Dr. Michael L. Hargrave (ERDC CERL) conducted geophysical surveys at the New Philadelphia site, Pike County, Illinois, during three brief visits in April, May, and June 2004. Alleen Betzenhauser (University of Illinois) assisted with the data collection in May. The objectives of this work were to identify subsurface archaeological features associated with 19th century occupation of the site, and to provide field school students with an introduction to the use of geophysical survey in archaeological field research. The geophysical work was conducted in support of the ongoing, NSF-funded field school and related investigations at the New Philadelphia site conducted by the University of Maryland, Illinois State Museum, University of Illinois, and the New Philadelphia Association.
Archaeological features such as pits, privies, cisterns, domestic architecture, etc., represent localized disturbances to soils that would otherwise comprise relatively homogeneous deposits (at the spatial scale relevant to archaeological sites). Features frequently contain organically enriched fill that is darker in color or different in texture than the surrounding soils. It is often this visual (and, to some extent, textural) contrast that permits archaeologists to detect features during excavation. Similarly, geophysical techniques can detect subsurface archaeological features that contrast with the surrounding soils in terms of electrical resistance, magnetic, or other properties. Factors that can create a geophysical contrast include soil compaction, particle size, organic content, artifact content, burning, and moisture retention. Remnant magnetism and magnetic susceptibility are particularly relevant for magnetic feature detection. Heating iron oxides (present in many soils) above ca. 400 degrees Centigrade results in a permanent change (remnant magnetism) in the objectís magnetic field. Human occupation often introduces burned and organic materials to the local soils and increases magnetic susceptibility. In general, any human action that involves the localized disturbance of the soil is potentially detectable by geophysical techniques. Localized disturbances associated with tree roots, rodents, and other natural phenomena, as well as recent cultural activities (vehicle ruts, plow furrows, etc.) are also often detectable.
In a geophysical map, cultural features (as well as other discrete disturbances) may appear as anomalies, i.e., spatially discrete areas characterized by geophysical values that differ from those of the surrounding area. Prehistoric features such as pits and hearths are typically characterized by a very low contrast with the surrounding soil matrix. Historic features frequently contain metal artifacts and architectural debris (brick, mortar, stone footings, etc.) and thus typically exhibit a stronger contrast with their surroundings.
Several other factors can make it difficult to identify anomalies associated with low contrast features. All geophysical surveys are to some extent affected by noise, a seemingly random component in the data attributable to the instrument itself, the operatorís field technique, or variability in the siteís soil, rocks, etc. Clutter refers to non-archaeological, non-random, discrete phenomena that complicate feature detection. Clutter can include plow furrows, rocks, tree roots, rodent burrows, and modern metallic debris. At some sites, anomalies associated with clutter can be stronger and more numerous than anomalies related to cultural features.Magnetic Field Gradient Surveys
Two geophysical techniques were used at the New Philadelphia Site: magnetic field gradiometry and electrical resistance. The magnetic survey (Bevan 1998; Heimmer and De Vore 1995; Scollar 1990) was conducted using a Geoscan FM36 gradiometer. This instrument is manufactured by Geoscan Research, a small firm in Great Britain that produces geophysical instruments and software optimized for archaeological applications. The gradiometer can measure exceedingly subtle disruptions in the earth's magnetic field that were caused by prehistoric and historic-era cultural activities, as well as recent cultural and natural phenomena.
The Geoscan FM36 records the difference or gradient between the values measured by two fluxgate sensors that are positioned at slightly different (.5 meter) distances from the ground surface. To collect data, the surveyor walks along a predefined transect, carrying the gradiometer in one hand. A sound emitted by the instrument's automatic trigger allows the surveyor to distribute the data collection points at regular intervals. The surveyor must take care to keep the tube containing the two magnetic sensors perpendicular to the ground surface. Deviations of the instrument from the perpendicular are manifested in the data as slightly anomalous readings. The overall effect of such anomalous values is to decrease the signal to noise ratio, making it less likely that very subtle features will be detected.
In preparation for the survey, a metric grid comprised of 20 by 20 meter blocks was established at the site. Blocks of this size represent a widely used data collection unit for many geophysical studies, particularly those conducted using Geoscan instruments. The blocks were oriented approximately 45 degrees east of magnetic north. This deviation from the New Philadelphia historic town plat (which is oriented to the cardinal directions) was necessary to prevent the obfuscation of linear features as a result of magnetic processing techniques (the zero mean traverse routine in Geoplot software can remove linear features that are parallel to the data collection traverses).
