Test Excavations at the
Crater Site, 31WK155
Wilkes County, North Carolina
by
J. Ned Woodall
Archeology Laboratories
Wake Forest University
Winston-Salem, North Carolina USA
March, 1995
Table of Contents
- Abstract
- Acknowledgements
- Background
- Site location
- Geology
- History
- Excavations
- Methods
- Stratigraphy
- Features
- Artifacts
- Pottery
- Lithics
- Tools
- Summary and Conclusion
- References Cited
Excavations at the Crater Site (31 WK 155) in the summer of
1994 were conducted by the Archeology Laboratories of Wake Forest
University. Sixteen 2m squares were excavated by hand, and several
shovel tests and auger borings also were done to test for buried
remnants of the site. The work indicated that the Crater Site
belongs to the late Woodland, and much of it has been disturbed by
cultivation. Other than a few postholes concentrated within a
relatively small area no subsurface features were found. The site
does not appear eligible for the National Register of Historic
Places.
The field work at the Crater Site was done by participants in
the 1994 field school in archeological methods. I would like to
thank the students involved for their infectious enthusiasm, good
humor, and especially their careful and commendable efforts to
retrieve as much information as the site would allow. Participants
were Eric Biederbeck, Bill Duncan, Elaine Griesheimer, Bruce Idol,
Karen Sell and Karen White.
Our work at the site was made possible by the cooperation of
the landowner, Jim Crater, and the helpful services of the Ronda
town manager, Judy Anthony. My thanks also go to the citizens of
Ronda who showed us their kind hospitality and willingness to help
in the inevitable minor problems that arise in a project of this
sort.
The Crater Site was discovered in 1991 in the course of an
archeological survey of the Yadkin River floodplain between Elkin
and Roaring River (Rogers and Woodall 1992: 18). In 1992 a survey
was commissioned by the town of Ronda as part of the planning
process of a wastewater treatment facility designed to serve the
community. Along with a number of newly discovered sites the
Crater Site was re-surveyed at that time (Woodall 1992). The 1992
work included the use of shovel tests and backhoe trenches and
those subsurface tests, along with the surface indications of a
late Woodland occupation, suggested that a substantial and possibly
well-preserved archeological site was present. As part of the
management recommendations for the town of Ronda it was recommended
that the Crater Site be more thoroughly examined in order to
determine its eligibility for the National Register of Historic
Places.
In the summer of 1994 the Archeology Laboratories continued
its research program on the Woodland cultures of the upper Yadkin
Valley. Because the Crater Site seemed to fit our research needs,
and because testing of the site would benefit the town of Ronda, an
agreement was reached between the site's landowner, the town of
Ronda and the Wake Forest field party. In return for convenient
housing for the field party and other amenities, Wake Forest agreed
to conduct the necessary testing of the site. That work was
completed in three weeks, during late June and early July of 1994.
Immediately southeast of the town of Ronda is a broad
elliptical floodplain created by the Yadkin River bending to the
south. The floodplain is bisected by Hughes Branch; the Crater
Site is located in the western half of the floodplain, between
Hughes Branch and state road 2303 which crosses the Yadkin River
here (Figure 1).
When originally surveyed in 1991 the site appeared to be
confined to the riverine portion of the floodplain, along the crest
and slopes of the low remnants of a natural levee. The 1992 survey
identified cultural materials (primarily fire-cracked rock) over a
much larger area, including portions of the backswamp currently
drained by a narrow ditch extending from state road 2303 to Hughes
Branch. In the summer of 1994 the original assessment was found
more accurate, with the bulk of surface artifacts present in an
oval scatter 170m by 50m, the long axis paralleling the river
course. In the autumn of 1994, after the corn crop had been
harvested and the entire floodplain was clear of crop cover, the
area was surveyed twice again, and once more the surface materials
were concentrated near the river. Thus, the site boundaries as
given in the original survey report for the town of Ronda (Woodall
1992:Fig.1) overestimate the horizontal surface distribution of the
site.
