four less intensively excavated sites, a single ocn()onent v«bs the
objective of the investigations. In contrast to the three intensively
excavated sites that shared few general traits. They differed in
physiographic position, intensity of occupation, midden ccsposition, and
feature visibility. Tte Hickory site (22It621) was a multiocnponent midden
mound with a thick Middle Archaic to Woodland organic midden located in the
floodplain. The Aralia site (22It563) was on the sloping sur&ce of the
vall^ wall adjacent to the floo^lain and within 400 m (1,300 ft) of the
Hickory site. It had only one major occupation, i.e. the Henson parings
horizon. The Beech and Oak sites were in the floodplain near the valley wall
and contained multiple ocsponents throuc^xxtt the prehistoric period. Biiddle
and Late Archaic features, visible in the less organically stained midden,
dominated. Site 22It606 was hi^ on a teirraoe outlier overlooking the
floodplain. While it contained multiple conponents, the occxpation was the
Late Wbodland/Mississippian.
Exoept for 22It606 material recovered from the other three sites was about
equal. This disparily between 22It606 and the other three sites probably
reflects the lack of an excavated midden at the fanner.
376
The hi^iest muter of oeraniics was recovered fron the Hickory site. The
Beech/Oak and Aralia had sijnilar amounts, and Site 22It606 had a^iprcnciinately
half the anount of these assemblages. Major ocnixxients in the assemblages
vary. 22lt606 and Aralia produced large anoounts of Late Wbodland/
Mississippian and Henson ^[xrings ceramics, re^sectively. The nulti-oonponent
Hickory site contained almost no Late Moo^and/Mississippian ceramics, while
the Beech/Oak ceramic assemblages contained 13.8% from that period. These two
latter sites contained similar hi^ amounts of Gulf Foimational and Middle
Woodland ceramics.
The chipped stone assemblages varied widely, and the Beech/Oak site
contained far more than any other of the sites. Site 22It606 agzdn had the
smallest nunter. All sites had over 30% of their chipped stone assemblages
coiposed of projectile points/knives, a figure hitter than at the larger
sites. The debitage at the Beech/Oak, Hickory^ Aralia was similar.
Althou^ the BeechAkUc site produced 30 times more ground stone tools than
Site 22It606, there was a wide range between the other three sites.
The hi^i^t ntnber of features was recorded at the Beech/Oak site (53) .
Site 22It606 produced 29 and the Aralia 12. Only four features were
encountered at the Hickory site. Pits were the most frequent type of featiires
at all sites. Postmolds were present at the Beech/Oak and 22It606 sites,
however, no patterns were detected. There were no prepared areas at these
sites, and few hearths were identified. One ceremonial cache of purposefully
broken blades and bifaoes was present in the Late Archaic component at the
Beech/Oak site.
The three sites vhich were tested varied widely in location and type. The
Mud Creek site (22It622) was a low mound in the floodplain of the Tcmbigbee,
near Site 22It606. The Stadlax site was located in a plowed field 64 km
(40 mi) south of the Mud Creek site on a low terrace edge near Amory. The
Dogwood Mcund, also near Annory, Ms, was a Middle Wbodland burial mound on the
edge of the floodplain.
The amount of material recovered from these sites varied widely, with the
Hid Creek site producing almost nine times that of the Smilax site. None of
the sites produced many ceramics. The Dogwood Mound had the hi^iest nunber,
with only 306 specimens. The most frequent artifact type %«as d^itage. The
Mud Creek site had over 9,000 specimens, and, surprisingly, the burial mound
had over 3,000. Features at these sites were scarce, with only a total of 11
encountered. The Sknilax site produced no cultural features.
The materials and features recovered in this project provide a good sample
of the archaeologic6d record of the Archaic stage ooctpations of the tapper
Tbmbigbee Valley and a few ^ situ samples of some of the post-Archaic
occipations. Of all the deposits investigated, there were 11 intact midden
deposits (18.3% of the ocmponents encountered) and 189 features (62.6% of
those leoovered) that could te affiliated with a cultural ocmponent (Table
148) . A brief summary of the distribution and contents of the archaeological
material by ccnponent follows.
The sample of the Archaic stage incliided 10 intact middens and 148
features. Early Archaic diagnostic projectile point/knives were recovered
from seven of the 11 sites investigated (Table 149) and included five chipped
stone clusters and 12 pit features (Table 147) . The hic^iest number of
diagnostic projectile point/knives were recovered from the Poplar (22lt576)
and Ilex (22It590) sites. The Poplar site produced 46 diagnostic projectile
point/)cnives and 13 features cissociated with the Early Archadc period. The
Ilex site (22It590) deposit contained more diagnostics (65) but fewer features
(only two) . The Hickcnry site (22lt621) had intact midden but no features, and
the W2dnut site (22It539) had a lew density midden and no features. The
377
pyi.j
148
ooltnEal affiliatian of
of tl» 11 8it
Labe
Pelao-
Indian Early
2flt5M -
Archaic
Middle
- QuI? -
_ Ftanna- l«»dland
Late _ tioMd Middle late Misaiesippian
4
mm
IM
IM
F(65)
144
F(l)
141
141
F(2)
141
F(2)
22It576
m
IM
F(7)
m
F(79)
144
F(3)
144
141
141
F(2)
22lt590
m
IM
F(2)
IM
F(9)
144
144
F(2)
14!
F(l)
141
141
22It563
A
A
A
A
Early:14t
Late:IM
F(9)
141
Ml
A
22It621
M4
IM
F(l)
IM
144
144
141
141
A
22It623/22It624
A IM
IM
F(3)
IM
F(5)
141
F(2)
141
141
F(l)
141
22It606
A
m
F(3)
MM
144
144
141
MM
F(9)
141
22It621
A
141
144
141
144
141
MM
144
22M
A
A
A
A
Ml
141
141
A
-
22M
A
A
A
A
Ml
141
141
A
Total Conpcnents
3 7
7
7
10
10
10
6
Intact Middens
0 5
4
1
1
0
0
0
Features
0
17
125 F
27 BU
4 CRE
7
16
3
20
2
AFnAbeent INsintact Midden MMlixed Midden BU^Burial CRE=Cremation F^Featui^
378
i
sanple of the Early Ardiaic from the Beecii/Oak site (22lt623/22It624) was so
gnan that it cannot be evaluated. An intact midden was identified, hcK«ever.
These assemblages and features reflect low density occi^tians vdiich were
oriented to hunting and tool preparation and repair.
The Middle Archedc period produced the most diverse and rich
archaeological record. It was identified at six of the sites investigated and
the sanple includes four intact middens and 124 features. The Wailnut site
cf^pears to have been a major focod point of the Middle Archaic in the UEfier
Tonbic^see Valley and cont£dned a thick intact midden with 359 diagnostic
projectile point/knives (Table 149) , 44 features, 17 burials, and one
cremation. The Poplar site was also a major foced point for this p^iod and
oontained 63 features, 10 burials, and two cremations. This site did not
contain an intact midden but did produce 109 diagnostic projectile
point/knives. The Ilex site contained 120 diagnostics but only seven features
\^ch oould be associated with this ocnponent. The other three sites
containing Middle Archaic oonponents were much weaker than those just
described, and each contained less that 51 diagnostic projectile point/knives
and a tot^d of five features.
The range of features from the Middle Archaic ocmponents is greater than
for any other prehistoric ocrrponent encountered in this project, i.e.,
62 pits
39 fired aggregates (prepared hearths)
12 prepared areas
4 rock clusters
3 botanicjd clvisters
2 blade caches (quarry blanks)
1 chipped stone cluster
1 complex cluster
17 burials
3 cremations
This infonration alone reflects the change and concentration of activities
in the Upper Tcmbi^jee Valley from the previous Early Archaic period. This
project sampled more intact middens from this farmer period (five) tut
recovered only 17 features (seven times less) and almost four times less
diagnostic projectile point/knives. The swittness and degree of adaptive
change in the Middle Archaic period is the most dramatic recorded in the
prehistoric period.
The distribution of the features during the Middle Archaiic period at the
sites investigated is far from even. As previously described, the Walnut and
Poplar sites contained 107 (86.2%) of all the features, as well as all the
burials and cremiations. These sites cbviously were focal points of this
period vpstream from Pulton, Ms in the l^per Tcmbi^Dee Valley. The four other
sites vdiich were occvpied during this period were analler and probably were
used as satellite camps for resource procurement.
T\rings
harizcn. One radiocarbon date of 429 B.C. was obtained &an this ocnfxxient,
placing it in the early portion of period docunented fac the Henson icings
horizon. The ceramic an^ysis and seriation confirm this temporal plaoemmit.
The Middle Woodland period is hard to docunent even with the large nunbers
of ceramics, because sand tempering was used in both the late Gulf Fonnation2LL
and the Middle Woodland periods. However, limestone tempering was also
practiced during this p^lod and was limited to it. Sand- and limestone-
tempered oeramiics were recovered from all sites investigztted. Sand tempering
ocmcx)aed more than half of the complete assesnblages from all but two sites
(22lt606 and 22lt539) . In the other assemiolages, it ocmpoaed from 53.0-97.9%
of the tot£d oeramdcs and was hi^^iest in proportion at the Aralia site. TSie
Poplar site (22It576) contained the hi^iest number (8,638) of these specimens.
Only three features could be affiliated with this period: cne oeramdc cluster,
one rock cluster, and one stain of ocher in the burial mound (22No531) . The
presence of the burial mound confirms the participation in this cerempnial
activity and ties the residents of the tapper Tcnft>ic|bee Valley in with the
other known area mound groups, such as Iharr and Boyd.
Late Woodland period oeramujcs are somendiat easier to segregate, because of
the close association of grog and bone tempering. Grog tempering appeared
previously, but the percentages were low. In this study, grog tempering is
used as a lazy indicator of the Late Woodland. Bone tempering appears to have
been produced primarily in the Late Woodland period. Grog*-teapered oeramiics
were recovered from all sites investigated, and bone-tempered were recovered
from five. The highest number and largest percentage of these temper types
were recovered from Site 22It606. Bone tempering and grog tempering were also
hi^ in percentages at both the Walnut (27.6%) and Poplar (14.1%) sites, vdiich
suggests the importanoe of these localities during this period.
Twenty features were recovered from this period, i.e., 18 pits, one
ceramic cluster, and one hearth. Most (13) of these were from 22It606. Six
radiocarbon dates were obtained fraa the Late Wxjdland component at 22It606
(Table 150). These dates range from A.D. 1,090 to A.D. 1,538 and leeKi to the
inference that the Late Woodland period in the Ujpper Ttmbic^see valley
coincides with the same time frame as the Missies ippian period in other areas.
The Mississippian period is not easily recognized in the Upper Tcmbi^bee
valley, especially in the central valley. The only indication of it is
shell-tempered pottery, and shell tempering is not a sure mar]oer of the
Mississippian period. It is, however, the only measure available in the
information recovered in this project and will be used.
Shell-tempered oeramiics were recovered from the seven sites vhich were
extensively or moderately excavated. None of the tested sites produced any
shell-tatpered specmens. The distribution of shell tempering follows that of
grog and bone. The hi^iest proportion is 11.7% from 22It606, with similar
percentages frcm the Walnut and Poplar sites (8.8% and 8.0%, reflectively) .
The amounts recovered frcm other sites axe anall (less than 2.0%) . The small
percentages and same distribution pattern as the grog- and bone-tempered
ceramiics suggest that shell tempering is most likely associated with the Late
Woodland ceramic assemblage, rather than with the Mississippian stage
occupation.
383
REVIEW OF mKJflHESES AND RESULTS
r..:
SPATIAL ANALYSIS OF THE MUXIf ARCHAIC PREPARED AREA
1
With the \i«ealth of archaeologiced infoonation recovered and the natural
limits to the amount of analysis that can be perfarmed in ai^ single project,
it iifBs not possible to perfora ocnplex ^atial analyses during this stu^.
HoMBver, it has been possible to perform one test case to demonstrate the
utili^ and limits to two different methods of spatial analysis in addressing
a ^>ecific research question.
One of the most interesting phenanenon that was encountered in this study
was the large ocnplex "prepared area" features associated with the Middle
Archaic occupation of the midden mounds. These were present at the Walnut
(22It539) / the Poplar (22It576) and the Ilex (22lt590) sites. They may also
have been present at the Hickory site (22It621) but the deposits were
disturbed by amateur digging and therefore could not be detected. These
features were characterized by an area of yellowish orange silt loam vhich
contrasted with the surrounding dark brown organic midden matrix. The silt
loam matrix of the features contained many pockets of diarooal and fra^nents
of hard-fired silt loam. Artifacts were generally scarce within the feature
matrix. Included in the prepared areas were well-defined harchfired
aggregates vhich very likely were hearths. These averaged 50-100 cm
(19.7-39.4 in) in diameter, viere usually oval to circular and often had two or
more "layers" of fired silt loam v^ch was very hard and reddish orange in
color. The fired aggregates also occurred regularly outside of the prepared
areas in the general midden in the Middle Archaic middens.
The ocnposition of the prepared areas varied bctamen the sites at tdiich
they occurred. This is likely due to both variation between the users of the
features, as well as the different depositional and preservational
environnents. For exanple, those at the Walnut site were well defined and had
nultiple hearths, while those at the Poplar site were not well defined and had
few hearths. HcMever, one at the Poplar site had a center cley^lined post
hole, and both had an associated flexed burial. The prepared areas at the
Ilex site were even less well defined and were not roappable. The existence of
prepared areas at this site was inferred from the proximity of hearths and
burned sandstone conoentrations.
Other researchers have also encountered different varieties of the
prepared areas at other midden mounds: Brinkley (Otinger et cd. ) , Maim
(Dye and Watrln ) , East Aberdeen (Rafferty et al. ) and 22Mo752 (Benae
b) . The characteristics and preservation of these phenoroBna vary in
detail, but all appear to be generally similar. The interpretation of the
function of these features includes structures, areas of residential activity
and hi^ intensity activity areas. One of the problems in interpreting these
features is documentation of associated artifacts. These features are
contained in a midden vAiich has a hi^ volume of artifacts, charcoal and other
features and there is no delineating attribute tl^it identifies peripherid
areas associated with the fired clay areas. Therefore, it is not known how
nucdi, if any, of the surrounding artifacts are included within the prepared
area and how much to eliminate. Therefore, the questions of function and
activities associated with these prepared areem hem been a problem for the
past seven years.
