Late Prehistoric Subsistence Shifts in
Eastern Arkansas:
Evidence From Landscape/GIS Analysis
Jamie C. Brandon and R. Joe
Brandon
Paper
Presented at the 55th Annual Meeting of the Southeastern Archaeology
Conference Greenville, SC. Nov 11-14, 1998.
Draft
Copy: Do Not Cite Without Permission from jcb
©2003
PRELIMINARY
RESULTS: DO NOT CITE WITHOUT AUTHORS' PERMISSION
[TITLE SLIDE] The development of agriculture has long
been looked at by archaeologists as a bellwether trait of complex
societies (e.g., Childe 1936; Wittfogel 1957), and the Central
Mississippi Valley is certainly no exception (Smith 1989,1992a).
[MAIZE SLIDE] Here, maize-reliant agriculture was looked
at as just that--a trait. And this trait, whether present or absent-whether
turned on or off-diagnosed the presence of whatever cultural florescence
it was being used explain. In the Mississippi Valley the transition
to agriculture and the adoption of a maize-reliant subsistence
economy has been placed in variety of temporal contexts over the
years including: Early Mississippian (Morse and Morse 1983:203;
Smith 1989), Baytown (Brain 1976, 1971:60), Hopewell (Griffin
1960), and even Poverty Point (Willey and Phillips 1958). [MISSISSIPPIAN
CORN SLIDE] Of late, researchers have settled on the correlation
of maize as a staple with the Emergent Mississippian culture material
shift. Whether the researcher viewed agriculture as a result or
cause, it seemed inextricably tied to the then monolithic concept
of the Mississippian.
The last ten to twenty years, fortunately, has seen researchers
dismantle many of the great monolithic concepts of southeastern
archaeology-the monolithic adoption of maize included. It now
appears that the use of maize as a staple occurs as early as the
Late Woodland-as in the American Bottom (Hastorf and Johannessen
1994)-or never at all as research indicates in Northeast Louisiana
(Fritz and Kidder 1993). Additionally, corn reliance can occur
rapidly-as it does in the Middle Tennessee populations (Buikstra
1992)-or slowly over a long period of time-as the evidence from
the FIA 270 project seems to indicate (Milner 1996).
In the Central Mississippi Valley, and eastern Arkansas in particular,
we are beginning to assess the shift in subsistence economy and
its relationship with the greater Mississippian cultural phenomena.
CORN IN THE CMV
[CMV_SLIDE] In the Central Mississippi Valley, like in
the rest of the eastern United States, corn has come to be associated
with the striking material culture shift that occurs between the
terminal Woodland and the Emergent Mississippian. In our study
area this is not without reason: [ZEBREE SLIDE] Dan Morse's
extensive excavations at the Zebree site yielded fairy large quantities
of maize and no non-disturbed Woodland features had any of it
(Morse and Morse 1980). Over the Missouri boarder as well, the
Gypsy Joint site has yielded over 40 grams of corn cupules and
cobs-extremely large quantities of maize. . . more than any other
single site in the region (Smith 1978).
Recently an alternative view of the adoption of a maize-reliant
subsistence regimen has been offered from the quarter of bioarchaeology.
[ROSE SLIDE] Rose and co-workers (Rose et al. 1991) as
well as Lynott and co-workers (Lynott et al. 1986) have utilized
stable carbon isotopes to trace the adoption of maize as a staple
through differential absorption of carbon through C4 pathways.
Rose approached his data set-a massive sample of skeletal material
from many sites in eastern Arkansas-by mustering evidence from
dental carries, dental wear, and pathologies as well as stable
carbon isotopes. The result: Rose postulates a slow shift to maize-based
subsistence not reaching staple proportions until sometime in
the Middle Mississippian in Northeast Arkansas while the Southeastern
portion of the state may never have been fully maize-reliant.
The latter concept seems to fit well with very recent and ongoing
work of Gayle Fritz and T.R. Kidder (Fritz and Kidder 1993) just
over our southern boarder in northeast Louisiana. Through an intensive
floatation regime and key excavations they have [ACORN SLIDE]
effectively questioned the presence of corn in the Coles Creek
heartland, stating that corn may not become accepted until as
late as 1200 AD, and even then not becoming a staple. Instead
they propose a acorn, hickory nut "pull" on settlement:
placing sites on levee deposits of major rivers.