In each block, the survey began in the west corner and proceeded northeast and southwest along transects that were spaced at 1 meter intervals. Transects were marked using nonmagnetic tapes held in place by plastic tent pegs. The gradiometer was set for its maximum resolution (.1 nanoTesla). The survey area was in tall (ca. 20 cm) grass and field conditions were generally favorable.
In the gradiometer surveys, data values were collected at .125 m intervals as the surveyor moved along each transect. This strategy resulted in a medium-density survey (8 data values per square meter) and reasonably high-resolution maps.Electrical Resistance Surveys
Resistance surveys (Bevan 1998; Hargrave et al. 2002; Heimmer and De Vore 1995; Scollar 1990) introduce an electrical current into the ground and measure the ease or difficulty with which the current (measured in ohms) flows through the soil. Cultural features and other localized soil disturbances can be detected if they differ sufficiently from the surrounding soil in terms of their resistance to the passage of the current. The number and mobility of free charge carriers (principally soluble ions) are the primary determinants of electrical resistance. The simultaneous availability of soil moisture and soluble salts determines the free charge carrier concentration in the soil. The mobility of the soluble ions is governed by soil moisture content, soil grain size, temperature, soil compaction, and the surface chemistry of the soil grains (Somers and Hargrave 2001). In situations where the fill of cultural features hold moisture more readily than the surrounding soils, the pits may be manifested by low resistance anomalies. Alternatively, features characterized by relatively coarse or loosely compacted (well-drained) fill may be associated with high resistance anomalies. A pit that is manifested by a high resistance anomaly in one season can conceivably be associated with a low resistance anomaly in other seasons, when relative soil moisture is different. Concentrations of building debris (bricks, stone rubble, etc.) typically exhibit relatively high resistance.
The resistance surveys at New Philadelphia were conducted using a Geoscan RM15 resistance meter equipped with a PA5 probe array and MPX multiplexer. The instrument was configured with three probes spaced at .5 meter intervals, generally known as a parallel twin configuration. This probe spacing was selected in order to collect resistance data representative of the uppermost ca. .5 meter of deposits. This depth was selected under the assumption that features would be located immediately below the modern plow zone. In the resistance survey, data values were collected at .5 meter intervals north south along transects spaced at 1meter intervals east west. This strategy produced a relatively high-resolution survey (4 data values per square meter) that should be adequate to detect most features larger than .5 meter diameter.Data Processing
The magnetic and resistance data were processed using Geoplot 3.00, a software package developed by Geoscan Research (Walker and Somers 2000) for archaeological applications. Geoplot routines were used to identify and remove data defects, detect anomalies that could be associated with cultural features, and to cosmetically improve the appearance of the maps. None of these processing steps resulted in the creation of anomalies that were not present in the raw data.
The general processing sequence for the magnetic data was as follows. Data were first Clipped to remove extreme outlying values. The Despike routine was then used to further reduce the effects of isolated data spikes. The Zero Mean Traverse routine was used to set the background mean of each traverse to zero. This removed much of the striping that is often present in the raw data. The interpolation routine was used to achieve square pixels. A Low Pass Filter was then conducted to remove high frequency, small-scale spatial detail (i.e., to smooth the data). The Low Pass Filter is often very effective in improving the visibility of the larger, weaker cultural features. As a final step, the processed data were imported into Surfer 8.0 to produce the image maps presented here.
The resistance data were processed somewhat differently, although the processing objectives were similar to those of the magnetic surveys. Despike was used to remove localized extreme values that can occur when a probe contacts a rock or other hard object. A High Pass filter was then used to remove the effects of the geological background, thereby increasing the visibility of relatively small anomalies that could be associated with cultural features. The data were interpolated to provide a finer-grained appearance. Surfer 8.0 was used to produce the maps included here.
Results of the geophysical surveys are presented in this report as gray-scale image maps. In general, data quality is very good. Note that the processed resistance and magnetic data are bipolar, with a mean of approximately zero (mapped as 50% gray). Positive values range from 50% gray to black, and negative values range from 50% gray to white. Maps viewed on the computer screen are, of course, much higher resolution and more readily interpretable than are the maps provided here. Anomalies thought likely to be associated with cultural deposits were highlighted in color and labeled A, B, C, etc. Note that only the most obvious anomalies were singled-out in this manner. It is highly likely that many other cultural features are manifested by subtle or otherwise ambiguous anomalies. As excavation proceeds, it is likely that the investigators will be able to make increasingly reliable interpretations of anomalies that, at present, appear to be ambiguous.