Due to the inportanoe of this issue, spatieil analysis will be used in tliis
exanple stud^ to attenpt to identify patterns in the midden materiail
surrounding a prepared area that could be associated with it. The feature and
384
block selected for stud^ is Feature 120 at the Walnut site as this vies the
best defined, nost care&iUy excavated, had nore piece plotting and was the
best dated. Background infonnation on this feature vdll be briefly revieMed
here.
Hie walnut site was the largest midden mound encountered in the upper
valley; it was ca. 9,000 square meters in size and was 1.5 m (5 ft) above the
floodplain. The Walnut site had good cultural stratification due to thickness
of the sedimentary deposits. Its physiographic position in the floodplain was
at the confluence of two major tributaries which form the Tbnbi^bee
River (Mackey's and Little Brown Creeks) both of idiich have a coarse bedload
vhich contributed to the large size of this parallel bar and vertical
separation of the cultural material within the sediments.
In this site, the Middle Archaic oonponent occupied a thick and relatively
discrete zone between Levels 5 and 13 (Ensor and Studer : 5.77-5.78) .
Associated with this oonponent was a linear arrangement of burials in a
"oemetery" . The individuals were extended and often had more that one
individual in the burial pit. This is unique in all midden mounds, and
probably reflects the importanoe of this site above the others. Burials have
been encountered in sevinral of the 16 other midden mounds, but not organized
in this manner nor in the extended position. Baaed on this and other
information frcm the Walnut site, it appears that this was seme sort of a
focal point during the S^ces-White l^ings/Benton horizon and due to its
physiogpnphic position, this oenponent was isell sealed.
There were two prepared areas associated with the Middle Archaic ocnponmit
at this site. Feature 120 which was located in the highest portion of the
site and the other (Feature 6) was on the southeast edge of the high site
surface area. Other fired aggregates or hearths were enoountered throuc^iout
the site and are associated %dth this oocupation.
Feature 120 was identified prior to excavation by systematic coring of the
site. The excavation unit (Block C) was centered over it and the
docixnentation of this feature was the primary purpose of this excavation unit.
Block C was a 10x10 m (33x33 ft) block excavated in 10 an (3.9 in) levels by
2x2 m (6. 6x6. 6 ft) units. The top of the feature was buried 80-90 cm
(31.5-35.5 in) below the surface and was first enoountered in Level 9. The
feature was isolated and pedestaled leaving any yellow or red stains or
patches of fired clay idiich were connected to or part of the feature. At the
Ixttom of Level 12, the feature limits had stabilized and the feature was
cleaned and excavated.
A total of 26 different strata were identified in the feature, most
varying in color (red or yellow) and ocnpactxmss. Several charcoal
concentrations were also contained in the feature. The feature was
disoontinuQUs and was riddled with rodent burrows, root stains and pockets of
midden, all of which were vhich were identified and removed. Onoe the feature
was fully exposed, it was 20-30 cm (7.9-11.8 in) thick and deme-shaped in
profile and amorphous in outline covering a 6x6 m (19.8x19.8 ft) area
(Figure 7) . There were two fired areas, both oenpewed of 1-4 on (3. 9-1. 6 in)
thick lexers of fired silt loon: a larger one (120x75 an/47.3-29.5 in) in the
center and a smaller one in the northeast comer.
The four eurhaecmagnetic dates and one radiocarbon date indicate that it
was in xise between 6,250 and 6,050±35 years ago. The north hearth dated
6, for the lower layer of fired clay and 6, for the vpper l29er.
Tte center hearth archaeonagnetic dates were 6, for the lower layer of
fired clay and 6, for the upper layer of fired clay. The charooed area
in the ncutheast comer of the feature dated at 6, (DIC-) . From
385
these dates, it appears that the hearths in the pirepared area %«ere used
serially, not at the same time. The northeast fired area v«as used first, with
a 50 year difference between last firings. The center hearth area was then
used with 100 years between last firings. Ihe oentr2d hearth area was much
larger and thicker, perhaps accounting for the longer use period. The
charcoal 2qpparently was from the central hearth cleanings, although it was
adjacent to the northeast hearth.
The artifactual material contained in the feature consisted of 214
artifacts and 69 kg of wel^^ied material. The artifacts consisted primarily of
flakes and broken chipped stone tools and included eic^t projectile
point/knives (one oonplete Benton and one Big Sand^ fragment) . The weic^ied
material was dcninated by fired clay (67.6 kg) with sandstone being the most
ficequent intxoduoed rock type. The cultural material was concentrated in the
center part of the the foature around the center hearth area with only miner
amounts present in the outer area. The tsrue association of these artifacts
with this Mature cannot be ironclad.
The midden surrounding and covering the feature %«as rich in cultural
material and the separation of feature vs. midden was not clear-cut except for
the hard fired areas. There viere no postholes or pits defining an enclosing
structure, nor aany other patterning of stains or features. It is possible
that the size of the excavation units were not large enough to include
perimeter features such as poetholes, however, the 10x10 m (33x33 ft) size of
the unit surrounding f\eature 120 is larger thmi any previously r^xsrted
structure for this time period.
In an attenpt to understand the context of the prepared area with their
aerially used hearths, artifact patterning in the surrounding midden was
recorded in the field through piece plotting and the use of 10 an (3.9 in)
levels for general midden excavation. In the preliminary report, Ensor and
Studer (:87) note that \ihile arti&K±8 were relatively scarce in the
pr^xured area, they noted that artifax±8 appeared to be concentrated around
the western perimeter the majority of which were chipped stone fragments and
projectile point/loiives.
The main issue involved in this issue is determining which artifacts in
the midden matrix around the feature are associated with the feature and vhich
are not. Proximity to the feature is probably ixportant, but how close and
vhere is the line drawn both horizontedly and vertically? The piece plotting
done by the excavators concentrated cn chipped stone tools, but abundance of
artifacts in the midden caused problems in Isolating all of these artifact
types. For example, there were 54 projectile point/knives (both whole and
fraepnents) piece plotted in the 30 on (11.8 in) of midden in the block
surrounding the feature. However, this is only 20% of the total projectile
point/knives from the this 30 an (11.8 In) of midden excavated in this block.
With 80% of the projectile point/knives unplotted, the patterning seen in the
plotted qmcimens probably is not valid.
Due to these problems of lack of a clear-cut periphery and hand-plotting
specimens in a hi^ density midden, a ^>atial analysis of all the material
(plotted and ui^lotted) from the midden surrounding the feature was conducted.
A file was made of all the non-feature materied from the level in vhich
the top of Feature 120 first appeared (Level 9) through its base (Level 13) in
Block C (10x10 m/33x33 ft) . This was then organized into the smallest
excavation proveniences (2 m x 2 m x 10 cm/6.6 ft x 6.6 ft x 3.9 in units) for
analysis. Then frequencies of artifact groups were then calculated for each
of the units. These groips were projectile point/lcnives, other chipped stone
tools, oores/prefonns, "(imbination" (projectile point /knives, other chipped
stone tools, ooces/prefanns) , debitage, and ground stcxie. In additicm, the
total amounts of counted specimens (stone tools and debitage) and the total
amount of iiiei^ied materials (introdiioed rock and fired clay) were calculated
for each level.
This data was then analyzed tising both SAS and SYMAP program packages.
The objective of this study was to present this data in such a manner that if
patterns existed in the data th^ could be identified. The first method was
using the SAS plot {arograms. This generated a sinulation of the flcxa: of the
excavation unit divided into the 25 2x2 m (6. 6x6. 6 ft) units with a graduated
symbol in the center of each depicting the frequency of the designated data
category. The choice of synhols that could be reproduced using only black and
white were limited to "pikes'* and "patterns”. Studies of these plots did not
reveal any identifiable patterns except for the abeenoe of midden in the area
of the feature. The qpiJce plots were the easiest to read, but the large
noiber and close praximil^ of the spikes made it difficult to Interpret. As
can be seen in the exanples of these plots in Figure 54, the rows are
difficult to identify and (xmpare at each level. However, stucfy of the spike
printcxxts seemed to indicate that each level had a different distribution
pattern and patterns between levels %«re not reveaded.
In an effort pursue this issue, the data were analyzed through the SYMAP
oonputer mapping program package. TVio mapping programs were used: contour and
tre^ surface analysis. The m^ing area (or window) was the 10x10 m
(33x33 ft) block and the same data (artifacrt categories by level) used in the
SAS analysis was entered. Contour imps display data by interpolating a
continuous surface in the regions th^ are ix> data points, basing the
values on the distance to and the values of nei^iboring data. Tte contour
lines drawn identi:fy areas of relatively higher and lower frequencies of
qecimens. Trend surface maps display data differently in that the dixec:tion
of "trends" in the frequency of specimens is di^layed cpver the whole uhit.
The biggest difference in the two mapping methods is that cxxxtour mape are
more sensitive to frequencies within the blcxk v4iile trend surface maps
display the interpolated "tendencies" or patterns of the feequanciee.
In using these mapping programs, cxie inportant factor to be cxxusidered is
that the feature was excluded from the data and the volixnes of the 2x2 m
(6. 6x6. 6 ft) units axe unequal. As depth increased, less midden was excavated
and more feature was left pedestaled. This effec:ts the distribution maps in
that the units ccntaining the feature axe considered in the mapping program as
areas of lew frequency rather than areas excluded due to the expansion of the
feature. To partially oonpensate for this factor, the "footprint” of the
feature has been marked on the maps so that it c:an be exmsidered in the
following interpretation.
Tb get a general pattern for the distribution of stone tcx>ls and d^itage,
the first pattern considered was the total amount of exunted specimens in each
level. The contour maps of this data axe presented in Figure 55a and Figure
55b as well as the trend surface maps in Figure 56. In ocxnparing the trend
and cxntour maps for the same levels, the relaticxiship between them is clear.
The smoothing efte± of the trend surface contrasts to the more patc^ and
detailed oontcur maps.
A preliminary studty of the cxntour maps by level reveed.s that >4iile there
are patterns in the distribution of specimens, most do not hold through
oonsecutive levels. For example, in Level 9, where only the top of the
feature in the center of the blcxk was e>qx>sed, the material was oonoentrated
east, northeast, and west of the feature. However, in Level 10 material is
concentrated southeast and west of the feature. In Level 11 material is
oonoentrated north and west of the feature. In Level 12 material is
Level 10 Classes; in numbers
0 = 3 - 27 3 = 76 - 99
1 = 28 - 51 4 = 100 - 123
2 = 52 - 75 5 = 124 - 147
Level 11 Classes: in numbers Level 12 Classes: in numbers
0 = 0 - 28 3 = 86 - 1 13 0 = 50 - 67 3 = 102 - 1 18
1 = 29 - 56 4 = 114 - 141 1 = 68 - 84 4 = 1 19 - 135
2 = 57 - 85 5 = 142 - 170 2 = 85 - 101 5 = 136 - 152
Figure 55a Contcur maps of counted specimens in Levels 9-12,
Block C, 22It539. (Feature 120 blocted out) .
Level 9 Classes: in numbers
0 =
50
- 63
3 =
90 - 102
1 =
64
- 73
4 =
103 - 115
2 =
77
- 89
5 =
116 - 129
Level 13 Classes; in numbers
0 =
0-37
3 =
112 -
148
1 =
38 - 74
4 =
149 -
185
2 =
75-111
5 =
186 -
223
Figure 55b Contour map of counted specimens in Level 13, Block C
22It539 (Feature 120 blocked out) .
concentrated east and west of the feature. In Level 13 materia] is
concentrated east, south and west of the feature. 'Oxis analysis indicates
that throu^MUt the 200-year use of this feature atid buildup of mi Aten around
it, consistently more materials were deposited on the western periphery that
any other peripheral area. Knowing that the noctiieast hearth was used first
and the oentral hearth used last does not aid in understanding the differing
position of other areas of material ocncentration between levels.
The trend surface maps generally reflect the oonoentrations described
above. However, the increased genocalizaticn or lack of qjecificity decreases
the sensitivity to the changing patterns. The trend of more materials in the
western part of the block is reflected in these maps, as well as the
variatiens in the other levels.
In an effort to identify patterns between levels and relate then to the
specific artlfcx± categories to the totals used in the first contour stud^,
contour maps of one tool type, projectile point/knives, were selected. These
are presented in Figure 57, and it can be seen that when oenpared to those for
the totals, the patterns for each level are not similar. Nhm viewed singly
between levels, the ocncentration of projectile point/knives tends to be on
the western and northern areas of the block.
Maps were made of the other artifact categories with the sane general
results: a general trend for more material to be located along the western
perimeter with variance in the other areas of oonoentraticn between levels.
Overall, the use of STMKP for spatial analysis to address the patterning
of material within the confines of a 100 square meters (119.6 square yards)
block with an {qproodmately 36 square meters (43 square yards) feature in the
center did reveal sene patterning. First, the consistent presence of higher
amounts of material in the midden on the viestem periphery was documented.
Also, these maps doconented the consistent changing of other areas of
oonoentraticn throuc^iaut the buildip of midden. tCiile this is a pattern in
itself, it is difficult to interpret. The problem of the unexcavated feature
areas being included in the intapolation of the data by the SYMAP programs is
trovhlescme, and undermines the credibility of the patterns depicted in the
contour and trend surface maps.
After using SYMKP to address this research question, it appears that the
best use of this package would be at the site level of stuc^ not a single
excavation unit. For exaiple, the question of site activities during the
Middle Archaic Sykes-4tiite parings as opposed to the Benton hcarizon at the
Walnut site could be analyzed throuc^ contour and trend surface mapping by
using all midden material from the designated levels from all excavation
units. This larger area and sample size is more conducive to these mapping
programs than the restricted area of one small excavation xmit. These
programs are designed to reveal general trends and patterns in large
geographic areas and likely could produce spurious patterns v4xen confined to a
small unit with a large fMture.
In ^ite of the problems with SYMAP, it was more useful, by far, than the
SAS plotting. The latter method identified no patterns and the plots were
difficult to stud^.