The geological setting of the Crater Site is discussed fully
in the survey report issued earlier (Woodall 1992). To summarize,
the floodplain containing the site is composed of Holocene alluvium
deposited by the Yadkin River, primarily Buncombe sandy loam. This
friable sandy soil yields to Congaree silt loam in the northern
portion of the floodplain, away from the river in the poorly
drained backswamp. The site itself occurs near the river on a low
rise of loose sandy loam.
It is important to note that the topography of the floodplain
is not similar to that observed downstream in Yadkin and Surry
counties. In those counties the floodplains, especially the larger
ones, exhibit the classic surface profile of a well-defined natural
levee paralleling the river course; the levee declines more or less
gradually as one moves away from the river, through the backswamp,
until the surrounding residual soils of the steep bordering uplands
are reached. In the Ronda floodplain, in contrast, the area
between the backswamp and the river shows an undulating surface
with a weakly defined natural levee, and the levee itself is
interrupted by declivities formed at oblique angles to the river.
As was suspected in our original survey (Woodall 1992:3,9) and
confirmed by the excavations reported here, the floodplain has been
subjected to recent periodic episodes of aggradation and
degradation created by overbank flooding. This has affected the
integrity of the site, and may have been influential in selection
of settlement locations in prehistory.
Above the town of Ronda the Yadkin River receives several
large tributaries that drain the southeast-facing flanks of the
Blue Ridge Mountains. These tributaries are quite short (the
mountains are easily visible from the Yadkin) and most are deeply
incised. In consequence they are prone to rapid rises when storm
systems from the Atlantic or Gulf of Mexico slide along the Blue
Ridge front and deposit heavy rains from orthographic lifting. The
runoff is channeled into the Yadkin, and the sudden infusion of
water causes the river to rise rapidly and overflow its banks,
often with disastrous results. (It was to control such flooding
that the Kerr Reservoir was constructed a few miles above Ronda in
the 1960's.) During the 1994 field work several residents of the
area recounted in vivid detail the ferocity of the 1916 and 1940
floods on the town and its surrounds, including the floodplain
containing the Crater Site.
The history of Ronda and its vicinity is reviewed in the
survey report (Woodall 1992). For the floodplain containing the
Crater Site, local informants indicate that it has been under
cultivation for at least 100 years. The only exception to that use
is the reported construction of a small-scale pottery industry
immediately north of the site in the backswamp of the floodplain.
According to local informants a building once was present here, and
a small pond that lies adjacent to the site's access road is a
result of quarrying for clay. During the field season a systematic
survey of this sector of the floodplain was conducted, and yielded
only a dozen small fragments of brick and a tiny glass sherd. It
was said that the building that once was present was swept away by
a flood in the first half of this century.
Prior to excavation a datum was set inside the tree line
bordering the river course. The datum is a metal rod with the
arbitrary horizontal coordinates N2E100. All excavation units are
keyed to this datum and bear a N and E coordinate, indicating the
southeast corner of the square. The excavations concentrated on
the low levee crest north of the datum, and then extended in a
northern line to almost bisect the Buncombe loamy sands of the
riverine portion of the floodplain (Fig. 1). Altogether 16 squares
were excavated, each 2m by 2m, although several were clustered to
ultimately form two four-meter squares separated by a 1m wide
wheelbarrow path (Fig. 1).
For each square the plow zone, usually about 30cm thick, was
removed and discarded without screening. Excavation below this
proceeded by 15cm levels, adjusted in certain squares to
accommodate cultural stratigraphy. All fill below the plow zone
was passed through quarter-inch mesh. Feature fill was troweled
and retained for flotation; this soil was processed at the field
camp, passed through one-eighth screen and then permeable cloth.
Photographs were taken of each level of each square, and profiles
were drawn to scale and photographed of at least one side of each
square. Written records were kept of each square's excavation,
along with standardized data forms for each level and/or feature
excavated. All artifacts were washed, sorted and cataloged in the
field camp, and they, along with attendant documentation, are
permanently curated at the Archeology Laboratories of Wake Forest
University.