By comparison the northeast Arkansas evidence is more ambiguous
and when we examine the testing and mitigation projects in eastern
Arkansas we find a more mixed message. . . [ZEBREE SLIDE] Aside
from the early component at Zebree (Morse and Morse 1980), maize
has been recovered in several early Mississippian contexts such
as [LITTLE CYPRESS SLIDE] Little Cypress Bayou (3CT50:
Dicks and Weed 1982), and [DEROSSITT SLIDE] DeRossitt (3SF49:
Spears 1978), but only in very small amounts (i.e., 3-5 cupules)
while many other early Mississippian sites, even with floatation,
do not yield corn: such as [CORN_SITES SLIDE] Mangrum (3CG36:
Klinger ) and Steele (3MS351: Brockington and Dicks 1982). In
the middle Mississippian Period Cherry Valley (3CS40: Soday 1966)
provides the only negative evidence while MacDuffe Place (3CG21)
Gypsy Joint and Zebree all yield Middle Mississippian context
maize.
Brogham Lake (3CT98: Kilnger et al. 1983), with its Tchula, Baytown,
Middle and Late Mississippian components, offers a radically different
picture. At Brogham Lake, 77 corn kernels and 44 cupules were
recovered from Baytown features comprising 28% of the edible floral
remains recovered. The Mississippian components combined yielded
over 400 specimens of corn. But again comprising about 21% of
the recovered edible flora. Although other Woodland period sites
have yielded corn specimens, Brogham Lake is the only one to yield
them in these high amounts. Of course, I'm obligated to point
out that none of this corn has been directly dated via AMS. .
. and we are relying on cultural material as our guide to age.
[CORN SLIDE] Given preservation problems in the region and
recovery inconsistencies this hardly compares to the Louisiana
data, and does not confirm nor refute Rose's hypothesis-especially
as most of the corn has not been directly dated. We are left with
three large Maize yield sites: the multi-component Zebree and
Brogham Lake, and Gypsy Joint. Only one of which-Zebree-has an
Early Mississippi manifestation.
In order to get at this problem from yet another line of evidence,
we have begun to look at settlement pattern changes from the Late
Woodland to Middle Mississippian using GIS technology and a coarse,
state-level analysis. As maize, [KNOTWEED/SUNFLOWER SLIDE]
unlike its agricultural predecessors of the floodplain weed
complex, requires high-maintenance and certain soil conditions:
a shift to a maize-reliant subsistence economy may lead to a shift
to farming areas advantageous for maize growth.
MATERIALS: SITE, SOILS AND GEOMORPHOLOGIC
DATA
[MATERIALS _SLIDE] Site data for the Central Mississippi
Valley in the state of Arkansas was provided by the Arkansas Archeological
Survey, via their AMASDA data base (Donat 1995; Hilliard and Riggs
1986). [SITES SLIDE] Sites listed under the myriad of phase,
culture, and complex names were conglomerated under the temporal
designations: "Late Woodland," Early Mississippian"
and "Middle Mississippian." As shown here, we also divided
the project area along this line-the 35 parallel- which closely
approximates the division between the Central Mississippi Valley
and the Upper portion of the Lower Mississippi Valley in Arkansas.
This also coincides with some cultural manifestations with few
Coles Creek sites occurring north of this parallel.
These data-now grouped by cultural affiliation-were then brought
into the GIS packages: the UNIX-based GRASS 4.1, developed by
the COE and the Windows-based Grassland1.1®.
[BUFFER SLIDE] As farming areas are not necessarily coincidence
with living areas, a 2 mile buffer zone was placed around each
site (1 mile shown here in black, the second mile in white). Thus
soil conditions within reasonable farming distance was incorporated
into our analysis.
Then coincidence tables were generated for each set of sites (Late
Woodland, Early Mississippian and Middle Mississippian) versus
the soils and geomorphology layer.
Soil association and geomorphology data was supplied by the Center
for Advanced Spatial Technologies (CAST), at the University of
Arkansas. [SOILS.250K SLIDE] The relatively "coarse"
state-wide soil association map (1:250k) was used, as finer scale
county soils data are still being completed.
[GEOMORPH SLIDE] The geomorphology layer, also provided
by CAST, is a digitized version of the maps prepared by Roger
Saucier (1994) of the USCOE Waterways Experiment Station in Vicksburg,
Mississippi. It allowed us to assess the age of landforms in addition
to providing data regarding another "pull" on settlement
patterns: the ecotone.