Level 9
Classes: In grams
2 = 83 - 106g
3 = 107 - 130g
Level 10
Classes: in grams
2 = 100 - 128g
3 = 129 - 156g
0 = 33 - 57g
1 = 58 ' 82g
0 = 42 - 71g
1 = 72 - 99g
Level 12 Classes: in grams
0 = 62 - 89g 3 = 145 - 172g
1 = 90 - 1 leg 4 = 173 - 199g
2 = 1 17 - 144g 5 = 200 - 227g
Figure 56 Trend surface maps of all weighed material, Levels: 9,
10, and 12, Block C. (feature 120 out).
pM^GBivincNen'
The envizonoaental hypotheses addressed in these investigations focused on
tescwrature/iiDistuie fluctuations and periods of landscape stability/ flux in
the Early and Middle Holooene. Soil develofment was investigated.
It was hypothesized that the Early Holooene was cooler and moister than
present; the floo^lain vas characterized by an unstable surface; and the
Tcnfcigbee was a bnduded stream. These hypotheses were scqpported primarily by
sedin^ analyses which demonstrated a higher sedisoentatim rate in the Early
Holooene (8-17.6 cm/oentury) than any other period. Particle size analysis
doconented that the Early Holocane sediments contained more coarse materials
than deposits of later periods. The hic^ sedimentation rate and the coarse
particle size suggest that the floodplain landscape of the Early Holooene was
unstable with more wlde^iread lateral erosion and deposition in the flooi^lain
than at any other period.
This hypothesis was also partially 8^ppesrted by palynological evidence
from 22It590. m the Early Holocene (Early Archaic) depraits of 22It590,
boreal type pollen was recovered in sev^al saa(>le8, indicating a cxx>l and
moist climate. Sasples selected from oontenporaneous sediments from other
sites investigated did not contain this lype of pollen. Therefore, the boreal
pollen most likely represented a relic oomnuni^ near that site, rather than
predominate conditions. Scarcily and preservation biases of the macro-
botanical remains precludes their use in testing other environnental
hypotheses.
The second paleoenvironnental hypothesis tested ias that the Middle
Holocmne was a period of maxinun post-glacial wazmth, probable dryness, and
landscape stability. This hypothesis was supported by three lines of
evidence: 1) the development of deep polygonal cxacks in the Early Holooene
deposits, 2) a decrease in the sedimentation rate, and 3) a decrease in
seidUment particle size. The soil of the Late Pleistooene/Early Holcxmne
deposits developed deep cxaofcs in the Mid-Holooene. These cracdcs are strong
indicators of a xeric climate. The development of a paleosol in the Early
Holocmne dqiosits also indicates that the landscape had 2dready stabilized
prior to this xeric period. More humid cxmditions about 6,000 B.P. resulted
in severe flooding and erosion of the upper portion of the paleosol down to
the structural B horizcm.
The third area of paleoenvironmental investigation was the development of
anthropic epipedons in the sediments of the three sites investigated: 22It539,
22It576, and 22It590. These soils are distinctive features of the sites and
serve as indicators of long-term habitation. The detailed chemical and
ptysical analysis of these epipedons has produced baseline data for this
umlque product of human cxxxpation. The well-drained, organically enriched
soils of the midden mounds in the f Icxxi^lain have been subjected to an extreme
amount of bioturbation by trees and burrowing animals. This heis modified soil
genesis in these organically rich deposits. Extremely hi^ amounts of
phosphorus and the cxating of grain surfaces with organic minerzd bonding
resulting from human habitation has retarded normal soil development
processes. The black, greasy midden has remained an A horizon for 5,000 -
6,000 years.
From the studies performed on site famaticxi and soil development
prooesses, the following sequence of events cxi flocjdplain midden mound sites
such as 22It539, 22It576, 22It590, 22IT621, 22It623 and 22lt624 is supported
and graphically represented in Table 151 and Figure 58:
393
Figure 57
Contour map of projectile point/knives. Levels 9 - 12,
Block C, 22lt539. (feature 120 blocked out)
STAGE
PHASE
MISSISSIPPIAN
WOODLAND
QULF FORMATIONAL
* Eltimated Dale
Millar
Hanson Springs
Littla Baar Craak
Banton
Sykas-White Springs
EVA-Morrow Mountain
Kirk
Greenbriar
ANTHROSOL
Paiaosol Truncation
PALEOSOL
Blue Clay
Pleistocene
Deposits
Figure 58
Sequence of events of ffoodplnm mtdden mound sites
Dates B.P.
range
nu 151
Cultural chrcoolomr of the Upper \
Stage Perl^ Culture
madlan
Paleo-Indian
Late
12,000-10,000
Archaic
Early
Kirk
Poplar
10,000-7,500
Middle
Eva-Morraw Mountain Hickory
7,400-6,200
7,000
Sylces-White Springs Walnut
Sylces-White Springs/
6,100
Benton
Walnut
6,000-5,840
5,900
Benton
Walnut
5,700-5,300
5,530
Late
Little Bear Creek
5,300-3,500
Gulf
FOonational
Middle
Wheeler
Brolcen
3,500-2,500*
Punplun
Late
Alexander
Henson
2,500
parings
Woodland
Middle
Miller
2,000-700*
Mlssissippian
700-500*
* ^srf-imaf-Ad (all others frcm direcrt: dates) .
1. Initially, the sites viere relatively stable flocxSplain landfonns such as
levees or point bars or outlieirs of fragnents of the valley terraces. !Bie
general floo^lain landscape was relatively unstable, and thick fluvial
deposits were laid down, the xxpper portions of vhich contain cultural
materials and signs of oocvpation during the Early and Middle Archaic
periods (10,000 - 8,500 B. P.).
2. The landscape stabilized during the Eva/Morrow Mountain period (ca.
7,500 - 6,300 B.P.) , and a soil formed with identifiable A, structural B,
and C horizons.
3. Xeric conditions, which caused desiccation cracks and a polygoi^ network
in the soil which permeated the existing structure, ensued during the
Syloes-White Springs/Benton periods (ca. 6,300 - 5,000 B.P.).
4. An onset of hmid conditions (ca. 5,000 - 4,500 B.P.) caused flooding and
scoured off the A and eiposed the B horizons of the paleosol.
5. Deposition resumed, adthough at decreasing rates, until the present.
Cultural materiails contained in the basal deposits of these new sediments
are Sykes-White Springs /Benton.
CULTURAL ADAPTKnONS
Several hypotheses pertaining to the adaptations made by the Archaic
populations to the environnent and resources of the Upper Tcmbi^see Vadley
were tested in this project. The hypotheses were actually on two levels:
those specific to the lithic assanblages reoon/ered and those specific to all
cultural infomation recovered.
LITHIC SPBCTFIC HCTOIHESES
OVERALL HyPOOHESES
The hypotheses addressed the nature of initial settlement, adaptive
changes, and the possible causes of cbemrved differenoes in the archaeological
record during the Middle-Late Archaic and Late Archaic-Gulf Famational
interfaces. Testing the overall hypotheses used all the lines of infomation
produced in the project.
The first hypothesis tested concerned the nature of the initial settlement
and occupation of the TcDobi^aee Vialley (Kirk throuc^ Eva-Morrow Mountain
oocipatians) . It was postulated that the settlement was generalized and
became increasingly refined through tine. The hypothesis vas generally
srpported by docimentation of an increase in density and diversity of cultural
material and site use throu^ the Early and Middle Archaic (Kirk through
Benton) . The four investigated sites of the Early Archaic appear to hatve been
used intermittently, and few site features (10) %mre recovered. The largest
feature type was the chipped stone clusters associated vdth stone tool
manufacture and repair. Site use apparently ««as durt-tem cenpeites
associated with hunting and tool production and maintenance.
The eaepectation that early occupations of the valley would eodtibit use of
qmcialized tools of higher quality raw materials than those of generalized
tools was not supported. The abundant locally available chert was used for
both specialized and general bifacial ly manu&tctured tools, and this pattern
changed little throu^ time. The locally available ferruginous sandstone wus
used for abrading and anvil purposes throughout the Archaic, and it was used
with increasing frequency, e^mcially during the Eva Moirrcw Mountain period.
The secuni set of hypotheses addressed the transition and adaptive changes
during the Middle Archaic (Eva-Morxow Mountain and Sykes-^tiite Icings/
Benton) . Alternative hypotheses ocnceimng the developsent of the intense
oonpation of the midden mount included; 1) development out of the preceding
Eva-Morrow Mountain period, 2) intrusion by an outside group, and 3) diffusion
of ideas frcm an outside area. Studies in^cated that portions of two
hypotheses were svpported. Lithic studies suggest that there is sufficient
continuity in the assenblages of these ocnpcnents to indicate that population
displacement did not occur.
The hypothesis that the change in lithic raw materials used to manufacture
bifaoes in the Benton period was due to lithic rescurce depletion vms not
stpported. Studies conducted in the area indicate that while ocopation of
the floodplain sites did intensify ca. 6,000-7,000 B.P., and floodplain
gravel bars were being buried in the Tcnbi^aee Valley, there were sufficient
sipplies of lithic raw material were available. Hi^ quality raw materials
were available throu^Kut the prehistoric period in the e^qpoeed tributary
stream valleys adjacent to the floodplain.
The change in raw materials occurred during the Benton period when Fort
Payne chert quarry blanks or prefoms inported from the Middle Tennessee
Valley were almost exclusively worked into finished bifaoes. Fort Payne chert
had previously been inported and used, but at a frequency of less than 5%.
Finished tools of Fort Payne in these early periods were rare. IXuring the
Sykes-Nhite Springs/Benton periods, use of Fort Payne chert increased to
ef^pgnacinately 15%. Other tc»ls continued to be manufactured fron local
cherts. The real arisons to post-Benton assemblages could not be
made because of the lack of integral material.
Threads of continuity run throuc^ the eissenblages during the 4,000-5,000
years duration of these Archaic periods. The stylistically sensitive
artifacts (heifted bifaoes) of the Early Archaic throu^ Syfces-Mhite Springs
are relatively similar in method of manufacture and notching technique.
Diffacenoas do exist, however, in the flaking of prefanns. Prefomns for
Graenbrlars and Kiiks required an additional seriM of flakes as ooncered to
Evas, Morrow Mountains, and cypress Credcs. ^oes-Vtiite Springs forms %iere
manufactured in a very similar manner to Greenbriars and Kirks, althou^ they
were the most ocaplex to manufacture.
It «ms hypothasized that the big base canfie of the ^oes-Mhite
Springa/Banton period were occupied year round. The paucity of macrofaotanical
reaadns did not allow confident testing of this hypothesis. The plant remains
from all locations were dominated by charred nutahi^ fragnents, especially
hickory. Other plant remains, eqiecially seeds and cane were recovered at
several sites. These qpring, sunmer, and fall seasonal indicators were not
bultressed by macrobotanical signs of winter oocupanoe. Preservation %«s
poor, however, and the hypothesis of year-round oocupation of the sites cannot
be dismissed.
The lithic studies edso addressed population mobility during the Middle
Archaic. The expected attributes in the lithic assemblages which would
reflect a decrease in mobillly from pre- and post-Benton occupations were not
present. This included attributes relevant to technological efficiency, raw
material use, tool oomplexity, or tool stsaviardi zation . There were some
changes through time seen on the assemblage level, sudi as an increase in
ground stone use and increase in density of lithics. The hi^iest artifacrt
density was present in the Middle Archaic Eva/Morrow Mountain period. The
primary characteristics of the lithic assemblages as measured by the specific
artifact attributes or assenblages at all sites and in ed.1 time periods
examined were HCMOGENEITY and STABILITY.
The third hypothesis tested in the lithic studies was a set of
alternatives addressing the cause of the "Benton phenomenon": trade,
inmigration, or changes in socio/political organization. The studies
supported the hypothesis that trade in lithic raw material , ^ecificedJy Fort
Payne chert, was directly correlated with the S^ces-Mhite Springs/Benton
period. Trade in this raw material had been conducted in limited amounts
prior to this period. The inportation of large numbers of quarry blanks
(preforms) , vAuch were then finished into bifaoes, added a new dimension to
the trade. The quarry blanks were often cached ani showed edge blimting from
travel, and surfaces were polished from rubbing during transportation. In
addition, the associated preforms were often made from the same core or
"block." Although infonnatian on the post-Benton assemblages is very limited,
it appears that trade in these qua3ny blanks decreased markedly, and bifaoes
were again made of local chert.
Alternative hypotheses concerning iinnigration of outside groups and
socio/politiced changes were not supported. Other than a change in raw
material for bifaoe manufacture, homogeneity and continuity were documented
throughout the Arcfaic lithic assemblages.
The next typothesis tested was the nature of the transition from the
Middle Archaic to the Late Archaic (Benton to Little Bear Credc) . TVro
€dtemative hypotheses were tested relating to natural evolutionary change and
398
the departure of the possible intrusive group and a return to the
pre-inbcuslve %«By of life. The intrusive group hypothesis was effectively
dismissed, whereas the second tiypothesis was supported, edthouc^ the one
intact Little Bear Grade ociiponenL provides marginal support. There are
several differences between Benton and Little Bear Grade oenponents, i.e.,
elindnation of inported quarry blanks for bifaoes, change in bifaoe
raanufactore to anall loced cobbles and elimination of the large multi-hearth
prepared areas. The Little Bear Grade oopponent was present at 10 sites, but
the midden was mixed and aoo«±ilaijca could not be identified except at one
site (22lt623/22It624) . At the sites with mixed Late Archaic middens, the
nuaber of Little Bear Greek bifaoes vies usually high, and probably indUrates
both oontinued site use during this period and the pl^ical streni^ of this
type point.
Several research questions were adso adoed in this stu^ on the
descriptive level rather than the explanation level of hypotheses. The first
question focused on the Henson ^srini^ phase oenponent oenmic assenblage
recovered &cm 22It563. The detailed modal analysis docunented that this
assenblage is from the earlier portion of the i^pan of time during which the
Alexander series was produced (ca. 600 B.G.) . Incising motifs were quite
varied in the ceramic assenblage, with 36 variations documented. However,
punctating was dominant and in^cates overlap with earlier deoorative motifs.
Modal analysis provided a data set of an early Alexander ceramic assenblage
that can be used in future conparatlve studies.