Boundary lines of the historic-period blocks, lots, streets and alleys of the town site have been overlain onto the data image maps by Dr. Christopher Fennell of the University of Illinois using graphics software. Fennell has also provided illustrations showing where, subsequent to the geophysical survey and analysis, associated excavation units were placed by the field school to further investigate particular areas of anomalies and a number of significant features were uncovered (see Figures 3, 4, 9-12, 16 and 17).
Geophysical survey at New Philadelphia was conducted during three visits: 27 April, 26-27 May, and 14 June 2004. Overall, 30 magnetic grids (12,000 square meters) and 18 resistance grids (7,200 square meters) were surveyed. Three areas were investigated: (i) portions of the area once covered by Block 9, Lots 5 and 6; (ii) portions of Blocks 3, 4, 7, and 8; and (iii) portions of Block 13, Lots 2-4.I. Block 9, Lots 5 and 6
Five magnetic and four resistance grids were surveyed here on 27 April (Figure 1, Figure 2). The objective of this initial survey was to assess the potential usefulness of electrical resistance and magnetic field gradiometry at the site. A second goal was to detect evidence for a structure that, based on archival and oral history sources, was believed to have been located there.
Several relatively large resistance anomalies were identified along the northern edge of the survey area, and several smaller anomalies located in the south-central area were viewed as possible building footings (Figure 2). These were not, however, singled out as high priority targets, given the absence of clear evidence for architectural remains. The presence of long, slightly curving linear anomalies in the resistance data was viewed as possible evidence for the subtle remains of early architectural terracing or a rather unusual result of historic plowing.
The magnetic survey area extended one grid further south than did the resistance survey area (Figure 1). The southern-most three grids included a number of relatively strong magnetic anomalies. Some of these were distributed in linear patterns, although these alignments did not seem to intersect at the right angles that might be expected for the in-situ remains of walls. The southern-most magnetic grid included an east-west oriented strong anomaly comprised of several dipoles (a dipolar anomaly is a paired positive and negative associated with a strong magnetic value and often indicative of metal). At the time of survey this was interpreted as a possible pipeline or other infrastructure feature.
On balance, results of the initial geophysical survey at New Philadelphia (Figure 1, Figure 2) indicated that electrical resistance and magnetic gradiometry surveys would be productive. It was recommended that larger contiguous areas be surveyed in order to achieve more interpretable results. The excavation teams later placed excavations units in the area of anomalies in the southwest corner of Block 9, Lot 5 (Figure 3) and uncovered the remains of a storage space that had later been as a refuse pit during the late-nineteenth century and early twentieth century (Figure 4).II. Blocks 3, 4, 7 and 8
On May 26 and 27, Hargrave returned to the New Philadelphia site to collect additional data. At this time each of the students had an opportunity to assist in data collection. The students played a major role in collecting the resistance data. Preliminary results and interpretations are described below.
Figure 5 shows the results of the resistance survey; Figure 6 shows the same data with selected anomalies highlighted. Thus far, the resistance data appear to offer the best evidence as to the possible location of architectural features. This is because strong magnetic anomalies often do not have dimensions that are coterminous with the actual feature or artifact. Obvious examples are the datum markers that are manifested in Figure 3 by anomalies that appear have diameters of several meters. On the other hand, the resistance anomalies tend to be subtle, and some of them overlap with the linear soil features (plow furrows).
Figure 7 shows the magnetic data. The strong, discrete black and white monopole anomalies as well as the dipole anomalies are likely to be associated with metal artifacts. The fainter gray discrete anomalies could also be associated with bricks or rocks that are somewhat more magnetic than the surrounding soil, or with relatively small pieces of metal buried at relatively greater depth. In general, concentrations of discrete magnetic anomalies are certain to be artifact concentrations, presumably associated with discard areas and/or habitation loci. These concentrations of anomalies correlate well with the distribution of ferrous metal from the Controlled Surface Collection conducted at the site in 2002 and 2003. The geophysical data appear to provide indications of more discrete concentrations and should thus be more useful than the surface collection data in locating architectural remains.
Figure 8 shows the magnetic data with an overlay of selected resistance anomalies. Resistance anomalies that occur in areas where there are few magnetic anomalies are problematic. If the high resistance anomalies were associated with construction materials (stone footings, etc.), one would expect metal artifacts to also be present. It is conceivable, however, that early structures could occur without abundant metal.
Resistance anomalies that occur in the presence of numerous magnetic anomalies are good candidates to be associated with architectural remains. Resistance anomaly A, for example, is highly likely to be architectural (Figures 6 and 8). Anomaly A may relate to the foundation of a structure that was much larger than a nearby cabin at the site (which is not an original structure of the site).