ANALYSIS
Analysis of these co-occurrence tables show a Late Woodland preference
for the Kobel association ( 2=92.31) in both Northeast and Southeastern
Arkansas. Kobel soils are commonly associated with depressional
to nearly level-floodplains near channels and oxbows of the White,
Black, Cache and St. Francis Rivers in Eastern Arkansas: prime
ectonal environments.
In the Early Mississippian sample, the distribution looks quite
different. High Chi Square values for both the Sharkey-Alligator-Tunica
association and the Dundee-Bosket-Dubbs associated were returned
for Northeast Arkansas. 36 of the 118 Early Mississippian sites
were located on the Sharkey-Alligator-Tunica complex, a soil infamous
for fine-grain texture (It'll dull up your shovel in a hurry)
. . .It's soil description reads aptly: "clay: structureless,
massive, firm, very sticky" and anyone who has worked in
the Mississippi Valley can attest to the truth of that statement.
Additionally, 51 of the 117 (43%) Early Mississippian sites in
Northeast Arkansas occur on the Wisconsin age braided stream surfaces:
also an interesting correlation.
The Middle Mississippian, however, returns high Chi Square values
for Dundee-Bosket-Dubbs association ( 2=41.32: 37 out of the 114
sites: 32%), the prime loams of the project area. Other soil association
of preference is The Amagon-Dundee (n=20) association (another
loam dominated association). Together these two associations account
for 50% of all Middle Mississippian sites. Additionally, an interesting
negative correlation is observed in the Middle Mississippian period
with the Calloway-Henry-Grenada association. Although this soil
type occurs as frequently in the project area as does the Dundee-Bosket-Dubbs
and Amagon-Dundee soils combined, only a single site assigned
to the Middle Mississippian occurs on this association. The Calloway-Henry-Grenada
association is dominated by silts with a hard frigpan situated
on the smoother portions of the loess belt of the Mississippi
Valley, including areas on stream terraces. . . We don't know
what this means yet.
In the southeastern portion of the study area, the Late Woodland
period sites return high Chi Squares for both Sharkey-Alligator-Tunica
and Kobel association soils. And Skarkey-Alligtor-Tunica soils
also remain the preference in both Early and Middle Mississippian:
making up 40% of the distributions in both cases. Interestingly,
the Coles Creek phenomena in particular shows a drastically different
distribution pattern with the overwhelming number of sites (n=111,
60%) occurring on the Rilla-Herbert association soils: the natural
levee and point bar deposits along present and past channels of
the Red and Arkansas Rivers. The geomorphic layer confirms this
diagnosis: a preponderance of sites situated pointbar deposits
and floodplain backswamp.
In summary, in the northeast Late Woodland sites appear to prefer
the oxbow related Kobel associations. There is an apparent shift
with the early Mississippian, but not to the prime Dundee soils,
rather the preponderance of sites end up on the Skarkey clays
of the floodplain-tough if fertile farming. By the Middle Mississippian
sites are situated on the loamy Dundee soils which incidentally
have high soil capabilities ratings. Our take on all this is that
is supports Rose's hypothesis more than it detracts from it. Clearly
Late Woodland sites are not situated on prime agricultural soils,
while the Middle Mississippian sites are. What happened in between
is likely a product of the multiple direction pull of other environmental
and social factors on settlement patterns. . . but I'll address
that in a minute.
Interestingly in southeastern Arkansas, chi squares show an importance
of the finer textured floodplain soils (Sharkey-Alligator-Tunica)
and the Rilla-Hebert association throughout the cultural sequence.
These soils represent geomophically the levee and pointbar deposits
that would yield acorns, hickory nuts and the ectonal environments
suggested by Kidder and Fritz (Fritz and Kidder 1993).
GIS AND THE MULTIPLE VARIABLE "PULL"
ON SETTLEMENT
[ANALYSIS_SLIDE] Using settlement patterns to infer subsistence
can be a dangerous thing. Although we are all aware of the extent
to which corn figured into native cultural life as sign and symbol.
. . intricately involved in cosmology and self-identity in native
society. We would be foolish to think it-as seen by proxy through
environmental factors such as the agricultural productivity of
soils-was the only pull on native settlement systems. This would
reduce us to a type of environmental determinism too easily acquired
when using spatial tools such as GIS. What should be stressed
here is that we are not trying to find a single "prime-mover"
cause for settlement shifts. . . We expect that corn was but one
of many variables pulling on settlement location in the Mississippian
period. [GATHERING SLIDE] The long-holding pulls of ectonal
environments (such as oxbow lake), riverine transportation routes,
and perhaps social hierarchy and group membership are still there.