The second research question oonoemed the Moodland ceramic sequences.
The small ceramic 2U3aeBblage from 22Mc^31 was characterized as standard Miller
I and seriated to have been made between A.D. 1 and A.D. 300. The Late
Woodland ceramic assenblage from 22It606 was classified as Miller III. This
was one of the first Late Woodland oeramic asaenblages from the Upper
Ttanbicbee Valley to undergo foonal modal analysis. These subphaaes, located
65-100 km (40.3-62 nd) from the defined exanples, exhibited expected and
documented differences. The tapper Ttnbigbee Miller III assenblage dififerenoes
included frequent amoothing of cord-narked surfaces, inclusion of bene along
with grog as a tenpering agent, and a higher peroentmge of shell-tenpeied
pottery. The assenblage was similar to other terminal Wxidland ceramics in
that the major types are the same, and shell- and grog-tenpered pottery have
similar stupes. The radiocarbon dates of this Miller III oeramic assenblage
extends to A.D. 1,200, vhich is 100 years later than those dates in the
central portion of the Ttanbicbee Valley.
MODEL GP ADAPTIVE SYSTQ6 IN THE UPPER TCYBIGBEE VALLEY
(10,000-5,000 B.P.)
The investigations conducted in this project in the tapper Tcnbicbee Vi0.1ey
of the midr-South have produced information vbich can now be integrated and
interpreted into a model of Archaic adaptations and adjuustments. A sunmary
eluci^tion of this model follows.
LATE PAEED-IMDIAN STAGE
(13,000-10,000 B.P.)
The Upper Tcnbi^bee Valley in the period of early post-glacial climate was
deeply incised and then rapidly filled in response to the fall and subsequent
rise in sea level. The floodplain was likely sparsely vegetated vd.th open
areeis of coarse sandy material. The Tcmbigbee was probably a braided stream.
399
BackMoter areas did exist \ilBin suitable for hunan oociciatian,
sodi as point bars, parallel bars, and Iswees along the many streaaiB. The
tanerature mbs cooler and aioister than today, but tnnner that during the
previous periods. Oocqpation was initiated in the UEPsr Tcnbigbee Valley in
this kind of physical environaent during the late Pal^Indian stage.
Evidenoe of this oocupanoe was recovered at four sites investigated 22It539
(two Daltons) , 221t576 (one Quad and two Daltons) , 22ItS90 (five Daltons) , and
221^21 (one Ciaaberland) . Unfortunately, all of these taiporally sensitive
bifaces were found out of context. However, the preeence of these artifacts
does dncuBHnt that this area was oocqpiad during this stage, most likely at
these sites. It is inportant to note that all of the sites containing late
Palao-Indian markers were also oco^ied during the Early Archaic. If presence
and frequency indicate relative age and density of occupation, the Hickory
site (22It621) %tas oocqpied first, followed by the Etsplar site (22It576) . The
Ilex site (22It590) had the most intense Dalton occupation, followed by the
Poplar and Walnut (22It539) sites. The most inportant ai^pects of the late
Faleo-Indian oco^pation are:
1. The selecticn of the floodplain for ooevpation
2. Bis^ correlation with Early Archaic occupation
3. lew density of cultural material
4. The landscape was dynamic and unstable
There are conflicting hypotheses e9q>laining the low nunber of Paleo-Indian
sites. Nuto and Gunn () state that there was severe flooding of the
valley epproodmately 10,000 B.P. vhich effectively sooured awey moet
Paleo-Indian remains. Pettry agrees that a scouring episode oocurxed, but he
places it much earlier - in the Pleistocene period. Pettry predicts thet the
lower deposits of the paleosol deposits, which are culturally sterile, were
laid down in the Late Pleistocene. A major scouring episode was not observed
in the paleosol at the sites investigated, but such soouring oocuned beneath
it. If Pettry is correct, the sediments of Paleo-Indian age, at least at
these sites, were present and had not been eroded enmy. However, Alexander
(b) has dated an erosional episode at 9,030±340 B.P. above 22lt590 and
22T. Based on the materials recovered in this stud^, this occupation was
li^t and was not present below the Early Archaic deposits in the areas
excavated. A brief review of edl sites identified in Itawenba County in the
Tcshli^bee Valley revealed that there were no other Pedeo-Indian conponents
disoovered, and all were in floodplain environnents. This leads to the
hypothesis that Paleo-Indian occupation of the Upper Tcnbi^bee Valley was
minimal and limited to the floo^lain.
In oensidering conflicting Pettry and Muto and Guim theses, the evidence
xeoovexed supports Pettry. Hewever, the original depositicnal context of
these markers was not docunented in this study, and the specimens could have
been curated or obtained fron other sites outside the floocplcdn. fXuiher
research is necessary to resolve the issue.
EARLY ARCHAIC STAGE
(10,000 - 7,500 B.P.)
The environmental oonditions during the Early Archaic period were similar
to thoee of the preceding one. However, during this period, several trends
culminated. First, the floodplain was "filling up" with loose coarse
400
sedijnents, and the local relief was diminishing. Elevations were being buried
under alluvium, vAiile others were beccming better defined. Ihe Tcnbigbee and
larger flooc^lain tributaries began to occi;^ more stable courses. As a
result, resident pqpulaticns were restricted to potential site locaUt.ies
whidi> protruded well above the floo(%>lain. Deposits during this period
ooRtinued to be coarse, sandy loans.
Remains of the Early Archaic were the earliest intact deposits encountered
in these investigations. Ccoponents were present in seven of ihe sites
investigated, five of \diich had intact middens (22It539, 22It576, 22It590,
22It621, and 22It623/22It624) , but one of these was not sampled (22It623/
22It624) . All sites were located on the flooc^lain, and all material was
contained in the sane paleoeol unit. The cultural material consisted
primarily of lithic material, and 17 features (12 pits and five chipped stone
clusters) could be associated (Table 148} . The assemblages at e2K:h site have
individual differences in size and composition. It e^ipears that all were
short-term campsites, flhe chipped stone clusters unique to the Poplar site
appear to be well-preserved remnants of tool manufacture for "retooling”
activities. This inplies that expooed cobbles were available nearby. Such
cobbles are not available toda^, suggesting that the Early Archaic floodplain
is buried.
The activities inferred fron the functionad anadysis of the lithic
assemblages can be associated with hunting, meat processing, stone working,
and wood working. Ground stone is not frequently found in these assemblage.
The matrix of the assemblages is a coarse sandy loam containing little
organic materiad. Charcoal was scarce and limited to charred hickory
nutshells. The sediments developed into a soil with a polygonad cracked and
filled network. However, these characteristics develop^ after the Early
Archaic occupations.
Fran the information recovered &an the Early Archadc d^xjsits the
following refined hypotheses are proposed:
1. The environment was graduadly warming.
2. The landscape was stabilizing through r^ud filling and stream channel
maturity.
3. Lithic ratw materials in the flooc^lain were being buried, and well-dradned
elevations suitable for occupation were decreasing in nuoaber.
4. The Early Archaic populations were the first to inhabit the floo^lain
regularly.
5. The sites were used for short-teim occupation by anall groups.
6. Tool manufacturing, tool kits, and tool function established during this
period were to continue for the duration of the Archaic.
KIIXXE ARCHAIC STAGE
(7,500-5,000 B.P.)
During the Middle Archaic in the Upper Tcmbi^see Valley a change in the
adaptive strategy of the popudatioi is docunented. The landso^m apparently
stabilized during the Eva period (7,500-6,300 B.P.). After stabilization,
xerlc conditions developed. Nesic conditions returned subsequently, and
flooding occurred approodmately 6,300 B.P. throu^iout the floodplain. The
upper portion of the soil was removed, and deposition resumed. This
dqxjsition has continued to the present, althou^ at a diminishing rate.
There are clear indications that cultural adjuustments were made to adapt
to the changing envlrormental and landscape oonditicxis. The sites occupied
401
WL.
during the Early Archaic continued to be ooctpied. However, the use of these
sites changed. Hie dranatic rise in mnbers and kinds of features beginning
in the EvarMorzow Mountain period (Table 146) testifies to those changes. The
33 features include storage, cooking and habitation facilities, v^ork areas,
and burials with grave goods possibly organized in a oanetery, as well as a
cremation containing zoomorphic beads.
This pattern of site use intensified in the following Sykes-White
^srings/Benton period and probably indicates both a larger site population and
longer tasn site use. These d^usits containing these materials rest
unocmfomiably on the truncated paleosol. Seven oonponents were investigated.
TWO were intensely occupied floodplain base camps (22It539 and 22It576) , four
were smaller floo^ladn carps (22ItS90 , 22It623/22It624, 22It622, and
22It621) , and one was an overlodc of the floodplain (22It606) . A total of 159
features were associated with these settlements, far more than with any other
period of oocipation. The features include lar^ refuse pits, prepared hearth
areas (fired aggregates) , large prepared clay with multiple hearths, burials,
and a full range of oth^ feature t^pes. This ooopation is clearly distinct
frcRi any other in the Ujpper Tcnbic^see.
In addition to the dramatic increase in the nunber of features, the matrix
surrounding them was highly altered. The Sjkes-White ^irings/Bentcn deposits
were organically enriched to the extent that they are still dark hrcwn and
greasy. There was an increase in charcoal (especially wood and dharred
hickory nutshells) from almost nothing to up to 13% of the cubic voluoe of the
midden. As the upper paleosol oontaining the Eva Morrow Mountain ocnponent
has been disoolored by organics moving down from above, the original
characteristics of the midden were difficult to measure. It appears, however,
that organics and charred plant material are more abundant than in the
previous Early Archaic midi^, but less so than in the following S^ces-White
Springs/Benton. The erosion of the upper portion of this deposit, which
contained a dark midden (A) and cultural material, precludes a determination
of other possible assemblages.
The hypotheses vhich can now be generated oonoeming the Middle Archaic
occupation in the Upper Tonbigbee Valley follow:
1. The environment oontinued to warm and reached a peak of aridity ca.
7,500-6,300 B.P. Stream regimens and the geanorphic landscape stabilized.
2. Humid conditions ensued initiating a flooding episode vhich scoured even
the hic^iest elevations in the floo^lain, and subsequent deposition has
oontinued to present.
3. Cultural adjustments to these cxxiditions were made by the occupants
initially during the Eva/Morrow Mountain period, and these adjustments
intensified during the Benton period. These chmiges included longer
residence by more people, the use of a few semipermanent main base camps,
and the oocupanoe of smaller sites for more ^pecisdized purposes.
4. These adaptations to the flooc^lain probably relates to increasing warmth
and dry ccnditions of the Altithemal climate episode. The uplands
apparently were not capable of providing the previous subsistence support,
vhich forced (or "pulled") people into the flooc^ladn.
lAIE MCBMC STMaE
(5,000 - 2,500 B.P.)
The envizonment fron tie onset of more hiinid (xnaitlons, «f:pn»iinately
6,300 B.P., signaled tiie asellnratlcn of post-glacial cldjnste. Most
paleoanwrironnental meearrh (Deloourt ; ttiitehsad ) agree that modem
conditions were reiKdisd approKimstely 5,000 years ago. Palaoenvironsental
data acquired in this stui^ cannot onnfinn or deny these assunptions. The
cultural infonnaticn recovered in this project reflects a change in flcoc^lain
site use at this same time.
Site disturbance of poet-Benton cultural material at the sites
investigated calls the nature of tiie transition to the Late Archaic and
following into question. It appears tirnt the initial Late Archaic ocaponent
is Little Beak Grade. However, it is not possible to docment the assuiption.
The stylistic bifaoes of this period %«eze d»ve the Denton ocaponents, but in
nine instanoes, they were mixed and contained nmterial frem otiier periods. In
the one site (22It623/22It624) which contained both a Benton and little Beak
Qnek assaablage, there is a amooth traimition between them.
The Little Bear Grade ocaponent included six faatuzes. Five were refuse
pits similar to those aaaocii^ed with the Denton oonpation at this site, but
there also was a oermonlal blade cache %d.th chert imported from the MidMest,
as well a green slltstone atlatl %ieight. These were not present in any other
asaeablage. Mo major timnges were docimmnted in the litiiic asseahlagBiD
either, and, tizxn all indications, there wes cultural continuity between these
periods.
Site use apparently changed. No longer are the large prepared areas
constructed nor does the midte contain as much charcoal from fires. Prepared
hearths, so frequent in the previous periods (40) , were not found, nor were
burials enoountered frceii this period. The availghle limited data suggest that
there was an adjustment to the improving conditions and an increased mobility,
although the flooc^latn area continued to be used and oocipled.
The hypotheses concerning the adaptations or adjustments made by the Late
Archaic population in the Upper Tosblgbee Valley follow:
1. Tim onset of more inesic conditions following the xeric Altithennal period
reduced the need for the nucleated settlement pattern of the Middle
Archaic. Populations became more dispersed, and site vise was more
homogeneous.
2. Subsistence resources on the valley terraces and in the vplands became
available again and were incorporated into the subsistence round.
3. Ceremonial practices changed to include the use and destruction of exotic
bifaoes.
GUIF FGPMATIGNAL STMZ
(2,500-2,000 B.P.)
By the time of the Gulf Fopnational, modem environmental conditions had
beocrae established, and the Upper Ttnbigbee Valley had a physical, biological,
and climmtological milieu, much like extant conditions. TUn oonpcxients of the
Gulf Formetional stage were enoountered in this stud^. These included both
Wheeler (middle) and Henson Springs (late horizon ocaponents) . One
undisturbed Henson tarings phase midden was identified at the Aralia site
(22It563) , and 17 features at five sites Mare also recxvezed. Die Aralia site
pfTcvided the primary source of infatnation for this period. The infarmation
frcm this assenblage indicated that Aralia mbs a base camp at tdiich multiple
activities took place. These included stone tool manufacture, food
processing, and Mood and hide ptooessing. Stone tool manufetcturing techniques
noir includ^ the splitting of snail oabbh^ by tiie anvil technique to
manufacture bi^KXS. Separate activity areas Mere identified by midden stains
and ooncentrations of different cultural materials.
Continuity with previous periods was docmented by all material gathered
on the Gulf Pormational stage. An increased reliance on seeds and acorns,
which are considered a seoond-line food resources, suggests some stress on
available food resources. Trivial adjustments were surely made to minor
environmental and population level fluctuations.