Resistance anomaly B is a fairly large (several meter) locus that also has a magnetic expression. It should be investigated as a possible architectural feature (footings, chimney, etc.) (Figures 6 and 8).
Anomaly C is visible in the resistance data as an apparent square or rectangular shape suggestive of a small structure (Figure 6). However, the absence of magnetic anomalies is troubling, so it is viewed here as problematic. Some low-effort investigation using soil cores and/or shovel tests is recommended.
Anomaly D is a fairly large (several meter) area of high resistance that also exhibits a few small magnetic anomalies. Anomaly D may simply be a component of the north-south oriented furrow or terrace complex, but it warrants investigation.
Anomaly E is an area of weak linear resistance anomalies suggestive of architecture. However, there are few magnetic anomalies at that locus, and E may simply be a component of the furrow/terrace complex. Anomaly E has a low probability of being architectural, but should nevertheless be investigated using soil cores and shovel tests.
Anomalies F and G are prominent positive resistance anomalies. They appear to be associated with a few, relatively weak magnetic anomalies. Investigating anomalies such as F and G will be productive in that it will help the excavators learn to interpret similar phenomena in the resistance data at this site.
Anomalies H and I are similar to F and G. The former appear to be aligned with the track of an old road (called King Street in historic maps) and thus may simply be components of that feature. They could be potholes filled with gravel or looser soil, etc. However, H and I are located in an area of abundant magnetic anomalies, and this may increase the likelihood that they are concentrations of architectural debris, etc.
Anomaly J is similar to H and I, but is associated with the edge of a track of an old alley (called High Alley in historic maps), rather than associated with the area of King Street.
On balance, the geophysical survey in Blocks 3, 4, 7, and 8 was highly productive. The historic streets and, to a lesser extent, the alleys are clearly discernable in the resistance data, and at least faintly visible in the magnetic data. Many of the highlighted resistance anomalies appear to be located along the streets and alleys, as would be expected.
Excavation teams later placed a unit in Block 7, Lot 1, in the area of a visible anomaly in the magnetic data map and an area that yeilded relatively high concentrations of surface artifacts in an earlier pedestrian survey (Figure 9). This excavation unit revealed a significant feature of stone foundation remains (Figure 10) that date from the late-nineteenth century to early twentieth century. Excavators also placed units in Block 8, Lot 4, over the area of anomalies B and C (Figure 11). They uncovered the remains of a house foundation that dates from the mid-nineteenth century (Figure 12).
Excavations were also conducted in Block 3, Lot 4, just north of the area covered in the geophysical survey of that lot area, based in part on densities of artifacts uncovered in an earlier pedestrian survey (Figure 16). The remains of a lime slaking pit for producing lime plaster were uncovered (Figure 17), which also dates to the late-nineteenth century.III. Block 13, Lots 2-4
Figure 13 shows the results of an electrical resistance survey conducted on 14 June 2004 in the area once covered by Lots 2-4 of Block 13. The objective was to map the remains of a ca. 1855 structure that may have been used as a small hotel or guest house. A number of resistance anomalies are identified and labeled in Figure 15. No excavations have yet been undertaken in this area.
Figure 13 and Figure 14 both show the data without and with (respectively) the use of a High Pass Filter. This filter calculates the mean for a "moving box" centered on each data point, and subtracts that mean from the value in question. This is done for each data value in the map. The High Pass Filter removes generalized variation (often related to the natural soil), allowing clearer identification of anomalies likely to be associated with archaeological deposits or other localized phenomena. (Note that the resistance survey results in Figure 2 and Figure 5 were also processed using a High Pass Filter).
In Figure 13 and Figure 14, the unfiltered map in Figure 6 shows a faint, slightly L-shaped anomaly that could be the footprint of the structure. This anomaly is identified as K in Figure 15. It would be useful to excavate several transects of soil probes or shovel tests across K, extending well beyond its limits, to determine if one can detect soil or fill characteristics that are associated with the anomaly. Anomaly K seems quite large, however, so it may also be productive to focus on the some of the higher resistance anomalies that occur within its limits. Note that these smaller anomalies appear more discrete in High Pass Filtered data, as shown in Figure 15.
L is a rather large, strong positive resistance anomaly. It could be a concentration of building debris or (less likely) a deposit of loose soil, possibly a pit. The High Pass Filtered data suggest the feature associated with anomaly L may have an irregular shape.