Interestingly we can see differing patterns regarding settlement
and major stream networks as seen here between a portion of southeastern
Arkansas (with sites situated within these 2 mile stream buffers)
and northeast Arkansas ( we see situations more on lands between
the north-south running rivers). Also note the clustering of sites
against the upland escarpments-a major ectone in eastern Arkansas.
[CORN1_SLIDE] What should also be addressed here is the
symbolic aspects of corn as linked with Mississippian society.
. . If we are to break away from our trait-oriented way of looking
at the past we must recognize the dynamics of cultural change.
. . The non-use of corn as a staple in the Early Mississippian
of Eastern Arkansas would not mean these people were any less
Mississippian. Only that corn was used as a Mississippian symbol,
not as a functional crop necessary to support a swollen population.
We cannot deny that sites such as Zebree offer us maize in Early
Mississippian contexts, but as Jerry Rose says "that doesn't
mean they are putting it in their mouths." Rose and co-workers
also point out a maize-fed sub-adult buried in an elite contexts.
. corn could have been used to help create a dominate ideology
of Mississippian elite only to become a staple for the masses
through time. [Corn_Maya Slide]
As Hastorf and Johannessen (1994) point out in the recent Corn
and Culture in the New World volume: "Shift in the food systems
served to carry messages about new social identities and new solidarities,
and thus had an active role in the process of cultural definition"
(p. 441).
Interestingly, corn can also be seen to inform the referent about
his (or her) relationship to the landscape itself: corn is often
seen in native mythologies as being neither cultural nor natural,
but more like a mediator between the two. Corn is basically a
natural substance (plant-life) that needs cultural intervention
to survive.
CONCLUSIONS
[Corn_field Slide] This is simply the first step in an
ongoing research vein that will continue with my co-author's thesis
work. Not only will other environmental and social variables be
examined as a part of this inquiry into prehistoric environmental/cultural
dynamics in Eastern Arkansas, but methodological issues regarding
GIS analysis will also be addressed.
1) Asses the Reliability of the Data
Source: [Sites_Slide]
Although Arkansas is one of the acknowledged leaders in database
management of site file. . but we cannot kid ourselves as to it
absolute accuracy. [SBARNES SLIDE] as attested to by this
lonely site coded for cultural affiliation as "Barnes"
but hundreds of miles outside the distribution of the Barnes phase.
As GIS analysis in archaeology commonly relies on this source
of data- state site files- we should begin to assess its relative
accuracy. We have chosen a stratified random sample from the Arkansas
state site files. . stratified as to include certain percentages
from each cultural affiliation designation. We are then pulling
site files and inventories to look for the presence of actual
diagnostics. This procedure is a tedious one, but we in the end
will be able to assess the impact of this data's accuracy on our
analysis.
2) Survey Area Bias: [Proj_Area
Slide] As the Arkansas Archeological Survey also keeps a current
GIS database on project areas, we will be able to insure that
our results are not simply the result of soil types surveyed.
As most projects are not random in their distribution, we expect
some shifts in analysis to occur.
3) Issues in Resolution: [ Soils_250k Slide]
Coarse, state-level data such as used here are also common in
GIS analysis. . .but does this level of resolution stand up when
finer scale data is available? or does it serve as a perfectly
fine, sufficiently resolvable level of analysis. The data scale
used here was 1:250k, but currently the NRCS is completing large
sections of the digital county-level soil layers. . . .Some of
these are currently available, although in unapproved form, to
CAST. Although other researchers (i.e. Gillam 1995,1996) have
employed this level of data in archaeological site distribution
analysis, we feel that it is imperative for GIS users to evaluate
the usefulness of coarse-level analysis. My co-author, R. Joe
Brandon, is currently addressing resolution by running similar
analysis to this one on county-level data sets. . No matter if
our conclusions are similar or different, we will have made strides
toward understanding what is an "appropriate level of analysis"
for research questions such as these.
So there, I've tried to pack a bunch of things into one: our preliminary
analysis of settlement/soil shifts in eastern Arkansas, a tirade
about single variable approaches, a urging for social considerations
in analysis, and a critique of GIS problems along with future
direction of our research. I should thank you for your time.
I would also like to thank the Arkansas Archeological Survey for
support and site data, CAST for data and facilities, and James
Davidson, Kathy Cande, and Dr. Robert C. Mainfort for their thoughtful
comments. Thank you.