VCXXliLAIC AND MISSISSIFPIAN STAGES
(2,000-500 B.P.)
Woodland oociqpations were docixnented at all of the sites investigated.
These Included floocplain and first terrace positions as well as an ipland
overlook site. All of the middens had been disturbed, however nine sub-midden
features were encountered at 22lt606, and two were identified at 22It576. In
eidditlon, a burial mound from the Middle Woodland stage was investigated
(22M) .
The burial mound provided the only information on the Middle Woodland. It
contained a Miller I ceramic assaiblage vhich %«as similar to that of the
Central Tonbic^oee Valley. The actual burials were not encountered, since
historic buries intruded. Hence, little information on this period was
reocvered.
Information on the Late Woodland period was recovered from 22It606. It
appears that maize, althou^ present, was not a diet staple of those
occupants. Wild foods, siMl^ to those of previous periods, was the dominant
form of subsistence remains. The ceramic assemblage was identified as Miller
III and was different in surface treatment and tenpering from those of the
central part of the vallQ^. Radiocarbon dates also suggest that these ceramic
types were in use for at least 100 years longer than further downstream in the
central valley.
The Mlssissippian occupation of the Tombic^oee Valley was not well
represented, although shell-tenpered pottery was recovered at six sites. This
stud^ provided little information on the activities conducted during this
stage in the Ipper Tombi^bee Valley. In addition, the knowledge of these
later stages of prehistoric occupation of the Upp^ Tcnbigbee Valley is also
limited by the lack of intact deposits. There is a fertile field for future
research to address the nature of the use of vallQ^ resources in relation to
the large agricultural Mlssissippian villages in the uplands near Tupelo.
RBCOItiEMDATICMS FOR FUTURE RESEARCH
Studies perfarmed as a part of the salvage excavations have yielded
significant data. The adaptive processes employed by Archaic populations are
well documented and can serve as a basis for further studies into
understanding the relationship of human oocupenoe to changing climatic
conditions. A stud^ of pr^iistoric adaptations in adjacent upland areas is
essential to provide oonparisons and contrasts with floodplain occupance.
404
Although archaeologists have long disregarded materials which are
doconented to contain mixed ten{x>ral marioers, as was done in this stucty, it is
essential to initiate investigation of such sites. Available information
could be ocrqpared to that from the Archaic. Such tasporally mixed deposits
contained the bulk of the recovered material , and iaportant information is
contained in than. Ihe presenoe of oeraidcs, the oonpressed tine of ceramic
technology, and surface treatments actually prediqaoees these archaeological
deposits to be described as "disturbed” by the archaeologist. Archaic
dBfXJsits eodiiblt an oiposite effect. Ihe stylistic bifac»8, which are
tenporal markers, have shown long periods of use, and the reasons for change
are still uncertain. The Archaic cxmponents, ihich span thousancis of years,
TOBcg be more "mixed” than Wtodland or Mississippian ccnponents vhlch span
hundbreds of years. The thousands of available ^leciin^ from these later
stages diould be intensively studied.
Infonnation fron all the midden sounds in the Tcmbigfaee and Tennessee
florxiplains should be cxxnpiled and analyzed. A cursory review of the
excavation reports reveals striking similarities over a 7,000 year span.
Analysis of several studies vaa^ yield broadly applicsable generalizations
oonceming site selection and use in floo(%>lain milieux cTver protracted
periods.
Ihe Benton period in the Middle Tennessee and Ujpper Tombi^^bee Valleys
should be intensively studied. The Walnut phase in the Tcxnbichee and the
Seven Mile phase in the Tennessee Valley are cjuite similar. They appear to
have been involved in heavy trading of Fort Payne chert. The nature of trade
ocnmodlties, volune, and intensity may be revealing of differences or
similarities between the two areas.
One of the most fruitful areas for future research involves further
analysis of information edreaciy produced. Far more ciata were prociucjed than
could be analyzed within the soope of this project. Exanples for further
analysis incltide statistical analysis of the ceramic data and ^xitial analysis
of the midden and feature data.
These collections are capable of answering maiy additional (juesticns on
the nature of the prdUxtoric adaptations in the mid-Scxtth. The nddden mound
materials have cxntributed uniquely to the void in our understanding of early
human occupatixxi of this region. ArtifeK±s recxTvered can provide the grist
for more definitive interpretations in future reseeux:h.
i5
inferences
Abu-AgwB, F.
Paleosols in the buried stream valleys of the loessial region of
Mississi|]{>i. PhD. dissertation. Agroncny Department, Mississippi
State University.
Adovasio, J.M. , J. Donahue, H.B. Rollins, J.L. Yedlowski, and R.C. Carlisle
Data reoovery of five rock shelters on the Tennessee-Tcnbic^see
Waterway. Paper presented at the Tennessee-Tombigbee Coordination
Meeting, Coluniius.
Abler, Stanley A.
Proiectile point fom and func^on at Rogers Shelter. Missouri
Ardiaeolo^cal Society Research Series, 8.
Pattern and variety in extended ooalescent lithic teohnology. PhD.
dissertation. Department of Anthropology, University of Missouri.
Colvnfcia University Microfilms, Ann Azix>r.
Stone Itool Analysis in the Elbee. Report submitted to the National
Park Service.
Abler, Stanley A. and Robert C. Christensen
A pilot stuc^ of Knife River flint procurement and reduction at Site
32DD508, a quarry and workshop location in Dunn County, North DaJcota.
Contribution 186, Department of Anthropology and Archaeology,
University of North Dakota, Grand Forks.
Alexander, Lawrence S.
a Phase I ychaeoloqic^ Recormaissanoe along tte Tusctefeia J^ver
Water^ied, Aloom and Prentijs OountiiM, M^lssiippE office of
Arciiaeologlcal Research, Uhiversity of Alabaroa. Report of
Investigations 40.
Organic soil carbon by reduction of chronic acid. Soil Science
40:311-320.
Amick, Daniel S.
Lithic raw roateriad vari^d>ility in the centred Duck River Basin:
reflections of Middle and Late Archaic organizational strategies.
Unpublished Master's thesis. Department of Anthropology, University
of Tennessee, Knoxville.
Anderson, David G.
Prehistoric selectioi for intentional thermal alteration: tests of a
model enploying southeastern archaeologicEd. materials.
Midoontinental Journal of Archaeology 4:221-254.
28:193-99.
Antevs, Ernst
Late Quaternary Climates in Arizona. American Antiquity
Asch, N.B. , R.I. Ford, and D.L. Asch
PaOeoetfancbotany of the Roster site: the Archaic horizons. Illinois
State Museixn. Report of Investigations 24.
Association of Official Analytical Chemist (AOAC) .
Official methods of analysis, 12th edition. Sec. 2.019-2.036.
Atkins, Steve and Jeannie MacMahan
The Zabski site, Merritt Island, Florida. Florida Anthropologist
20:133-145.
Atkinson, James R.
^:pendix A: Test excavations at the Vau^m Mound site. In
Arch^loqical Survey and T^t Excavatioarts in the Uiyer-Central
Tcnbiqbee Riveg^alley: Alioeville-toluDp£us Lo^ and Dam and
Duxuncknent Areas, Alabama and Mississippi, edited by Marc D. Rucker,
FP. 115-158. Report submitted to the Mationed. Park Service the
Department of Anthropology, Mississippi State University.
A cultural rescwroes survey of selected oonstruction areas in the
Tennessee-T^icfcee Waterway. Alabama and Miss^ippi, vol. 1.
Report suhmitted~to the U.S. Amy Corps of Engine^s, Mobile District
by the D^>artment of Anthropology, Mississippi State University.
Atkinson, James R. and Jack D. Elliott, Jr.
A Cultural Resources Survey of Selected Construction Areas in the
Tennessee-Topfcic^jee Waterway; Alabama a^ Mississiroi. Report
submitted to the U.S. Amy Corps of Qigineers, Mobile District by the
Department of Anthropology, Mississippi State university.
Atkinson, James R. , John C. Phillips and Richard Walling
The Kellogg village site investigations, C^y County, Missmsi^i.
Report submitted to U.S. Amy Corps of Engineers, M^ile Distri^ by
the Department of Anthropology, Mississippi State Ikiiversity.
Bartram, William
Travels of William Bartram, edited by M. Van Doren. Dover Press.
Behn, Jeffrey A. and William Green
Collections from the Copperas Creek site (11-F-lOO) , a multi-
ccmponent site in the Central Illinois Veilley. Rediscovery 2.
Illinois Association for the Advancement of Archaeology.
Bense, Judith A.
Archaeological testing investigations at 58 sites in the River and
Canal Sections of the Tennessee-Tcwfcicj^ Waterway. Report submitted
to the National Peurk Service by the Office ot Archaeologiccil
Research, university of Alabama.
a Archaeological Investigations at Site 22It581, Itawairba County,
Mississippi. Report submitted to the National Park Service by the
Office of Archaeological Research, University of Alabama.
b Arcbaeo
Willow smoke and dogs' tails; huntex>-gatherer settlement systems
and archaeologiced site formation. Anerican Antiquity 45:4-20.
Bones; Ancient men and modem myths. Academic Press.
The Archaeology of Plaoe. Journal of Anthropological
Archaeology 1:5-31.
Long-term land use patterns; seme implicaticxis for archaeology. In
L^u T.iivMir Punctated; Essiys in Honor of George Irving Quin^,
edited by R.C. Dunnell and D.K. Grayson, pp. 27-53. Universl^ of
Michigan, Museum of Anthropology.
Binkley, Kenneth M.
Excavation of Eleven Archadc and Woodland Sites in the Divide-Cut
County,
im*
Archaeologlced Research, University of Mississippi.
Blakeman, Crawford, H.
Aixhaeologioal investigations in the
74 seasm. ’Report submitted to the National Park Service by the
Department of Anthropology, Mississippi State University.
976 A cultural resource survey of the jUaerdeen Lock and Dam and Caned
Section emeas of the Tennessee-Toroblgfaee Waterway; . Report
submitted to the National Park Service by the Department of
Anthropology, Mississippi State University.
'Z^9h
Archaeological Report 5.
Brose^ David S.
Functional analysis of stone tools; a cautionfury note on the role of
anijnal fats. American Antiquity 40(1) ;86-93.
Brose, David S., Ned J. Jenkins, and Russell Weisnon
An arcteeologiced reconnaissance of the Black W2un:ior/Lcwer
Tcnisic^Dee Valley and Mobile Delta, Alabama. Report subnitted to
the U.S. Amy Corps of Engineers, Mobile District by the Department
of Geography and Geology, University of South Alabama.
Brown, James A. and Robert K. Vierra
Overview of Koster site rese^u:ch. Paper presented at the 44th
Annual Meeting, Society for American Archaeology, Vancouver.
What happened in the Ardiadc? Introduction to an ecological approach
to Koster site archaeology. In Archaic Hunt^s and Gatherers in the
American MidwMt; Multidisciplinary Studies in the Archaic, "edited hy
James L. Phillips and James A. Brown, pp. 165-196. Academic P*ress.
Bullen, Ripley P.
The Qreuige period of peninsular Florida. In Fiber-tapper^
in the scxttheastem Uhited States and northem Coltnbia; its Origins
Oontext, and Significance, edited by R.P. Bullen and J.B. Stoltman,
pp. 9-33. Florida Anthropological Scx:iety Publications 6.
BuUen, Ripley P. and Adelaide K. Sullen
The Sixnmer Haven site, St. Jc^ins County, Florida.
Anthropologist 14:1-15.
Florida
Sullen Ripley P. and H. Bruoe Greene
Stratigraphic tests at Stallings Island, Georgia.
Florida Anthropologist 23:8-28.
Buntley, G.J., R.B. Daniels, E.E. Gamble and W.T. Brown
Fragipan horizons of the Menf^s-Loring-Grenada sequence in
West Tennessee. Soil Science Society of American Joumed 41:400-407,
Caddell, Gloria May
Plant resources, archaeological plant ranains^ and prehistoric
plant-use patterns in the oenteal Tcnbi^bee River Valley; Part I.
Biocultural studies"!!! the GainSville LaJce area. Report submitted
to the U.S. Any Corps of Engineers, Mobile District by the Office
of Archaeological Research, University of Alabama.
Caldwell, Joseph R.
Trend and tradition in the pr^iistory of the eastern United States.
American Anthropological Association, Memoir 88.
Callahan, Errett
The basics of bifaoe knajping in the eastern fluted point tradition:
a manual for flint knappers and lithic and lithic analysists.
Archeology of Eastern North America 7 (1) :661-683.
Cambron, James W. and David C. Hulse
Handbook of Alabane Archaeolcw, pa
Archaeolo^cal Research Association
part I; point t^s.
ion of Alabama, (revised) .
Carlson, David L.
Hunter-gatherer mobility strategies: an exanple feom the KOster site
in the lower Illinois valley. PhD. dissertation. Northwestern
University, University Microfilms, Ann Arbor.
Chapnan, Jefferson
The Rose Island site and the Bifurcate Point Tradition. Report
submitted to the Tennessee Valley Authority by the Department of
Anthropology, University of Tennessee.
Archaic period research in the Lower Little Tennessee River Valley.
Report si±mitted to the Tennessee Valley Authority by the Deparbnmt
of Anthropology, University of Tennessee.
Chapman, Jefferson, Patricia Cridl^sau^, Lucy Fol^, emd Larry Kimball
Archaic Period Research in the Lower Little Tennessee River Valley.
Report submitted to the Natlohal Paric Service by the Dep>artinent of
Anthropology, University of Tennessee. Report of Investigations 18.
Chapman, Jefferson and James Adovasio
Textile and bemketry inpressions from loehouse Bottom, Tennessee.
American Antiquity 42(4) :620-625.
Christenson, Andrew L.
Maximizing clarity in economic terminology. American Antiquity
47:419-426.
Collins, Michael B.
Lithic technology as a means of prooessual inference. In Lithic
technology: making and using stone tools, edited by Earl Swanson,
pp. 15-34. Aldine, Chicago.
Collins, Michael B. and Jason M. Fenwick
Heat Treating of chert: methods of interpretation and tlieir
application. Plains Anthropologist 19 (64) :134-145.
