Content uploaded by Ted Nelson
Author content
All content in this area was uploaded by Ted Nelson on May 05, 2020
Content may be subject to copyright.
The Hightower Village Site (1Ta150):
The Kymulga Phase (A.D. 1500-1650) and Its Place in the Coosa River Valley
By
Ted Clay Nelson
1
Presented at the 74th annual Southeastern Archaeological Conference in Tulsa Oklahoma,
November 9, 2017
1
Department of Anthropology, University of Alabama, tcnelson@crimson.ua.edu.
Swanton once wrote that the Creek town of Coosa was “one of the best established points
along De Soto’s route” (Swanton 1939:206). In fact, it seems that much of Swanton’s and the De
Soto Expedition Commission’s proposed route was established with the 16th century chiefdom
and town of Coosa located near Childersburg, Alabama (Error! Reference source not found.).
This would be because that is the location of the 18th century town of Coosa (DeJarnette and
Hanson 1960). Assuming the 16th century location of Coosa corresponds to the 18th century
location grossly underestimates
the complex history of the
American Indians now known as
the Creeks and historically as the
Creek Confederacy or Creek
Nation.
In fact, Swanton’s assumption
was proven inaccurate in 1960
with the excavation of the
Childersburg site (1Ta1) by
David Dejarnette. The works of
Marvin Smith, Jim Knight, and
many others have further
highlighted the complex social
history of Muscogee Creek
groups after European contact,
especially along the major Alabama Rivers. We now understand the Indians known as Creek are
Figure 1. Swanton's DeSoto Route (after Swanton 1939).
an amalgamation of multiple ethnic groups with diverse group identities. The Kymulga phase in
the 16th and 17th century Middle Coosa River Valley adds to this complexity. While it is
mentioned in various works (Hally, et al. 1990; Little 2008; Smith 2015; Walling 1993), the
Kymulga phase represents an infrequently examined Lamar people that can help to further
understand the early coalescent processes that occurred after 1540. I examine the Hightower
Village site, a Kymulga phase site in present day Talladega County, Alabama as an introduction
to this Coosa Valley occupation. I then use preliminary examinations of a Kymulga phase
ceramic assemblage to fit the Hightower Village site within the larger culture history of the Early
Historic period in the Coosa River Valley.
In general, the story of the Creek Indians along the Coosa River is a story of survival
through migration and coalescence. Throughout the study of past societies, much time and
energy have been spent in trying to understand how new societies develop. Such ventures have
led to theories in coalescence, polity formation, and theories of ethnicity. In Mississippian
culture, in order to emphasize new group cohesion and to legitimize the authority of the ruling
elites, certain parts of the past would be selectively forgotten while other aspects may be further
embraced (Blitz 2016:58). Such acts of selective social memory correspond strongly to changes
in material culture seen in many mound sites. At Moundville, many houses surrounding the
mounds were leveled in order to expand the plaza, older mounds that did not fit in the planned
structure were leveled, and ceramic styles were changed in order to emphasize polity cohesion
and legitimize a specific social order (Blitz 2016; G. D. Wilson 2007).
Mississippian polity formation demonstrates that processes we often describe for historic
Native American groups were not completely novel experiences that just occurred after
European contact. However, for mid-16th century Indian groups, we do not speak of polity
formation but about coalescence, a process where multiple groups come together in a loosely
unified political body for
mutual protection and
economic benefit (Kowalewski
2006). This social process, seen
in multiple places around the
world, has been shown to have
some social and material
signatures. For instance, in the
interest of unity, emphasis may
be placed on kinship or clan
structures that cross cut
communities, on yearly
renewal processes that
emphasize group integration,
and on mechanisms that
incorporate multiple social identity groups into a loosely unified socio-political structure (Beck
2013; Kowalewski 2006).
The 16th and 17th century Middle Coosa River Valley is an important time period and
region for understanding the early formation of the group we know today as the Creek Indians.
Before European contact, the areas in and around the Coosa River consisted of multiple
Mississippian Chiefdoms that belonged to general archaeological culture groups recognized
through ceramic similarities. The main archaeological culture relevant to my research is the
Figure 2. Lamar and Dallas Culture Areas
Lamar culture which spreads over all of Georgia and into parts of South Carolina, North
Carolina, Tennessee, Alabama, and Florida and dates from about AD 1350 to about the time of
European Contact (Williams and Shapiro 1990:4) (Figure 2). The large geographic area that the
Lamar culture encompasses has also been broken down into smaller phases based on variations
within the standard Lamar pottery tradition based primarily on grit tempered ceramics with
complicated stamped and incised designs (Halley 2008; Little 2008; Smith 1987). While other
Lamar culture phases will be discussed later in reference to the coalescence of Muscogee groups
in the Coosa Valley, I now turn to a description of the Kymulga phase and the Hightower Village
site in particular.