The stratigraphy at the Crater Site consists of three,
sometimes four levels distinguished by texture, color and artifact
content. In the several units closest to the river, i.e.
excavation units (E.U.s) 1-8,and 16, these are a 25-35cm thick plow
zone of loose yellow-brown sandy loam with occasional artifacts; a
dark brown, almost black clayey sand with artifacts, usually 10-
15cm thick, and a sterile compact yellow-brown sandy clay below
extending to an undetermined depth. In excavating these squares
the plow zone was removed, the dark stratum (referred to as the
midden zone in the field notes) removed as a unit, and then a 15cm
level excavated below into the sterile soil. This removal of part
of the sterile soil allowed any features present to be easily
observed, mapped and excavated separately. These consisted of
several postholes, described below.
As our line of squares extended farther to the north the
stratigraphy changed, and the new sequence was particularly evident
in E.U.s 11-16. Here the modern plow zone is again 20-30cm thick,
consisting of a rather friable yellow-brown sandy loam. Beneath
this is a thin zone of coarse laminated sand, varying in thickness
from 2-10cm. Under this is a darker brown clayey sand 10-15cm
thick with occasional artifacts, and under this is a very dark
brown clayey sand containing artifacts and resting on a compact
yellow-brown sandy clay. This sequence can be seen in Fig. 2a.
The deepest zone, the sterile yellow-brown sandy clay, extends
to an unknown depth; in E.U. 11 a deep test was made, with
excavation in a 1m by 1m square extended to 106cm below surface,
then an auger boring to slightly over 2m below surface--no
artifacts, charcoal or even river pebbles were recovered from the
homogenous deposits, which steadily increased in moisture content.
The artifact-bearing stratum clearly represents an old soil
horizon which was the land surface at the time of the site
occupation. This was subsequently buried and then plowed, and the
darker brown clayey sand then represents an old plow zone, probably
created by animal traction as judged from its thickness. The
underlying darker soil has, at the interface of the two, scars left
by the plow and, in E.U. 15, a badly rusted piece of iron was
present at the interface. This old plow zone itself was buried
under at least 30-40cm of alluvium. When plowing began again after
that flooding episode, the lower portion of the new alluvium was
not affected, leaving the laminated coarse sands undisturbed. This
sequence was not seen in E.U.s 1-10 because here, on the crest of
the levee remnant near the river, soil deposition was less during
the later flooding episode and failed to bury the artifact-bearing
soil as deeply. This burial of the cultural stratum probably
explains the occurrence of surface materials mainly on the riverine
portion of the floodplain; away from the river the site lies
underneath the modern plow zone.
Because of its burial under recent alluvium the boundaries of
the Crater Site are difficult to estimate. Our excavation units
were augmented with shovel tests and auger borings, 26 in all, with
shovel tests to 50cm below surface and auger borings to 1m. These
were concentrated along the eastern and western periphery of the
floodplain between Hughes Branch and state road 2003, and none
produced artifacts, charcoal or even the markedly darker soils
which contained cultural materials in our controlled excavation
units. Particularly evident along the western portion of the
bottom are thick deposits of coarse laminated sands below the
modern plow zone, indicating high energy alluviation processes
there, corresponding to the stratigraphy seen in our more northerly
excavation units.
Features at the Crater Site were found only in those
excavation units near the river. These consisted of postholes and
a circular area of red clay lying just beneath the plow zone, 60cm
in diameter and containing tiny bits of burned bone. This feature,
found in the center of E.U.16, was only 3-5cm thick, and probably
represents a short-term hearth area--the soil was not hardened by
the fire, and unfortunately no charcoal was present.
Twenty-five postholes were found below the dark stratum in
E.U.s clustered near the river, on the levee crest. All but one
were in the two four-meter square blocks created by contiguous 2m
squares, and this is the only portion of the excavated site that
produced any sign of subsurface occupational features. Figure 2b
shows three of these postholes near the northeast corner of E.U. 4.