Connaway, John M. , Samuel 0. MoGahey, aixl Clarence H. Wdob
Teoc Cre^: a Poverty Point site in Carroll County, Mississippi.
Mississippi Department of Archives and History, Azchaeologicel
Report 3. Jackson.
Cotter, Jchn L. and John M. Corbett
Archaeology of the Bynvm Mounds, Mississippi. National Park Service,
Archaeological Research Series 1.
Crabtree, Don E.
Grinding and smoothing of stone artifacts. Tebiwa (17)1:1-47.
Crabtree, Don E. and B. Robert Butler
Notes on experiments in flintknapping: 1. heat treatment of silica
minerals. Tebiwa 7 (1) ;l-6.
Cridelbaugh, Patricia A.
An analysis of the Morrcw Mountain caiponent at the Icehouse Bottom
site and a reassesanent of the Morrow Mountain Ccnplex. Urpublislied
Master's Thesis. Department of Anthropology, University of Tennessee.
Curren, Caleb B.
An archaeological survey of the Warrior Valley in Tuscaloosa County,
Alabama. Site forms on file at Mound State Monument, Moundville.
Davis, R.P. Stephen Jr. , Larry R. Kimball, and William W. Baden (editors)
An archeological survey and assessment of aboriginal settlement
within the lower Little Tennessee River Vedley. Final Report
sutmitted to the Tennessee Valley Authority.
Day, P.R.
Particle fractionation and particle size analysis. In Methods of
Soil Analysis, Peu± I., edited by C.A. Black. Agronomy 9:552-562.
American Society of Agixmcny, Madison, Wisconsin.
DeConick, F., D. Ri^i, J. Mauoorps, and A.M. Robin
Origin and micrcmorphological nomenclature of organic matter in
sandy Spodosols. In Soil Microsccw# edited by G.K. Rutherford,
263-280. The Limestone Press, Kingston, Ontario.
OaJamette, David L. John A. Walthall, and Steve B. Wdmberly
Archaeologiced Investigaticns in the Buttahatchee River Valley II:
excavations at Stacks Bluff shelter. Journal of Alabama
Archaeology 21:99-119.
Oelcourt, H.R.
Late Quaternary vegetation history of the eastern Hi^iland Rim and
adiaoent Cixoberland Plateau of Tennessee. Ecological Monographs
49(3): 255-280.
Delcourt, Paul A. and Hazel R. Deloourt
tote Pl^stooene and Holocene distributioned history of the deciduous
forest in the southeastem United States. Verof^entlidning des
Geobotanischem Institutes der EIH, Stiftung Rubel (Zurich) 68.
Dodd, Walter, A. Jr.
The wear and \ise of battered stone tools at Armijo Rodcshelter.
In Lithic Use-wear Analysis, edited by B. Hayden, pp. 231-242.
Academic Press.
Dragoo, D. W.
Adena and the eastern burial cult. Archaeology of Eastern North
j^merica 4:1-9.
Dye, David H.
An Alexander phase in the Tennessee Valley and adjacent areas.
Paper presented at the 30th Amiu2il Meeting of the Southeastem
Archaeological Conference, tfenphis.
A model of tote Archaic subsistence systan in the Western Middle
Tennessee Valley during the Bluff Cre^ Phase. Tennessee
Anthropolgist 2:63-80.
Primary Forest Efficiency in the Western Middle Tennessee Valley.
Unpublished PhD. dissertation, Washington University.
Dye, David H. and Jerry R. Calm
Alexander, Tchefuncte, and Black Sand: an early Gulf tradition in
the central Tonhic^see Valley. Ps^»r presented at the Third Annual
Meeting of the Mid-South Archaeological Conference, Memphis.
Dye, David H. and Charlotte A. Watrin
Phase I md II Archaeological Investigations at the W.C. Mann site
(22Tb565) , Tiahominqo County, Mississippi. iReport submitted to the
U.S. Amy Corps of Engineers, Nashville District the Department of
Anthropology, Memphis State University.
Earle, Timotliy K. and Andrew L. Christenson (editors)
Modeling change in prehistoric subsistence eooncmies. Academic Press.
Eickmeier, Jan
Use of Carbonized Seeds in Archaeological Seed Analysis.
Missouri Archaeological i
[ Newsletter 283.
Elliott/ Jack D./ Jr.
A cultural resouioes survey of selected construction areas in the
Tennessee-Tcndbii^^Dee Waterway: Alabama and Mississippi, Vol. 2.
Report svdanitted to the U.S. Am^ Carps of Ehgineers, Mobile
District 1:^ the Department of Anthropology, Mississippi State
University.
Ensor, H. Blaine
An evaluation and synthesis of changing lithic tedinologies in
the Central Tcirbic^see Valley. Southeastern Archaeological
Conference, Bulletin 22:83-90.
Lithic morphology, technology, and use the central TOnbic^oee Valley
drainage: the Miller II and Miller III phases. Unpublished Master'
Thesis, Department of Anthropology, University of Alabama.
Gainesville I^e area litMcs: duorolow, tedmoloqy. and use.
Volvme III of archaeological, inivestiaatix »s in th e Gainesville Late
area of the Tennosooo-Ttanbigbee Waterway. Report submitted to the
U.S. Amy Corps of Engineers, Mobile District by the Office of
Archaeological Research, University of Alabama. Report of
Investigations 13.
Fairbanks, Charles H.
The taxonomic position of Stallings Island, Georgia.
American Anticmity 7:223-231.
Fitzpatrick, E.A.
Soils. Longman Inc., New York.
Flennikin, J. Jeffrey
Replicative Systems TUiadysis: A Model Applied to the Vein Quartz
Ar^facts &om the Hoko ^ver Site. Reports of Investigations 59 .
Laboratory of Anthropology, Washington State University.
Ford, James A.
Early Formative Cultures in Georgia and Florida.
American Antiquity 31(6) ;781-799.
Ford, James A. , Ehilip Philips and William G. Hatag
The Jatetown site in west-central Mississippi.
American Museum of Natural History, Anthropoloc
Papers 46:1-164
Ford James A. and George I. Quimky
The Tchefuncte culture, an early occupation in the Lower
Mississippi Vadley. Society for American Archaieology, Memoir 2.
Ford James A. and Clarence H. Wd±>
Poverty Point, a Late Archadc site in Louisiana.
American Museum of Natural History, Anthropologj
Papers 46:1-136
Futato, Eugene M.
Chipped stone bifaoe manufacture in the Bear Creek Watershed.
Southeastern Archaeological Conference Bulletin 22:77-83.
Patterns of lithlc resources utilization in the Cedar Cre^
Reservoir area. Southeastern Archaeology 2(2) : 118-131.
Futato, Eugene M., G.N. Caddell, C. Gayletdiaal-^lavis, H.S. Hale, K.R. l\jmer
Arcfaaeoli^ical investi^tioM in the Cedar Creek auad Ubper Bear Cre^
ReyrwoifT Rqxxrt stimtte^ to the Tennessee Vallsy Aothority by the
Office of Archaeological Research, Ufiiversity of Alabena.
Gagilana, Shenrood M. and Clarence H. Mdab
Archaic-Poverty Point transition at the Pearl River mouth.
Southeastern Archaeological Conference Bulletin 12:47-72.
Galm, Jerry R.
*rte euxhaeoloqy of the Curtis Lake site (34Lf58) , Ltf'lore County,
Oklahcnal Research Series 2. Archaeolo^cal Reseaixh and
Management, Ihiversiiy of Oklahcma.
Prehistoric Cultured Adaptions in the Winter Valley, east-central
Oklahoma. Ur^xiblished £hD. dissertation, Washington State Uhiversity.
Goodyear, Albert C.
A hypothesis of the use of cryptocrystalline raw materi^s amoyr
Pal^IndiMi groups in North Jneri^ Research Manuscript Seri^
i56. Institute of Archeology Anthropology, University of South
Carolina.
Tool kit entropy and bipolar reduction: a study of interassenblage
lithic variability among Paleo-Indian sites in the Northeastern
Ikiited States. Unpublished Ms.
Gould, Richard A.
Ihe anthropology of human residues.
American Anthropologist 80:815-835.
Gregg, Michael L. and Richard J. Graybush
Thermally altered siliceous stone from pr^iistoric context:
intentioned versus unintentional alteration. American Antiquity
41:189-192.
Greenland, D.J., G.R. Linston, and J.P. Quirk
Role of polysaccharides in stabilization of natural soil aggregates*.
Nature (London) 191:-.
Griffin, James B.
Culture Periods in Eastern Ikiited States Archaeology. In Ardmeology
of EMtem Ihited States, edited by J. B. Griffin, pp. 352-364.
^3ie Uhiversity of Chicago Press.
Eastern North Anerican Archaeology: A Simmary. Science 156:175-191.
Eastern ttiited States. In Chronologies in New World Ar^Murology,
edited by R. E. Taylor and C. W. Mercian pp. 51-70. Acacbmic Pr^s.
415
Griffin Jcdm W.
Fiber-taipeied pottery in the Tennessee Vall^. In Fiv>er-tg»^r**^
pott^ in the southeastern Ihited States and iKarthMn Colunbla;~Tts
^iflinf context, and significamoe. edited by R.P. Bullen anri .t,r.
Stoitmai, pp. 34-36. Florida Anthropological Society, Publication 6.
Griffith, M.A.
A pedological investigation of an archaeological site in Ontario,
Canada, I. An examination of the soils in and adjacent to a
fontier village. Geodenna 24:237-336.
Grissinger, E.H. and J.B. Murphy
Present "probleni" of stream channel instability in the bluff area
of northern Mississippi. Jcumed Mississippi Acadeny Science
27:117-118.
Grossman, R.B. aixi F.J. Carlisle
Fragipan soils of the eastern Iftiited States.
Advances Aqronony 21:237-239.
Hanson, Lee H. Jr.
Srgyi^ of Town Creek Waterdied, Dam No. 4^, Lee County,
Missj^TppTT Report submitted to the Soil Conservation Service .
Harlow, William D.
Fruit Key and TWig Key to Trees and Shrubs. Dover, New York.
Harradine, F.F.
Uie variability of soil properties in relaticxi to stage of
profile developnent.
Soil Science Society American Proce*»HTTiqs 14:302-311.
Hayden, Brian (editor)
Lithic use-wear analysis. Academic Press.
Henry, Don 0., C. Vance Haynes and Bruce Bradley
Quantitative variations in flake stone debitage.
Plains Anthropologist 21 (71) ; 57-61 .
Hilgard, E.W.
Report on the geology and agriculture of the state of Mississippi.
JacJcscxi, Mississippi.
Hill, Mary C.
A skeletal analysis of the hunan burials from 9Cla62. In
Cemochechob^: archaeological investigations at the waiter F. George
Dam Mound slte,~9Cla62, Clay County, Georgia, by Frank T. S^iell,
Vem» J. Kni^t, Jr., and Gail S. Schnell. 464-473. Report
submitted to the U.S. Army Corps of Engineers, Mobile District and
the Heritage Conservation and Recreation Service.
Hofinan, Jack L.
Contextual studies of the Middle A^ch^dc cotponent at Cave Springs in
middle Tennessee. Unpublished Master's Thesis. Department of
Anthropology, University of Tennessee, Kixaxville.
416
Hoideauexch, C. , A.R. Hill, D.M. Lapp, and S. Navaratil
Soil and envirocinental analysis of the Bobitaille site. In
Palaeoeoolxxiy and Ontario Prdiistory, edited by W.M. Hurley and C. E.
Hoindenreich. Department of Anthropology, university of Toronto
Research Project 2:179-249.
Hole, F.D.
Some relationship between forest vegetation and Padzol B horizons
in soils of Menominee tribal land, Wisconsin, U.S.A.
Soviet Soil Science 7:714-723.
Holzhey, C.S., R.B. Daniels and E.E. Gamble
Ibick Bn horizons in the North Carolina Plain: II. Physiced.
and ohemiccd properties and rates of organic additions from
surface sources. Soil Science Society American Prooeedinqs
39:-.
Holliday, V.T.
Morphology of late Holocene soils at the Lubbobk Late archaeological
site, Texas. Soil Science Society American Journal 49:938-946.
House, John H.
A functional typology for Cache River project surface collections.
In The Cache River archa^loqical project: an eagpertwant in
oontrart archaeology, edited by John H. House and Michael
B. Schififer. Arkansas Archeological Survey Research Series 8.
Hubbert, Charles M.
A cultural resource suirvey of the Bay Sprii^ segment of tte
Tetmessee-Tcn^ic^e Waterway. Report subnitted to the National
Park Service by the Office of Archaeological Research, University
of Alabama. Report of Investigations 3.
Hudson, Charles
The Southeastern Indians. University of Tennessee Press, Knoxville.
Hugie, V.K. and H.B. Passey
Soil surface patterns of some semiarid soils in northern Utah,
southern Idaho, and northeastern Nevada. Soil Science Society
American Proceedings 28:786-792.
Jackson, M.L.
Soil Chemical Analysis, Advanced Course. Published ]y author.
Soil Science Department university of Wisconsin, Madison.
Jane, F.W.
The Structure of Wood. A. and C. Bleuirk, London.
Jenkins, Ned J.
Subsistence and settlement patterns in the Western Tennessee Valley
during the transitional Archaic-Woodland period. Joumcd. of Alabama
a
b
a
b
Jenkins,
Jenkins,
Jenkins,
Archaeological investigations in the Gainesville Lock and Dam
Reservoir: . Report sufcniitted to the National Park Service
by the Department of Anthropology, University of Alabama.
Ilie Wheeler series and southeeustem prehistoxy.
Florida Anthropologist 18:17-26.
Pr^iistoric chronology of the lower Chattahoochee Valley.
Joujnied of Alabama Archaeology 24:73-91.
cainesville Lake ceramic description and chronology. Archaeological
investigations in the Gainesville Lake of the Ttennessee--^Itent)icfe^
Wtaterv»y. Report subnitted to the U.S. Amy Corps of Engineers,
Mobile District by the Office of Archaeological Research, University
of Alabama. Report of Investigations 12.
Gainesville Late area ceramic description aixl chroaiology. University
of Alalaama, Office of Archaeological Research , Report of
Investigations 11.