The Kymulga phase is found specifically in Talladega County, Alabama (Figure 3). It
dates roughly from AD 1550 to 1650 and possibly later. The Kymulga phase is also a Lamar
culture group, based on the
similarities in ceramic decoration
to other Lamar cultural phases;
however, tempers are more
diverse in Kymulga phase sites
than in other 16th and 17th century
Coosa River occupations. Many of
the ceramics are shell tempered,
grit tempered, or grog tempered
with different combinations of
these three ingredients (Williams
and Shapiro 1990). Most of the
sites attributed to this
archaeological phase are along
tributaries of the Coosa River and
were identified during Knight's
Coosa River Survey from 1983.
From this survey and previous research by Roger Nance and Lewis Larson, approximately
thirteen Kymulga phase sites had
been recognized within about a 30km radius of each other in the Coosa Valley (Knight, et al.
1984:13). Among these, only three have undergone any form of excavation, the Ogletree Island
Figure 3. The orange circle indicates the general location of the Kymulga Phase.
site (1Ta107), the Rogers-CETA site, and the Hightower Village Site (1Ta150).
The Kymulga phase site with the most extensive excavations is the Hightower Village
site (1Ta150), located north of Sylacauga on Emauhee Creek, a branch of Tallaseehatchee Creek
that runs into the Coosa River (Figure 4). It was occupied for most of the phase, from about AD
1500 to at least 1650 and probably later. The site was first identified as protohistoric by Knight’s
1983 Coosa survey due to the presence of iron artifacts identified as Spanish. At that time, the
village site was located in pastureland and was a popular location for looters to visit for day trips
from Birmingham. It was excavated in 1984, ‘85, and ’87 as part of the University of Alabama
Museums Expeditions under Richard Walling, Bob Wilson, and Harry Holstein. During these
excavations, 97 two meter by two meter units were completed and the artifacts are now housed,
mostly unprocessed, at the University Of Alabama Office Of Archaeological Research in
Moundville, Alabama. Prior to excavation a soil resistivity survey was executed on a portion of
the site (Appendix A). They used a Williams Soil Resistivity Device with four probes inserted
Figure 4. Map of excavations at Hightower Village site (1Ta150).
10cmbs and 1 meter apart. The survey was performed in ten 20 square-meters blocks and one
partial block. Appendix A is a reinterpretation of their data using Archaeofusion, a free remote
sensing data interpretation software. In 1984, decisions of where to place excavation units were
made using the resistivity data in order to improve chances of finding large features. In so doing,
at least 4 burials and three structures were located (R. C. Wilson 1987). The burials contained
glass beads, copper armbands, shell, ceramic, and stone artifacts. The 1985 excavations were
used in order to explore artifact rich areas found the previous year, expanding around certain
burials and structure
floors.
Before I started
examining the
Hightower Village site
ceramic assemblages, the
two most detailed
analyses came from
Knight, Cole, and
Walling’s (1984) report
on their Coosa Survey,
and an unpublished report on a single excavation unit by
Andrew Veech (1994). The survey sample Knight et al. (1984:98) procured from the site was
fairly small, containing only 89 sherds. Of these 79, or 88.7% are plain. The most common
decoration is incising with small amounts of brushed and complicated stamped decorations
following. Many of these designs crosscut multiple temper types. Knight et al. identified eight
Figure 5. Hightower site map with Veech's chosen unit identified.
different temper categories at the site: coarse and fine shells, Coarse Grit, Grit, Grog, and then
mixtures of these like Grit/Grog, shell/sand, and shell/grog. The most common temper type is
coarse shell, followed by grit, and then grog. The combination tempers never break 10 percent of
the relative frequencies. The next sample analyzed came from Andrew Veech’s analysis of the
1987 Unit N986 E1056 and was for some time the only detailed ceramic analysis from the site
that came from excavated material (Figure 5). The unit was chosen because the notes and field
drawings indicated that the unit may have been located on the eastern slope of a platform mound.
The notes mention that the archaeologists on site interpreted part of the south profile wall as
showing basket loads of dirt fill. Unfortunately, the unit did not focus on discerning any mound
layers in excavation. It was instead excavated in arbitrary 10cm levels. However, this method of
excavation may still provide some information about the ceramics from the site and any possible
time depth indicated by the different levels. In general, a quarter of the ceramics were just shell
tempered, another quarter were grit and grog, with small amounts of grit, grog, and shell
mixtures. If they are broken down into three tempers, they are 52% shell, and then 34% Grit, if
Grit supersedes Grog. In terms of decoration, the distribution is very similar to that of Knight’s
surface collection. Most decorations crosscut temper groups with no discernable pattern.