BIBLIOGRAPHY
Belmont, John
1967 The Development of Agriculture in the Lower Valley. Bulletin
No. 4 of the Southeastern Archaeology Conference p. 16-33.
Buikstra, Jane; Jill Bullington, Douglas Charles, Della C. Cook,
Susan R. Frankenberg, Lyle W. Konigsberg, Jospeh B. Lambert, and
Liang Xue.
1987 Diet, Demography, and the Development of Horticulture. In
Emergent Horticultural Economies of the Eastern Woodlands, William
F. Keegan, ed. Center for Archaeological Investigations, Occasional
Paper 7, Southern Illinois University.
Brain, Jeffery P.
1971 The Lower Mississippi Valley in Prehistory. Report submitted
to the National Park Service, Southeast Region and the Arkansas
Archaeological Survey.
1976 The Question of Corn Agriculture in the Lower Mississippi
Valley. Bulletin No. 19 of the Southeastern Archaeological Conference
p. 57-60.
Brockington, Paul E., A. Merrill Dicks, and Wayne P. Glander
1983 the Buffalo Creek Archaeological Project: Vol. II. Excavation
of the Steele Site (3MS351), Mississippi County, AR. Report submitted
to the USCOE, Memphis District.
Childe, V. Gordon
1936 Man Makes Himself. Watts, London.
Cochran, Roy J.
1982 Site Distribution at Mangrum. In The Mangrum Site: Mitigation
through Excavation and Preservation. Arkansas Archaeological Survey
Research Series No. 20.
Dicks, A. Merril, and C. S. Weed
1985 Archaeological Investigation of the Little Cypress Bayou
Site (3CT50), Crittenden County, Arkansas. New World Research
Report of Investigations No. 82-21. Contract No, DACW66-82-0064.1.
Submitted to the USCOE, Memphis District.
Donat, Lela
1995 Computerized Site Files in Arkansas: Utilizing New Technologies.
In Archaeological Site File Management: a Southeastern Perspective.
David Anderson and Virginia Horak, eds. Readings in Archaeological
Resource Protection Series - No. 3 Interagency Archaeological
Service Division. Atlanta.
Fritz, Gayle J.
1994 Coles Creek and Plaquemine Landscapes. Paper presented to
the 1994 Southeastern Archaeological Conference, Lexington, KY.
1997 A Three-Thousand-Year-Old Cache of Crop Seeds from Marble
Bluff, Arkansas. In People, Plants, and Landscapes: Studies in
Paleoethnobotany. Kristen J. Gremillion, ed. University of Alabama
Press, Tuscaloosa.
Fritz, Gayle J., and Tristram R. Kidder
1993 Recent Investigations into Prehistoric Agriculture in the
Lower Mississippi Valley. Southeastern Archaeology 12(1) p. 1-14.
Griffin, James B.
1960 Climatic Change: A Contributory Cause of the Growth and Decline
of Northern Hopewell Culture. Wisconsin Archaeologist 41:2.
Harris, Suzanne E.
1982 Preliminary Ethnobotantical Analysis. In The Mangrum Site:
Mitigation through Excavation and Preservation. Arkansas Archaeological
Survey Research Series No. 20.
Hilliard, Jerry E. and John Riggs
1986 AMASDA Site Encoding Manual: Version 2.0. Arkansas Archeological
Survey Technical Paper No. 1.
Jeter, Marvin D., Jerome C. Rose, G. I. Williams, and A. M. Harmon
1989 Archaeology and Bioarchaeology of the Lower Mississippi Valley
and Trans-Mississippi South in Arkansas and Louisiana. Research
Series No. 37, Arkansas Archeological Survey, Fayetteville.
Kelly, John E.
1990 The Emergence of Mississippi Culture in the American Bottom
Region. In. The Mississippian Emergence, Bruce D. Smith, ed. Smithsonian
Institution Press, Washington DC
Kidder, T. R.
1992a Timing and Consequences of the Introduction of Maize Agriculture
in the Lower Mississippi Valley. North American Archaeologist
13(1).
1992b Coles Creek Period Social Organization and Evolution in
Northeast Louisiana. In Lords of the Southeast: Social Inequality
and the Native Elites of Southeastern North America Alex Barker
and Timothy Pauketat, eds. Archeological Papers of the American
Anthropological Association Number 3.
Kidder, T. R. and Gayle Fritz
1993 Investigating Subsistence and Social Change in the Lower
Mississippi Valley: the 1989 and 1990 Excavations at the Reno
Brake and Osceola Sites. Journal of Field Archaeology 20(3).