Archaeology of the Gainesville Late area; synthesis. Report
submitted to the U.S. Amy Corps of Engineers, Mobile District by the
Office of Archaeological Research, University of Alabama. Report of
Investigations 23.
The Wheeler series; space, time, aM external relationships. Paper
presented at the Third Annual Meeting of the Mid-South Archaeological
Conference, Matphis.
Ceramic chronology along the lower Tonibi^bee River and Mobile Delta
regions. Report on file at the University of South Alabama,
Department of Geology and Geography.
Ned J., Ccdl\:p B. Curren, Jr., and Marc DeLeon
Archaeological site surv^ of the Demoplois and Gainesville Lake
navigation channels and additional construction areas. Report
submitted to the Naticaial Paurk Service by the Department of
Anthropology, University of Alabama.
Ned J. and Richard A. Krause
The Tonbictfjee Watershed in Southeastern Prehistory. The University
of Alabama Press, Thscedoosa.
Ned J. and Christopher J. Peebles
A seriation of Late Middle Woodland-Late Woodland features frcm
the Gainesville Late area. In Archaeology of the Gainesville
Lake area; synthesis by Ned J. Jenkins. Volume V of Archaeological
Investigations in the Gainesville Lake area of the Tennessee-
Tcnbi^j^ Waterway . Report sdbmitted to the U.S. Amy Corps of
Engineers, MobileDistrict by the Office of Archaeological Research,
Lftiiversity of Alabama. Report of Investigations 23.
418
Jemujigs, Jesse D.
Qdckasaw and earlier Indian cultures of northeast Mississippi.
Ilie Journal of Mississippi Histo^ 3:155-226.
Ihe archaeological survey of the Natchez Trace.
American Antiguity 4:408-414.
Jermy, J.
Model of a rising profile in Nile Valley alluvium and its
agronanic and pedogenic inplications. Soil Science Societ
ADnerican Proceedings 26:588-591.
Jochim, Michael
Hunter-gatherer subsistence and settlement: a predictive model.
Academic Press.
Johnson, Jay K.
Lithic Procurement and Utilization Traiectories: Analysis Yellow
Creek Nuclear Power Plant Site
Keel^, Lawrence H.
Techniques and methodology in micrcMear studies: a critical review.
World Archaeology 5(3) :323-337.
tal determination of stone tool use. University of Chicago.
Rafting and retooling: effects cti the archaeological record.
American Antiquity 47:789-809.
Keeley, Lawrence H. and M.H. NevKxmer
Microwear analysis of ejqjeriroental flint tools: a test case.
Journal of Archaeological Sciax:e 4:29-62.
Kelly, Robert L.
Hunter-gatherer mobility strategies. Joum£d of Anthropological
Research 39:277-306.
King, Frances, B.
n.d. Key to Ccramon Dioot Woods of central Illinois. Typescript.
King, J.E. and W.H. Allen, Jr.
A Fiolooene vegetation record from the Mississippi River Valley,
southern Missouri. Quaternary Research 8:307-323.
Klippel, Walter E. and William B. Turner
Prehistory and Holocene land surface changes in the Nashville Basin.
Paper presented at the 48th Annual Meeting of the Society for
American Archaeology, Pittsburgh.
Knudson, Ruthann
Organizational variability in late Paleo-Indian assemblages.
Unpublished PhD. dissertation. Department of Anthropology,
Washington State University.
Lafferty, Robert H. and Carlos Solis
Phaise II of the Bay Springs segnent of the Tennessee-
Tcpfci<;^bee W&terviay. Draft report submitted to Interagency
Archa:TOl^ic2d Se^ioes, Atlanta.
Lzphan, M.H.
Genesis and morfAiology of desert soils. American Soil Survey
Association Bulletin 13:34-52.
Lee, Richard B. and Irving Devore (editors)
Man the Hunter. Aldine, Chiceigo.
Lewis, Sheila D. and Jenes D. Caldwell
Survey of the Tennessee-Torobiqfaee Waterway System, -.
Manuscript on file with the Mississippi Department of Archives
and History, Jackson.
Lewis, Thomas M.N. and Madeline Kneberg Lewis
Eva: an Archaic site. University of Tennessee Press, Knoxville.
Lloyd, Janet R. , Judith A. Bense and Jesse L. Davis
Tedlahatta quartzite quarries in the Escainbia River Drainage.
Journal of Alabama Archaeology 29(2) : 125-141.
Lcwe, EEhraim N.
Mississippi: its geology, geogregahy, soils, emd minered resources.
Mississi^i State Geol^ical Surv^ Bulletin 12.
Plants in Mississippi. Mississippi State Geological Bulletin 17.
Lurie, Rochelle
a Economic models of stone tool manufacture and use: the Koster
site. Middle Archedc. PhD. dissertation. Department of Anthropology
Northwestern University, University Microfilms, Ann Arbor.
420
b A functicxial analysis of artifacts fron sene pre-oeramic oenponents
In PreJi^tory of the Ayacucho Basin (vol 4) by R.S. MacNeish,
A. Nelkin-Tisner, R.K. Vierra, R. Lurie, and A.G. Cook.
University of Michigan Press, Ann Arbor.
Raw material selection and reduction at the Oak Forest site.
Paper presented at the 49th Annual Meeting of the Society for
American Archaeology, Portland.
n.d. Lithic technology and mobiliQr strategies: Ihe Koster Middle Archaic.
In Time, Energy amd Stmie Tools, edited by Robin Torrence, Cambridge
University Press.
Lutz, H.J.
Ihe conoentraion of certain chemical elements in soils of Alaskan
archaeological sites, itoerican Journal Scimice 249:925-928.
MoCenb, A.D. and W.E. Locmis
Subclimax prairie. Bulletin Torrey Botanical Club 71:46-47.
MoGahey, Samuel O.
^chaeoloqical survey in the Tcmbiqfaee River Drainage area, May-June
. Mississippi Archaeological Survey, Prelimina^ Report 2.
McKeague, J.A.
Humic-fluvic acid ratio, Al, Fe and C in pyrophosphate extracts as
criteria of A and B horizons. Canadian Journal Soil Science 48:27-35.
Mc±lichael, Edward V. and James H. Kellar
Archaeological salvage in Oliver Basin. University of Georgia,
Laboratory of Archaeology Series, Report 2.
Magne, Martin and David Pokotylo
A pilot stu^ in bifacial lithic reduction sequences.
Lithic Technology 10(2-3) : 34-47.
Mandeville, M.D. and J. Jeffery FlenniJcen
A oenparison of the flaking properties of Meha»4ca chert before and
after thermal pretreatment. Plains Anthropologist 19(64) : 146-148.
Marshall, Richeuid A. and John T. Glover
Archaeological survey of Tishomingo State Park and environs,
Tishomingo County, Mississippi. Department of Anthropology,
Mississi^i State University.
Martin, Alexander C. and W.D. Barkley
Seed Identification Manual, university of California Press, Ber)celey.
Mehringer, Peter J., Jr.
Pollen analysis of the Tula Springs area, Nevada. Nevada State
Museum Anthropology Papers 13:129-200.
Meixner, R.E. and M.J. Singer
Phosphorus fractions from a chronosequenoe of alluvicLL soils.
San Joaquin Valley, Califomia. Soil Science 139:37-46.
Menzel, R.G.
Qurichnent ratios for water quality modeling. In CREAMS: A
Field-Scale Model for Chemicals, Runoff, and Erosion from Agriculture
Management Systems. U.S. Agriculture^. Conservation Resources Report.
Moo^, Ula L.
Phytoliths as an interpretive device of paleoenvironments in
archaeological sites. Unpublished Master's Thesis. Department of
Geology, University of Montana.
Morrow, Carol A. and Richard W. Jeffries
Prehistoric chert exploitation at the Black Eeuoth site. Carrier
Mills, Illinois. Paper presented at the Conference on Prehistoric
Chert Exploitation, Carbondale, Illinois.
Curational theory as a means of deterndning the role of lithic
material in cultural systems. Paper presented at the 47th Annual
Meeting, Society for American Archaeology.
Morse, D.F.
The Robinson site and shell mound Archaic cultures in the middle
South. Unpublished EhD. dissertation. University of Michigan.
Muto, Guy R. and Joel Gunn
A study of Late Quaternary enviranments and ]^ly Man along the
Towbic^ee Mvec, Alabama aaad Mississippi. Final report submitted
to the National Park Service by Benham Blair and Affiliates,
C^clahoma City.
Nevxxmer, M.H. and L.H. Keeley
Testing a nethos of micxcwear analysis with experimental flint tools.
In Lithic Use-Wear Analysis, edited by Brian Hayden, pp. 195-206.
Academic Press.
Nielsen, Jerry J. and Ned J. Jenkins
Archaeological investigations in the Gainesville Lock and Dam
Reservoir: . Report submitted to the National Park Service,
Tallahassee.
Nielsen, Jerry J. and Charles W. Moorehead
Archaeological investigations in the Gainesville Lock and Dam
Reservo^, Tennesaee-ToBfei^ee Waterway. Report submitted to
the Natio^l Park Service, Tallahassee.
Oakley, Carey B. and Eugene M. Futato
Archaeological Investigations in the Little Bear Creek Reservoir.
Rqx>rt submitted to the Tennessee Valley Authority by the Office of
Archaeological Reseeuch, University of Alabama. Research Series 1.
Odell, George H.
l4icxo-weeur in perspec±ive: a synpathetic response to Lawretxe
H. Keeley. World Archaeoloqy 7(2) : 226-239.
Hie application of ndcro-viear analysis to the lithic oonponent of
an entire prehistoric settlement: methods, problems and functional
reconstruction. PhD dissertation. Harvard University, University
Microfilms, Ann Arbor.
A new and inproved system for the retrieval of functional
information &om microscopic observation of chipped stone tools.
In Lithic Use-Wear Analysis, edited by Brian Hsyden, pp. 329-344
Academic Press.
Toward a more behavorial approach to archaeological lithic
concentrations. American Antiquity 45:404-431.
A diachronic, view of lithic assemblages &om a stratified site
in the lower Illinois valley. Paper presented at the 47th Annual
Meeting, Society of American Archaeology, Minneapolis.
Odell, George and Frieda Odell-Vereeken
Verifying the reliability of lithic use-wear assessments by 'blind
tests' : the low-power approach. Journal of Field Archaeoloqy
7:87-120.
O'Hear, John W.
The W.C. Mam site (22lt565) . Report submitted to the Natimal Park
Service by the Department of Anthropology, Mississippi State
tftiiversity.
O'Hear, John W. and Ihomas L. Com
Archaeoloaical ssdvaoe excavations at the L.A. Strickland I site
Tennessee-Tcnox
An anthi
John Wi]
iloqical anal\
, New York.
3is of
Otinger, Jeffrey L. and Robert H. Lafferty, III
Hie depositions^, iaplications of Archaic structures at the Brinkley
Midden, Tishomingo County, Mississippi* Southeastern Archaeological
Conference Bulletin 22:101-110.
Otinger, Jeffery L., C.M. Hoffman, and R.H. Lafferty, III
The Brinkley site (22Ts729) ; archaeological investiqaticmis in the
Yellow Creek watershed, Tishimingo County, Mississi^i. Office of
Archaeological Research, University of Alabama.
Parker, Malcolm
The Owl Creek People: A Preliminary Report on the Hart site,
40 VM 14.
Pcursons, R.B. , W.H. Scholtes, and F.F. Rickens
Soils of Indian Mounds in northeastern Iowa as benchmarks for
studies of soil genesis. Soil Science Society American Proceedings
26:491-496.
Patterson, L.W.
Quantitative characteristics of debitage from heat-treated chert.
Plains Anthropologist 24(85) : 255-260.
Patterson, L.W. and J.B. Sollberger
Replication and classification of small-size lithic debitage.
Plains Anthropologist 23(80) : 103-112.
Peebles, Christopher S.
Determinates of settlement size and locaticai in the Moundville phase.
In Mississippian settlement patterns, edited by Bruce D. Smith.
Academic Press.
An overview of research of the Lubbub Creek archaeological locality.
^ Eycava^ons in the Lubbub Creeic archaeological loc^ty:
Prdiistoric agricultural ooninunities in West Central Alabama.
Volune edited ly C.S. Peebles, pp. 70-129. Final report submitted
to the Interagency Archaeological Services and U.S. Amy Corps of
Engineers, Mchile.
Peech, M.
Exchange acidity. In Methods of Soil Analysis, Part I.,
edited by C.A. Black. Agronomy 9:914-926. American Society of
Agronaiy, Madison.
Perman, John T.
ArchaeologicctL Survey in Mississippi, -. Mississippi
Department of Archives and History, Archaeological Report 2.
Pertulla, Timothy
Patterns of prehistoric lithic raw material utilization in the
Caddoan area: the western Gulf Coast plain. Paper presented at
the Conference on Prehistoric Chert Exploitation, Ceuhondale,
Illinois.
Peterson, Drexel A., Jr.
n.d. W.C. Mann site, 22Ts565 . Edited by David Dye. Final report
submitted to U.S. Amy Corps of Engineers, Nashville District by
the Department of Anthropology, Mississippi State IMiversity.
Peterson, G.W. and R.B. Cox^
A modified Chang and Jackson pccoedure for routine fractionation
of inorganic soil phosphates. Soil Science Society American
Proceedings 30:563-565.
Fettxy, D.E.
Soil Rerouroe Areas of Mississ^
Mississippi Agricultural Experii
State University.
L. Infomation Sheet 178,
ital Station, Mississippi
Phelps, David S.
IhcRias Creek ceramics in the central Savannah River locality.
Florida Anthropologist 21:17-30.
Pielou, E.C.
Ecol
jiccd Diversity. John Wiley and Sons, New York.
Phillips, Philip
Archaeological survey in the lower Ycizoo basin, Mississippi,
-. Harvard University, Peaboc^ Museum of American
Archaeology and Ethnology, Paper 25.
Purdy, Barbara A.
Investigations oonoeming the thermal alteration of silica minerals:
an archaeologiced approach. PhD dissertation. University of Florida,
University Microfilms, Ann Arbor.
Investigations oc»noeming the thernal alteration of silica minerals:
an archaeological e^roach. Tdoiwa 17(1): 37-66.