Unfortunately, most of the decorated sherds came from level 2, so no real pattern of change
through time could be determined.
My own preliminary analysis has found similar results. This past year, two undergraduate
students used ceramic samples from the Hightower Village site for class projects, focusing on
measuring attributes that I set up for them. The main sample examined the ceramic sherds that
could be found for levels 2 through 6 of Unit N950 E1070 that was excavated in 1984 (Figure 6).
Overall, about 80% of the sample was some combination of shell with some grit and grog,
almost 20% was Grit, Grit and grog, and less than 1 percent was solely grog tempered pottery.
There does seem to be a pattern of shell tempering becoming more common through time over
grit tempering but I only feel comfortable arguing that point after more comparisons of ceramic
assemblages across the site. The decorations are very similar to ones found by Veech, but does
include some punctated decorations, along with simple
stamping, complicated stamping, and
slipping as minority decorative techniques. The majority of decorations were curvilinear incised
and brushed designs. Thus far, the relative differences in temper type across the site are not that
different and correspond fairly well to the general surface survey conducted in 1983. Moreover,
the relative frequencies of designs and tempers are not altogether different to other sites of the
same phase in Talladega County. I now turn to a comparison of the Hightower Village site
ceramics to two other Kymulga phase sites.
Figure 6. Hightower site map with 1984 unit sample identified.
First, I compare my results to the Ogletree Island site, the only Kymulga phase site
located on the Coosa River itself, next to the mouth of Choccolocco Creek. It was excavated
prior to inundation in 1961 and 1962 by Ross Morrell. A single Nueva Cadiz bead was found in
association with a house floor which dates to about A.D. 1540. In total, three structures were
found with associated features. From these associated features, Richard Walling provides an apt
description of Kymulga phase ceramics from a site tentatively dated to 1540 to about 1600. The
ceramic analysis found 12 different temper types made up of combinations of fine and coarse
shell, sand or grit, and grog. Over half of the plain sherds analyzed contain shell, about 33
percent are grit or grog/grit tempers. Surface treatments include incising, complicated stamping,
brushing, and red slipping. Few incised sherds had identifiable motifs, but ran the gamut from
curvilinear designs such as guilloches and spirals, to rectilinear such as line-filled triangles.
Incising occurs mostly on shell tempered pastes while complicated stamping occurred on
primarily grit tempered pastes. I mention so much about the Ogletree Island site because it
provides a good comparison to the Hightower Village site, which has had even less analysis
performed on its ceramic assemblage than any of the other Kymulga phase sites. According to
Smith (2015) the Ogletree Island temper proportions is actually less grog, at only 5.5% of the
total when compared to shell and grit tempers, making its assignment to the Kymulga phase a
little dubious. However, I think a lot of discrepancy exists when discussing Kymulga phase
ceramics due to the inordinate amount of temper combinations. Attempting to boil the ceramic
assemblage of any Kymulga phase site down to just shell, grit, and grog, means arguing for some
tempers superseding others. It assumes some tempers would be introduced later and be the
deciding factor for strict temper categories. If we examine the grog only category, from the
Veech (13%), Knight (13%), and my samples (.71%), we see vastly different amounts within the
site. And yes, more grog tempering
seems to be present than at Ogletree
Island, but the difference is not as stark
as Smith reports. The second Kymulga
phase site of interest here is the
Rogers-CETA site, excavated in the
1980s and located on Talladega Creek.
Nance (1988) categorizes sherd
tempers as clay, shell, and sand with
incised and complicated stamped
decorations. Also, small mounts of
brushed, and cord marked designs are
present. Four blue glass beads were found during general
excavation and surface survey. If we were to examine just the differences in grog temper, we see
that about 6% of the Rogers-CETA assemblage is grog, 80% grit, and 12% shell. Nance proposes
the Rogers-CETA site was a late Lamar variant that would have been occupied during the early
part of the Kymulga archaeology phase, primarily the early to mid-1500s. If he is correct, then
the progressive changes in ceramic manufacture may indicate that grit was a preferred tempering
agent early in the phase, and then was progressively replaced by shell tempering and a minority
of grog tempering. However, this simplified explanation does not really explain the many multi-
temper categories found at all Kymulga phase sites. Still, the progressive switch to shell
tempered ceramics mimics the trend that has been reported for other parts of the Coosa River.