Klinger, Timothy C.
1977 Subsistence-Settlement Variability in the Mississippi Valley:
An Example from Northeast Arkansas. Unpublished M.A. Thesis, University
of Arkansas, Fayetteville.
1982 The Mangrum Site: Mitigation through Excavation and Preservation.
Arkansas Archaeological Survey Research Series No. 20.
Klinger, Timothy C., S.M. Imhoff, and R. J. Cochran, Jr.
1983 Brougham Lake: Archaeological Mitigation of 3CT98 Along the
Big Creek Enlargement and Diversion, Item 1, Crittenden County,
Arkansas. HPA, Fayetteville, AR. Contact No. DACW66-80-C0082.
Submitted to USCOE, Memphis District.
Lynott, Mark J., Thomas W. Boutton, James E. Price, and Dwight
E. Nelson
1986 Stable Carbon Isotopic Evidence for Maize Agriculture in
Southeast Missouri and Northeast Arkansas. American Antiquity
51(1) p. 51-65.
Martin, J. H. ; W.H. Lenoard , and D.L. Stamp
1976 Principles of Field Crop Production 3rd ed. Macmillan Press,
New York.
Morse, Dan F.
1969 Preliminary report on the archaeological investigation at
the Zebree site (3MS20), summer 1969. Ms. on file, Arkansas Archeological
Survey.
Morse, Dan F., and Phyllis A. Morse
1976 A Preliminary Report of the Zebree Project, New Approaches
in Contract Archeology in Arkansas. Arkansas Archeological Survey
Research Report No. 8
1980 Zebree Archeological Project: Excavation, Data Interpretation,
and Report on the Zebree Homestead Site, Mississippi County, Arkansas.
Contract No. DACW 66-76-C-0006. Report Submitted to the USCOE,
Memphis District.
1983 Archaeology of the Central Mississippi Valley. Academic Press,
New York.
1996a Northeast Arkansas. In Prehistory of the Central Mississippi
Valley, Charles H. McNutt, ed. University of Alabama Press, Tuscaloosa.
1996b Changes in Interpretation in the Archaeology of the Central
Mississippi Valley Since 1983. North American Archaeologist 17(1).
Perino, Gregory
1967 The Cherry Valley Mounds and Banks Mound 3. Central States
Archaeological Societies, Inc. Memoir No. 1.
Rose, Jerome C., Murray K. Marks and Larry L. Tieszen
1991 Bioarchaeology and Subsistence in the Central and Lower Portions
of the Mississippi Valley. In What Mean These Bones?: Studies
in Southeastern Bioarchaeology. Powell, Mary Lucas, Patricia S.
Bridges, and Ann Marie Wagner Mires, eds.. University of Alabama
Press.
Scholtz, James
1965 Summary Report on Archaeological Phase II Project at the
DeRossitt Site (3SF49). University of Arkansas Museum, Fayetteville.
Submitted to the Arkansas Department of Transportation and Highways,
Little Rock.
Shea, Andrea B.
1985 Ethnobotany of the Little Cypress Site. In Archaeological
Investigation of the Little Cypress Bayou Site (3CT50), Crittenden
County, Arkansas. New World Research Report of Investigations
No. 82-21. Contract No, DACW66-82-0064.1. Submitted to the USCOE,
Memphis District.
Smith, Bruce
1978 Prehistoric Patterns of Human Behavior: A case Study in the
Mississippi Valley. Academic Press, New York.
1989 Origins of Agriculture in Eastern North America. Science
Vol. 246, December.
1992a Rivers of Change: Essays on Early Agriculture in Eastern
North America. Smithsonian Institution Press, Washington.
1992b Hopewellian Farmers of Eastern North America. In Rivers
of Change: Essays on Early Agriculture in Eastern North America.
Smithsonian Institution Press, Washington.
1992c Origins of Agriculture in Eastern North America. In Rivers
of Change: Essays on Early Agriculture in Eastern North America.
Smithsonian Institution Press, Washington.
Soday, Frank
1966 Excavation of the MacDuffie Site, Craighead County, Arkansas.
Field Notes No. 15:4-5.
Spears, Carol
1978 The DeRossitt Site (3SF49): Applications of Behavioral Archaeology
to a Museum Collection. Unpublished M.A. Thesis, University of
Arkansas, Fayetteville.
Wittfogel, Karl
1957 Oriental Despotism: A comparative Study of Total Power. Yale
University Press, New Haven.