Purdy,
Baidsara A. and J.K. Brooks
Thermal alteration if silica minerals: an archeologic2d. approach.
Science 173:322-325.
Raab, L. Mark, Robert Cande and David W. Stable
Dd^itage graptis and Archaic settlement patterns in the ArJcansas
Ozarks. Midoontinental Jounal of Archaeology 4(2) : 167-182.
Radford, Albert E. , H.E. Ahles and C.R Bell
Manual of the Vascular Flora of the Carolinas.
Carolina Press, Chapel Hill.
Iftiiversity of North
Rafferty, Janet E., B. Lea Baker, and Jack D. Elliott, Jr.
Archaeological investigaticns at the East Aberde^ site (22Mo819)
_ multi-resource dlitelct, AlabanwTand M^sissippIT
Final report submitted 'to the National Park Service.
Rands, Robert L.
The Wells site: Poverty Point-Woodland remains on the Pearl River
Mississippi. Southeastern Archaeological Oonferenoe, Newsletter
6:15-18.
Reidhead, Van A.
Optimization and food procunenent at the prehistoric Leonard Haag
site, southeastern Indiana: a linear approach. PhD. dissertation,
Indiana IJhiversity, University Microfilms, Ann Arbor.
Ihe eccnonies of subsistence cdiange: a test of and optimizatiCTi
model. In Modeling change in prehistoric subsistence eooncmies,
edited by Tomothy R. Earle and Andrew L. Christenson, pp. 141-186.
Rick, John Winfield
Heat-edtered cherts of the lower Illinois vcdley. Northwestern
University Archaeological Progran, Prehistoric Records 2.
Rovner, Irwin
Potential of opal fAytoliths for use in paleoeoologic^d
reconstruction. Quantemary Research 1:343-359.
Rucker, Marc D.
Archaeological survey and test excavations in the Upper-Central
Ttanbic^aee River Valley : Alioeville-tolrwfi^ Lock and Djan
Inpoundaimt areas, Aiifcama and Mississippi. Report subnitted to
the National Park Service by the Department of Anthropology,
Mississippi State University.
Ruhe, R.V., R.D. Hall, and A.P. Canepa
Sangamon pedeosols of Southwestern Indiana, U.S.A.
Geoderma 12:191-200.
Schiffer, Michael B.
Behavorial Archaeology. Academic Press.
Sears, William H. and James B. Griffin
Fiber-tempered pottery of the southeeist. In Prehistoric Pottery
of the eastern United States, edited by J.B. Griffin.
University of Michigan, Museum of Anthropology.
Sheets, Payson
Behavorial analysis and the structure of a prdiistoric industry.
Current Anthropology 16(3) : 369-391.
J. Richard
Oak Island archaeology: prehistoric estuarine adaptations in the
Mississippi River delta. University of New Orleans, Archaeologiccil
and Cultural Research Program. Rf^»rt submitted to Jean Lafitte
National Historical Park.
Shenkel, J. Richard and George Holley
A Tchefuncte house. Southeastern Archaeological Conference
Bulletin 18:226-242.
Sheldon, Elisabeth S.
Botanical Remains frcm Archaic Period Sites in the Upper Tcnibigbee
Valley. Paper presented at the annual meeting of Society for
Econcmic Botany, Boston.
Shenkel,
426
Shoot, Michael J.
Forager Mcbilil^ and TechnologiraO. Organization. Paper presented
at the 49th Annual Meeting of the Society for American Arcdiaeology,
Portland.
Sknith, Brent W.
A Late Archaic-Poverty Point steatite trade network in the lower
Mississippi Valley: some preliminary observations. Florida
Anthropologist 34:120-125.
Sknith, Bruce D.
Prehiet^ic Patterns of Iftnai Behavior: A Case Study in the
Mississippi Valley. Academic Press.
Sknith, Eric Alden
Human adaptation and energy efficiency. Human Ecology 7(2) : 53-74.
Soil Taxonomy
O.S. pepartment Agriculture Handbook 436. U.S. Govenment
Printing Office, Washington, D.C. pp. 14-17.
Solis, Carlos and Richard Walling
Archaeological Inyes^gations at the Yauchorouch site (22C)
Clay Ooun^, Mississippi. Report submitted to the U.S. Amy Corps
of Bigine^, Mobile District by the Office of Archaeological
Research, University of Alabama.
Springer, M.E.
Desert pavement and vesicular layers of some soils of the desert of
Lohotan Basin, Nevada. Soil Science American Proceedings 22:63-55.
Stephenson, L.Q. and W.H. Monroe
The Upper Cretaoeaxis Deposits. Mississippi State Geological.
Survey Bulletin 40:61-94.
Stephens, C.G.
Climate as a factor of soil formation through the Quartemary.
Soil Science 99:9-14.
Stephenson, L.Q. and W.H. Monroe
Tte Upper Creataoeous deposits. Mississippi State Geological Survey
Bulletin 40.
Steponaitis, Vincas P.
Some preliminary chronological and technological, notes on Moundville
pottery. Southeastern Archaeologicad. Conference Bulletin 22:46-51.
Ceramics, Chronology, yid Ccnmunity Patterns: An Archaeological Study
at Moundville. Academic Press.
Stevenson, Marc G.
The formation of artifact assarblages at vrorkshop/habitation sites:
models from Peace Point in northern Alberta. American Antiguity
50(1) :63-81.
427
steward, Julian H.
theory of culture chcmget the methodology of multilinear evoL^don.
Uhiv^sTty of Illinois Press , Urbana.
Stoltman, Janes B.
!Ihe Late Archaic in the Savannah River regicn. In Fiber-tenpexed
Pottery ^ the southeastern United States and northern ColtroiaT"
its orlcq^, context, and significance, edited by R.P. Bullen and
J.B. Stoltm^, pp. 37-62. Florida Anthropological Society
Publication 6.
Groton Plantation; an auxihaeological study of a South Carolina
loccLlity. Hcurvard IMiversity, Peetbofy Museixn of American
Archaeology and Ethnology, Morx^gre^ 4.
Swanton, John R.
Indians of the southeastern United States. Bureau of American
Ethnology Bulletin 137.
Symonds, George W.O.
Tree Identification Book. William Morrow and Co. , Inc. , New York.
Tan, K.H.
Principles of Soil Chemistry. Marcel Dekker, Inc. , New York,
pp. 41-131.
Tan, K.H. and J.E. Giddens
Molecular weights and spectral characteristics of humic and fluvic
acids. Geoderma 8:221-229.
Ihonas, David H.
Archaeology's operational imperative: Great Basin projectile points
as a test case. UCLA Archeology Survey, Annual Report 12:31-60.
Archaeology. Holt, Rinehart, and Winston, New York.
Thomas, Prentice H. Jr., L.J. Campbell, C. S. Weed, M.T. Swanson, and
K. Bagley-Baumgartner
Archaeological Investigations at the Turtle Pond site (22It643)
Itawamba County, Mississippi. Report subnitted to the U.S. Army
Corps of Engineers by New world Research, Inc.
Thome, Robert M.
A cultural resouroe survey of the Divide-Qit section, Tennessee-
Tcmbichee~waterway . Tishcroin^ County, MissiBjsippi; .
Report s\±mitted to the National Peurk Service ty the Department of
Sociology and Anthropology, Iftiiversity of Mississippi.
Thorp, J.
The nature of the pedological record in the Qucurtemary,
Soil Science 99:1-8.
428
Tarxenoe, Robin
Time budgeting and hunter-gathexer tedmology. In Hunt
eoanomy in prdiistory: a EuropeMt perspective, editecT
Bailey, pp. 11'‘22. Cantoidge Uhiv^sity Press.
-gatherer
Geo?f
fnrickey, E. Bruce and Nicholas H. Holmes, Jr.
A chronologiced framework for the Mobile Bay region, (revised,
) . Journal of Archaeology 17:115-128.
Tringham, Ruth, G. Cooper, G. Odell, B. Vbytek, and A. Whitman
E>qperimentatim in the formation of edge damage: a new aqpproach
to lithic analysis. Joumed of Field Archaeology 1(1-2) : 171-196.
USDA Soil Conservation Service
Soil Supey Laboratory Methods. Soil Survey Investigations
R^xart 1. O.S. Govemnent Printing Office, Washington, D.C.
USDA Soil Surv^ Staff
So^ Survey Manual. Agricultural Handbook 18. U.S. Govemnent
Printing Office, Washington, D.C.
University of Alabama Miseums
A preliminary archaeological survey of the propoaed Gainesville
Lo^ and Dm Reservoir ^ the TcnEj^ee^RTver. Report submitted
to the National Paid: Service, Tallahas^.
Waddell, Eugene G.
Thomas Creek Punctate. Southeastern Archaeological Conference
Newsletter 9:3-5.
Walker, T.W.
Ihe significance of phosphorus in pedogenesis. In Experimaritad.
Pedology, edited by E.G. Hallsworth and D.V. Crawford, pp. 295-315.
Butterworths, London.
Walker, T.W. and J.K. Syers
Ihe fate of phosphorus during pedogensis. Geoderma 15:1-19.
Wadthall, John A.
Krdiistqric Indians in the Southeast, Archaeology of Alabama and
the Mlddie~South. University of Aledsaroa Press, University.
Walthall, John A., Clarence H. Wdsb, Stephen H. Stowe and Sharon I. Goad
Galena analysis and Poverty Point trade. Midoontinental Journal
of Archaeology 7:133-149.
Wsdthall, Jchn A. and D.L. DeJamette
Copena burial caves. Journal of Alabama Archaeolc
20:1-59.
Walthall, John A. and Ned J. Jenkins
Ihe Gulf Formational stage in southeastram prehistory.
Southeastern Archaeological Con^xenoe Bulletin 19:43-49.
429
LI
WBtacn, Patty Jo
In pursuit of prehistoric subsistence: a ooRparative account of sane
oontoiporaxy flotation techniques. Midocntinental JourmJ. of
Archaeology 1:77-100.
Mehb, Clairence H.
The Poverty Point culture. Louisana State University,
School of Geoscience and Man 17. (seccnd edition, reused) .
M^}b, Clarence H. , Janes A. Ford and Sherviood Gagliano
Poverty Point and the American formative. Ms. on file. Department of
Sociology and Anthropology, Auburn University, Montgcnery.
Wetb, Williem S. and David L. DeJamette
^ archaeol^iczd survey of Picikwick basin in the adjacent
paeons ot the stetes of Alabama, Mississippi, and Tennessee.
Smithsonian Institution, Bureau of Ethnology, BuTletin 129.
Weigel, R.D. , J.A. Holman and A.A. Palounpis
Vertebrates from Bussell Cave. In Investigations in Russell Cave,
edited Yr/ J.W. Griffin. National Park Ser^oe Pi±iircations in
Archaeology 13.
Weinstein, Richard A.
Arch^lxxrical Investigations along Moore's Credc, Alcorn County,
Mississippi. Report submitted to the itotional Park Service,
Atlanta. Coastal Ehvironments, Baton Rouge, La.
Weissner, P.
Style and social information in Kalahari San projectile points.
American Antiquity 48(2) : 253-276.
White, Nancy M.
Archaeological Survey at Lake Seminole. Report submitted to the
U.S. Army Corps of Engineers, Mobile District by the Cleveland Museum
of Natural History. Archaeological Research Report 29.
White, Nancy M. , Chung Ho Lee, and J.A. Bense
Archaeological Investigations in the Upper Tcrobi^feee Valley:
Mississippi: II « Report submitted to the U.S. Army Corps of
Engineers, Mobile District the Office of Cultural and
Archaeolo^cal Research, IMiversity of West Florida. Report of
Investigations 4.
Williams, Stephen
The WSring papers: the collected marks of iVitonio J. Wtoing, Jr.
Harvard University, Peabody Museun of American Archaeology amd
Ethnology, Paper 58.
Williame, Stephen and J. B. Stoltman
An outline of southeastern United States prehistory with particular
enphasis on the Paleo-Indian era. In The Quartemary of the United
States, edited by H. E . Wri^t, Jr. and David G. Frey. Princeton
university Press, New Jersey.
Winberly, Steve B.
Indian po
_ , frcpi Clarke County aid Mobile County, southern
Alabaneu Alaivima Museum of Natural History, Misean Paper 36.”
Wing, Elizabeth and Antoinette BrcMn
Pedeonutr ition : Methods and Theory in Prdiistoric Poodway s.
Academic Press.
Winterhalder, Bruce and E.A. Sknith
Hunter-gatherer foraqinq stoategiest ethnographic and archaeological
analyses. Ikiiversity of Chicago Press.
Wobst, H. Martin
Stylistic bdiavior and information exchange. In For the Director:
Research essays in the honor of James B. Griffin, edited by C.E.
Cleland, pp. 317-342. Anthropological Papers, Ikiiversity of Michigan,
Museun of Anthropology, 61.
Woodrick, Anne
^ analysis of the faunal rema^ from the Gaine5nn.lle Lake area:
Part II. Biocultural studies in the Gwin«*BvIlle Lake "area. Volume
4. Report submitted to the U.S. Amy Corps of Engineers, Mobile
District by the Office of Archaeological Research, Dhiveirsity of
Alabama. Rqxnrt of Investigations 14.
wynn. Jack T. and James R. Atkinson
Archaeology of the Qkashua and Self sites, Mississippi. Report
submitted to the U.S. Amy Corps of Engineers, Mobile District by the
Department of Anthropology, Mississippi State IMiversity.
Yamell, Richard A.
^Doriqinad Rela^caiships between Culture and Plant Life in the Upper
Great Laloes Regi^ Iftiiversity o£ Michigan, Anthropologiral Papers
23.
Yellen, John
Airhaeoloqical egaproaches to the present; model for reoonstiuctinq
the past. Academic Press.
Yuan, T.L.
Determination of exchangeable hydrogen in soil by a titration method.
Soil Science 88:164-167.
Zary, Jose{^ V.
Woodland management productivity. In Soil Survey of Itawemba County,
Mississippi. United States Department of Agriculture.
If you are looking for more details, kindly visit RF970 precise positioning type fiber.
For more RF970 Fiber Optic Perimeter Intrusion Detection Systemsinformation, please contact us. We will provide professional answers.
Link to Raycom