Figure 7. Location of Barnett Phase
Thus far, I have discussed the Hightower Village site ceramic assemblage and its relation to
other Kymulga phase sites surrounding it. Now I will turn to a larger comparison to sites and
archaeological phases found upriver. This comparison will help place the Kymulga phase into
the culture-history of the region and add context to the 16th and 17th centuries, when the
Hightower Village site was occupied.
The earliest relevant
Lamar Culture phase for
understanding Creek coalescence
is the Barnett phase, AD 1475 to
1550, located along the upper
parts of the Coosawattee River in
Northeast Georgia (Figure 7).
Marvin Smith argues that this
phase represents the primary
center of the Coosa Chiefdom that
was visited by De Soto and Juan
Pardo. Its ceramic assemblage
contains a mixture of Lamar grit
tempered ceramics and Dallas shell tempered ceramics,
creating some cultural tie to the Mississippian peoples of the Tennessee River Valley (Halley
2008; Smith 1989b). After European Contact, Smith describes these people as progressively
moving down the Coosa River into Alabama in the subsequent Weiss phase, Whorton’s Bend
phase, and Woods Island phase (Figure 8). His book published in 2000 convincingly argues that
Figure 8. Barnette phase (A.D. 1475-1550) location
Figure 8. Relevant phases on the Coosa River.
these subsequent phases represent the migration of the Coosa chiefdom out of Georgia and into
Alabama where they became the Creek town of Coosa and its sister towns. By the end of the 17th
century, the people of the Coosa chiefdom occupied an island in the middle of the river at the
Woods Island site (1Sc40) (Smith 2000). The ceramic assemblage at this site, which is also the
type site for the Woods Island phase, consists of mainly shell tempered wares with Lamar incised
motifs, rim applique strips, and some cord marked and brushed designs (Knight 1998; Smith
1989a). By the time the Weiss basin was occupied in the mid-16th century, Lamar bold incised
designs with minority complicated stamped decorations were being applied almost exclusively to
shell tempered pastes. In fact, this trend can be illustrated using Correspondence Analysis. Smith
(2015) provides relative frequencies of temper groupings on sites ranging from Northwest
Georgia to West-Central Alabama, and covering the protohistoric and early historic time periods.
The correspondence analysis of relative temper frequencies is based on ratios of grit, grog, and
shell tempers without regard to mixtures of paste ingredients (Appendix B). Most of the sites
distribute predictably, with most of the Mid-Sixteenth Century sites to the right, and the
Seventeenth Century sites to the left. Smith’s Hightower sample is an outlier due to the high
percentage of grog temper reported. While Rogers-CETA, Ogletree Island, and Hightower are all
part of the same archaeological phase, they are fairly spread apart due to the differences in grog
percentages among the samples. However, they still show some possible time lapses, due to
Rogers-CETA being further to the right, with Ogletree in the middle, and Hightower to the left.
In fact, the CA demonstrates that the Hightower Village site is an early seventeenth century site
with Ogletree Island most likely being a mid sixteenth century site. The Rogers-CETA site may
indicate that early Kymulga contexts are very contemporaneous with other late Mississippian
sites further up river. Moreover, it may be slightly contemporaneous with the Davis Farm site, a
16th century Barnett Phase context (Little 2008:75).
So what is the Kymulga Phase? An archaeological phase representing a long-term
occupation of the Middle Coosa River Valley in present-day Talladega. It is very unique in that it
lasts a very long time, possibly for over 150 years from before 1550 to after 1650. Ceramic
assemblages found in the phase are difficult to type because over eight different temper types are
always found, consisting of combinations of shell, grit, and grog. While it can be useful to
combine these types into three distinct categories, the jury is still out as to whether that overlooks
important information for understanding the ceramic manufacturing practices of the people
represented by the phase. Thus far, of the Kymulga sites found, the Hightower Village site was
most likely one of the last sites occupied in the region during the late 16th century. Most likely,
this site represents part of a historically known chiefdom that has yet to be fully identified. For
instance, Knight et al. (1984) argue that it may be part of the Talisi chiefdom visited by de Luna.
The similarities in ceramic manufactures may indicate that once the site was abandoned, people
moved up river to join their fellow Muscogee speakers from upriver. This phenomena is hinted at
in the increased use of shell seen at the Hightower site over the other Kymulga phase sites. The
use of shell tempering would indicate a slow change in ceramic making practices, which shows
increased efforts at uniting disparate peoples throughout the valley. The ultimate increase, too, in
brushed pottery throughout the valley may further indicate attempts at using material culture to
illustrate, either consciously or unconsciously, the progressing coalescences of groups in the
Coosa River Valley. I argue that Mississippian groups didn’t disappear, but simply changed with
the times and adapted to growing socio-economic pressures seen throughout the Early Historic
Period Southeast. They ultimately persisted and became what we know of today as Southeastern
Native American tribes.
