As part of the overall excavation strategy at the Clampitt site, several units were placed over magnetic anomalies which had been revealed by the gradiometer survey carried out in March of l991 (see above). Excavations were conducted at these locations in order to reveal the subterranean sources of the recorded anomalies. This "ground-truthing" exercise proved extremely fruitful, and the results are summarized in Table A-1.
The efficacy of magnetic surveys in detecting buried archaeological features is primarily based on the increased magnetic susceptibility of objects deposited within the features. Magnetic susceptibility is the physical property by which a material, when placed in a magnetic field, responds by producing a magnetic field of its own. Such an object, when placed in the earth's magnetic field, will add to the strength of the earth's magnetic field within a localized area surrounding the object. This local increase in the earth's magnetic field can be detected by the gradiometer as an anomalous high reading (Figure A-1). Human activities, most commonly heating, can artificially increase a materials magnetic susceptibility. This is especially true of materials containing iron oxides such as hematite, maghaemitite and magnetite. These iron oxides occur naturally in most soils and some rocks. The heating of soil under a hearth, the baking of clay in pottery production, and the heating of iron-rich rocks will all produce materials with an augmented magnetic susceptibility.
Figure A-2 illustrates the spatial distribution and generalized forms of the magnetic anomalies (i.e., light gray and white "spots" in figure) recorded during the survey in Grid numbers l through 8. As shown in Figure 13 (this report), only the eastern four grid squares (Grids 5 through 8) intersected the village itself and, consequently, it was within these grids that most of the testing of anomalies took place (the one exception was a single anomaly in the southwest corner of Grid 3 that was tested by Unit A).
The results of the magnetic survey were mapped in two different ways. Simple dot density maps were initially used to locate areas of obviously higher magnetic susceptibility (Figure A-2). Contour maps (Figure A-3) were then used to gain a better appreciation of the shape and intensity of the magnetic anomalies. In order to accentuate the magnetic anomalies, the contour maps were plotted using only the values derived from the 95th and 5th percentiles. Essentially, only the significantly high and low values from the survey data were used, effectively isolating the detected archaeological features from background "noise". Theoretically, a buried magnetized sphere will produce a magnetic signal known as a high- low dipole (Figure A-1) when measured from the surface. The maximum positive value of the detected signal actually occurs south of the source. The displacement is approximately one-third of the distance between the gradiometer sensor and the source of the magnetic anomaly. The deeper the source of the magnetic anomaly, the more displaced are the high readings.
Table A-1: Results of Magnetic Survey
Excavated Pit Features Excavated Pit Features
Undetected by Magnetic Survey Detected by Magnetic Survey
----------------------------- ---------------------------
Features 6,11,13,26,33,34,35, Features 1,2,3,4,5,9,10,12,
43,56 14,18,19,27,30,31,
32,42,49
Seventeen of the twenty-six pit features (65.456) excavated within
the area of the magnetic survey were successfully detected. On average,
the high magnetic readings recorded for these pit features were only 4 to
6 nanoteslas (nT) above the background soil magnetism, a considerably smaller
range than reported in other archaeological magnetic surveys. The successful
detection of the pit features was the result of a high instrument sensitivity
level (0.1 nT) combined with a fine scale sampling methodology. The
gradiometer successfully detected subsurface pits rich in cultural materials
but was unsuccessful in the detection of both the inner stockade trench,
which surrounded the village, and pits which contained a low density of
cultural materials. Figure A-4 illustrates the weights of cultural materials
in the detected pits versus those that were not detected. While the detected
features had, on avenge, a higher density of cultural materials, there was
still a degree of overlap between the two groups. A SAS regression analysis
was performed to determine what material, or combination of materials, were
primarily responsible for the strength of the magnetic signals emanating
from the detected features. Cultural materials recovered from these features
and included in the analysis were pottery, fire-cracked rock, chert (much of
which was heattreated), and limestone. The results of the analysis are
summarized in Table A-2.