The postholes were identified as such based on size and shape--each
was either troweled of its darker fill or vertically sectioned.
All were vertical or nearly so, and terminated between 56 and 80cm
below surface. Artifacts and bits of charcoal were found in most
of the postholes, and several contained complete or broken river
cobbles either in the posthole stain or immediately adjacent to the
stain. These stones are interpreted as wedges to secure the post;
one such stone can be seen in Fig. 2b.
Despite our best efforts, no unambiguous pattern was created
by the exposed postholes. It seems unlikely that they represent a
house. Although as many as six postholes could be seen in a linear
array, covering a distance of 2.5m, the line of posts did not
continue beyond this. Most postholes seems to be rather randomly
placed, but interestingly they did form loose clusters which
indicate they likely were part of integrated structures. Those
structures may have been drying racks or small crude temporary
shelters (perhaps lean-tos). Although flotation of the fill of the
postholes yielded flecks of charcoal from each, only one produced
large pieces of charcoal suitable for conventional radiocarbon
dating (Table I).
Table I: Radiocarbon Date from Posthole 16-3, 31WK155
Lab Number Uncorrected Calibrated Comments
Radiocarbon Range
Age B.P (2 sigma)
Beta-78598 1050+/-90 BP AD 790-1195 Small sample (< 1 g C)
Note: Dates calibrated with the program CALIB 3.0.1 (Stuiver and
Reimer 1993). Intercept with calibration curve: AD 1000.
Artifacts from the Crater Site consist of potsherds, lithics
(stone tools and debitage, or wastage from their manufacture) and
fire-cracked rock. Non-artifactual materials normally recovered
from riverine late Woodland sites such as mussel shell and animal
bone are absent, apart from a few minuscule bits of burned bone.
Even charcoal, normally found frequently in most Woodland site
excavations, was unusually rare at the Crater Site.
A total of 820 potsherds was found on the surface and in the
excavations at the Crater Site. A portion of these were not useful
for examination of surface treatment due to their small size or
eroded surfaces, and are listed as Unclassified in Table II. As
seen in Table II, the assemblage is dominated by plain and net-
impressed surfaces. Almost without exception the ratios of the
various surface treatments identified were consistent in the
vertical units of the excavations, implying that the favored method
of ceramic production did not change, at least in regard to surface
treatment.The great majority of the Crater Site sherds are between
4mm-8mm in thickness. Testing a relationship between surface
treatment and thickness, those over 8mm thick were subject to
separate surface treatment analysis. For these thicker specimens,
the predominant surface treatment was net-impressing (44 vs. 11
plain--10 could not be classified). Despite the suggested
relationship between vessel thickness and surface treatment, the
thick sherds were evenly distributed between the midden and the
plow zone.
Using the ceramics from E.U.s 1-7 as a sample, 162 sherds were
analyzed in regard to temper and surface treatment. Tempering
agents were, in descending order of occurrence, crushed quartz
(50.5%), sand (34.5%) and soapstone (15%). Of the sample selected,
110 sherds were undamaged and large enough for observation of
surface treatment. Using the four more common surface treatments
(plain, net-impressed, fabric-impressed and brushed), a chi-square
analysis was done of the relationship between surface treatment and
tempering agent. The contingency table thus created, with 6
degrees of freedom, yielded X=10.198, significant at the .05 level.
Examination of the table shows a weak tendency for the soapstone
temper to be associated with plain surface treatment.
Lithic materials were not abundant at the Crater Site. A
total of 1174 pieces of debitage were found in the excavations,
consisting entirely of flakes. Table III lists the flakes from
each excavation unit and level, broken down into primary (more than
half of the dorsal surface covered with cortex), secondary (some
cortex but less than half the dorsal surface covered) and tertiary
flakes (no cortex evident). Tertiary flakes were dominant,
indicating that tool finishing/refurbishing rather than initial
manufacture from cobble cores was the primary producer of the
materials.