References Cited
Beck, Robin A.
2013 Chiefdoms, Collapse, and Coalescence in the Early American South. Cambridge
University Press, Combridge.
Blitz, John H.
2016 Mound X and Selective Forgetting at Early Moundville. In Rethinking Moundville
and its Hinterland, edited by V. P. Steponaitis and C. M. Scarry, pp. 54-73. University
Press of Florida, Gainesville.
DeJarnette, David L. and Asual T. Hanson
1960 The Archaeology of the Childersburg Site, Alabama. Notes in Anthropology 4.
Halley, David
2008 King: The Social Archaeology of a Late Mississippian Town in Northwestern
Georgia. University of Alabama Press, Tuscaloosa.
Hally, David, Marvin T. Smith and James Langford
1990 The Archaeological Reality of De Soto's Coosa. In Columbian Consequences:
Archaeological and Historical Perspectives on the Spanish Borderlands East, edited by
D. H. Thomas, pp. 121-138. vol. 2. Smithsonian Institution, Washington, D.C.
Knight, Vernon James
1998 Aboriginal Pottery of the Coosa and Tallapoosa River Valleys, Alabama. Journal
of Alabama Archaeology 44:188-207.
Knight, Vernon James, Gloria G. Cole and Richard Walling
1984 An Archaeological Reconnaissance of the Coosa and Tallapoosa River Valleys,
East Alabama: 1983. Univeristy of Alabama Museums.
Kowalewski, Stephen A.
2006 Coalescent Societies. In Light on the Path: The Anthropology and History of the
Southeast Indians, edited by T. J. Pluckahn and R. Ethridge, pp. 94-122. University of
Alabama Press, Tuscaloosa.
Little, Keith J.
2008 European Artifact Chronology and Impacts of Spanish Contact in the Sixteenth-
Century Coosa Valley. Dissertation, Department of Anthropology, The University of
Alabama, Tuscaloosa.
Nance, C. Roger
1988 Archaeology of the Rodgers-CETA Site: A Lamar Village on Talladega Creek,
Central Alabama. Journal of Alabama Archaeology (34):1-240.
Smith, Marvin T.
1987 Archaeology of Aboriginal Culture Change in the Interior Southeast. Ripley P.
Bullen Monographs in Archaeology and History. University of Florida Press, Gainesville.
1989a In the Wake of De Soto: Alabama's 17th Century Indians on the Coosa River.
Alabama De Soto Commission, Tuscaloosa.
1989b Indian Responses to European Contact: The Coosa Example. In First Encounters:
Spanish Exploration in the Carribbean and United States, 1492-1570, edited by J. T.
Milanich and S. Milbreath, pp. 135-149. Ripley P. Bullen Monographs in Archaeology
and Histor. University of Florida Press, Gainesville.
2000 Coosa: The Rise and Fall of a Southeastern Mississippian Chiefdom. University
Press of Florida, Gainesville.
2015 Protohistoric Ceramics of the Upper Coosa River Drainage. In Archaeological
Perspectives on the Southern Appalachians, edited by R. A. Gougean and M. S. Meyers,
pp. 59-84. University of Tennessee Press, Knoxville.
Swanton, John R.
1939 Final Report of the United States De Soto Expedition Commission. vol. 71. 76th
Congress, 1st Session, Washington.
Veech, Andrew
1994 Title., in press.
Walling, Richard
1993 Lamar in the Middle Coosa River Drainage: The Ogletree Island Site (1Ta238), A
Kymulga Phase Farmstead. Bulletin Museum of Natural History 15:33-48.
Williams, Mark and Gary Shapiro (editors)
1990 Lamar Archaeology: Mississippian Chiefdoms in the Deep South. The University
of Alabama Press, Tuscaloosa.
Wilson, Gregory D.
2007 The Archaeology of Everyday Life at Early Moundville. University of Alabama
Press, Tuscaloosa.
Wilson, Robert C.
1987 A Preliminary Report on the Archaeological Investigations at the Hightower
Village Site (1Ta150), Talladega County, Alabama: 1984 and 1985 Field Seasons, edited
by U. o. A. Museums. University of Alabama Office of Archaeological Research.
Appendix A
Map of Soil Resistivity Survey
Appendix B
Correspondence Analysis Distribution based on Temper