The regression analysis indicated that pottery was the only significant source of a feature's magnetic field strength. This agrees with a direct comparison of pottery weights between those features detected by the magnetic survey and those which were overlooked (Figure A-5). There remained several problematic features in this analysis. The most glaring omission was the failure to detect Feature 6 in Grid 8. This feature contained 85 grams of pottery, far more than Features 3, 10, 14 and 42 which were detected by the survey (see Figure A-5). In the case of Feature 3, this discrepancy can be explained by the recovery of a unique artifact: a complete geode which was recovered from just below the plow zone in the northern half of Feature 3. Due to its unusual nature (intact geodes were rarely recovered during the excavation), its position was point-plotted prior to removal. The maximum positive value of the magnetic anomaly measured over Feature 3 occurred over the center of the pit rather than south of it as did the anomalies associated with the other features in Grid 8 (Figure A-6). As was stated above, the closer the gradiometer sensor is to the source of the magnetic anomaly the less is the displacement of the maximum positive magnetic value. The geode's position (near surface and in the northern half of the feature) suggested that it might have been the source of the magnetic anomaly. The geode was tested and found to have a high magnetic susceptibility which was caused by the deposition of iron oxides in its crust as a byproduct of weathering. The geode appeared to have been heated at sometime in the past causing an enhanced magnetic susceptibility.
Table A-2: Analysis of Variance
Sum of Mean
Source DF Suares Square F Value Prob.>F
-------------------------------------------------------
Model 4 8704.93 2176.23 7.81 0.0614
Error 3 835.20 278.40
C Total 7 9540.13
Root MSE 16.69 R-Square 0.9125
Dep. Mean 42.91 Adj. R-sq 0.7957
C.V. 38.88
Parameter Estimates
Parameter Standard T for H0:
Variable DF Estimate Error Parameter=0 Prob.>T
-------------------------------------------------------------
Intercept 1 14.74 11.90 1.24 0.3035
Pottery 1 0.03 0.007 3.73 0.0336
Chert 1 0.001 0.023 0.07 0.9499
FCR 1 -0.01 0.042 -0.30 0.7848
limestone 1 0.023 0.052 0.45 0.6863
The detection of the other three pottery-deficient features, in
contrast to Feature 6, can not be so easily explained. The degree of
magnetic susceptibility displayed by a pit feature is a complex
interrelationship between the pitas geometry, quantity of magnetic
materials and the relative location of those magnetic materials within
the pit. The detection of Features 10, 14, and 42 may have either been
fortuitous or the result of unrecovered cultural materials in the plow zone
above the pits. On the other hand, the magnetic signal produced by Feature 6
may have been concealed by the much stronger signals emanating from Features 1,
2, and 3 located nearby. Despite these reservations it does appear that all
features containing appreciable amounts of pottery (i.e. more than 100 grams)
were detected by the magnetic survey.
Overall, the marked success of the magnetic survey at 12 Lr 329 bodes well for the future application of this geophysical technique to other North American sites. Previously, magnetic surveys had been primarily applied to more substantial archaeological sites, mostly in the Old World, containing stone architecture and urban planning. Recent technological innovations in magnetometer design has produced a generation of machines capable of rapid survey at fine sensitivity levels. This technological improvement has lead to new surveying methodologies which are appropriate for use on the more "ephemeral" sites of aboriginal North America.
A small scale resistivity survey was initiated in May of 1992 in an attempt to trace the line of the inner stockade wall. Readings were taken at 0.25 meter intervals along eight north-south and eleven east-west transects. The north-south transects were spaced about three meters apart and the southwestern corner of this survey area was located at N15,E120. Nine of the east-west transects were each ten meters long and spaced at intervals varying from one meter to five meters apart; the southwest corner of this group was located at S3,EIS5. The remaining two transects were each fifteen meter long and three meters apart with the southwest corner at N17,E145.
The magnetic survey was largely unsuccessful in detecting the stockade trench due to the relatively shallow nature of the trench and the low quantities of magnetically susceptible materials in the trench. Resistivity surveys measure the ability of a volume of soil to conduct an electrical current. The primary factor in a soil's ability to conduct electricity is its ability to retain water, a property which is heavily influenced by the amount of organic matter it incorporates. Both the pit features and the trench were composed of a more highly organic soil than the surrounding soil matrix. This is the result of the decay of organic materials incorporated into the features prehistorically (i.e., food waste) and the deposition of organic topsoil into the less organic subsoil through human agency. Due to a miscalculation as to the probable path of the stockade walls many of the resistivity transects completely missed the stockade trench. The trench was only weakly indicated by the resistivity survey but a deep pit feature, Feature 71 (Figure 16), was detected as an extremely low resistivity reading.
Weymouth, John W.1986 Geophysical Methods of Archaeological site Surveying. In Advances in Archaeological Method and Theory 9. Edited by M. Schiffer, pp. 311-395. Academic Press, New York.