It is evident from Table III that the lithic materials are
contained in the midden and, where the midden has been plowed, in
the modern plow zone and surface. Of greater interest is the
unusual abundance of debitage in E.U. 3. This is not matched by an
abundance of potsherds in that unit, which suggests that stone tool
processing was especially emphasized here. More importantly it
shows that the site has not been severely disturbed and sorted by
erosion, in turn supporting the interpretation offered here that
the absence of features, charcoal concentrations, animal bone and
other common constituents of Woodland sites is not due to natural
transformations.
The debitage also was studied in regard to raw material, using
the classes felsite (various metamorphosed volcanics including
rhyolite, andesite, and argillite), chert, quartz and jasper. Totals
for the site are felsite=68% (by weight), chert=9%, quartz=21% and
jasper=2%. The debitage from E.U. 3, mentioned above, is 98% felsite
of similar composition, supporting the hypothesis that this lithic debris
is primary refuse. The chert found is the black glossy material commonly
known as Knox chert, probably originating in southwest Virginia or
eastern Tennessee. The felsite and chert (and probably the jasper)
are imported to the site from some distance. The white quartz
found may have come from the quartz quarry site discovered by the
original Ronda survey. The quarry yielded a Woodland projectile
point and thus was in use during that time period, and it is
situated less than a kilometer north of the Crater Site. Despite
the proximity of a quarry producing white quartz of unusually high
quality that material is represented by only 21% of the Crater Site
debitage.
A total of 29 stone tools, complete and fragmentary, were
found during the 1994 excavations. The majority (23) are
projectile points, and at least two others likely are the distal
tips of points. Only one Archaic point was present, with the
remainder consisting of Woodland triangular or stemmed (type
Randolph) arrow points. Table IV shows selected attributes and
provenience of these specimens. Type designations are after Coe
(1964). Interestingly, the Crater site yielded no retouched flakes
other than those shown below, not even flakes with use retouch
commonly found on Woodland sites of the region.
The Crater Site, 31WK155, is a Woodland riverine settlement on
the alluvial bottomlands of the Yadkin River. Dating ca. A.D.
1000, it is unusual in the restricted range of artifact classes,
the near-total absence of non-artifactual debris such as bone,
mussel shell and charcoal, and the scarcity of sub-surface
features. Lower portions of the site have not been affected by
post-depositional disturbances by flooding or plowing, and the
restricted range of materials likely reflects a site use pattern
distinct from other late Woodland sites previously reported. The
presence of multiple post holes and a single hearth on the riverine
edge of the site, coupled with an island in the Yadkin river at
that point suggest that fish were being obtained here, the island
offering an opportunity to easily obstruct one river channel by a
weir. In the absence of other evidence however this is
speculative.
In regard to the culture history of the region the site
indicates that the use of soapstone as a ceramic tempering agent
likely began about A.D. 1000 and increased in popularity after that
time. Also, the importation of lithic raw material from regions
across the Blue Ridge front, as well as from the Carolina Slate
Belt, was well established at that date. Both these observations
should of course be considered somewhat tentative because they are
based on a single radiocarbon date, but nothing in the artifact
assemblage serves to dispute that date.
The 1994 excavations, including the auger and shovel tests
described above, do not indicate additional significant information
could be obtained by further work at the site. It does not appear
to be eligible for the National Register of Historic Places, and
thus does not seem to warrant further consideration in the planning
process for the water treatment plant.
Coe, Joffre L.
1964 Formative Cultures of the Carolina Piedmont.
Transactions of the American Philosophical Society 54, part
5.
Rogers, Rhea J. and J. Ned Woodall
1992 An Archeological Survey in the Upper Yadkin River
Valley, North Carolina. Report on file, Office of State
Archaeology, Raleigh.
Woodall, J. Ned
1992 An Archeological Inventory of the Town of Ronda
Wastewater Treatment Facilities, Wilkes County, North
Carolina. Report on file, Office of State Archaeology,
Raleigh.
Library Index