Anomalies M, N, and O (and several unlabeled anomalies near O) are all very discrete high resistance anomalies located in or near a ditch that appears in Figures 6 and 7 as a linear low resistance anomaly. It seems likely that anomalies M, N, and O may not be in-situ features (given their presence in the ditch), but this remains speculative. One could attempt to locate the objects associated with these anomalies using a probe or soil core.
Anomalies P, Q, R, S, T, V, and W are all discrete high resistance anomalies located with the limits of K (the possible structure). These could be either localized deposits of building material (footings, chimney fall, etc.) or (perhaps less likely) pits with looser, drier fill.
Anomaly U is a trench-like high resistance anomaly. It is not certain what type of feature this may represent, but its north-south orientation is consistent with that expected for the structure.
Many other anomalies are present in the magnetic and resistance maps. Those mentioned here and highlighted in the accompanying figures are viewed as the most likely to be associated with archaeological features. Systematic investigation of these anomalies, as well as of a sample of those not singled out here, will allow the New Philadelphia project investigators to maximize the interpretive value of the geophysical maps.
A current trend in the use of geophysics in U.S. archaeology is an emphasis on large-area surveys. Large area coverage increases the reliability of interpretations and enables the investigation of past community plans and activity patterning. Given that New Philadelphia, like many 19th century communities, was partitioned into standard sized lots arranged along a symmetrical grid of streets and alleys, a large area geophysical study offers an excellent opportunity for the investigation of these topics. A large area survey will also result in a visually compelling image that will help students and members of the general public visualize the archaeological remains of the New Philadelphia community.
It is recommended that additional resistance and magnetic field gradient surveys be conducted during the second and third years of the New Philadelphia project. Additional magnetic survey will be useful because it will identify concentrations of metal artifacts that likely correlate (at least in general terms) with structure locations. This should be more obvious if and when much larger areas have been surveyed.
The resistance data appear to be very useful for identifying the historic roads, alleys, and architectural remains. It would be useful to collect resistance data all around the known building locations (based on the early aerial photographs), as this should help identify a series of structures and major features at those loci.
Investigations at New Philadelphia site are funded by a grant from the National Science Foundation's Research Experiences for Undergraduates program. The author would like to thank Dr. Paul Shackel (University of Maryland), Dr. Christopher Fennell (University of Illinois), and Dr. Terry Martin (Illinois State Museum) for the invitation to participate in the project, assistance with the geophysical fieldwork and background information about the site. Dr. Fennell overlaid the New Philadelphia streets and blocks onto the geophysical maps and provided the maps showing the locations of excavation units. Ms. Alleen Betzenhauser (University of Illinois) assisted with the collection of magnetic data in May. Thanks also to all of the students of the 2004 New Philadelphia Site Field School, who played a major role in collecting the electrical resistance data. Finally, my thanks to the New Philadelphia Association, particularly Mr. and Mrs. Armistead and Mr. and Mrs. Likes for their hospitality and ongoing efforts to preserve this important site.
Bevan, Bruce W.
1998. Geophysical Exploration for Archaeology: An Introduction to Geophysical Exploration. Special Report No. 1. U. S. Department of the Interior, National Park Service, Midwest Archeological Center, Lincoln, Nebraska.
Hargrave, Michael L., Lewis Somers, Thomas Larson, Richard Shields, and John Dendy
2002. The Role of Resistance Survey in Historic Site Assessment and Management: An Example from Fort Riley, Kansas. Historical Archaeology, 2002, 36(4).
Heimmer, Don H., and Steven L. De Vore
1995. Near-Surface, High Resolution Geophysical Methods for Cultural Resource Management and Archeological Investigations. U. S. Department of the Interior, National Park Service, Rocky Mountain Regional Office, Division of Partnerships and Outreach, Interagency Archeological Services, Denver, Colorado.
Scollar, Irwin, A. Tabbagh, A. Hesse, and I. Herzog
1990. Archaeological Prospecting and Remote Sensing. Cambridge University Press, Cambridge.
Somers, Lewis, and Michael L. Hargrave
2001. Magnetic and Resistance Surveys of Four Sites. In Phase II Archaeological Evaluation of 25 Sites, Fort Bragg and Camp MacKall, Cumberland, Harnett, Hoke, and Moore Counties, North Carolina, by B. Idol and S. Pullins, pp. 342-355. Report submitted to CERL by TRC Garrow Associates, Inc., Durham, NC.
Walker, Roger, and Lewis Somers
2000. Geoplot Version 3.00 for Windows. Instruction Manual Version 1.0 March 2000. Geoscan Research, Heather Brae, Chrisharben Park, Clayton, Bradford, BD14 6AE, UK.