SEAC 2001: Magnetometer and Auger Testing at the Berry Site, Burke County, North Carolina

by Thomas Hargrove and Robin A. Beck, Jr.

During the 1980's, Slide Charles Hudson and his associates revived debate on the explorations of sixteenth century Spaniards Hernando de Soto and Juan Pardo in the interior Southeast. In 1994, new evidence emerged to provide strong support for what has come to be known as the Hudson Route. John Worth's discovery and translation of the 1584 Domingo de Leon account, together with the identification of sixteenth-century Spanish ceramics and hardware at the Berry site in Burke County, North Carolina Slide ,indicates that the course of the Hudson route through the Carolina Piedmont is accurate. Further, on the basis of documentary evidence and the utilitarian nature of the Berry assemblage, Worth, David Moore, and Robin Beck have identified this site as Joara (Soto's Xuala), an important polity visited by both the Soto and Pardo expeditions.

Juan Pardo's first expedition departed from Santa Elena, located near modern Charleston, South Carolina, on December 1, 1566, and likely arrived at Joara by early January, 1567. Pardo renamed this town after his native city in Spain, as both were "located at the foot of a range of mountains, surrounded by rivers." At Joara, Pardo states that he found a large number of Indians and caciques and made a fort where Boyano, his sergeant, and certain soldiers remained with their munitions of powder, matchcord, balls, and maize to eat. Fort San Juan de Joara became Pardo's primary outpost in the interior, and was occupied by a contigent of about 25 Spanish soldiers who lived at the fort for more than a year and a half before it was overrun, probably during the spring of 1568, by local Native Americans.

Slide Hudson et al. originally argued that Joara was located on the upper Catawba River, near the present town of Marion, North Carolina; Worth's translation of the Domingo de Leon account leaves little doubt that this interpretation is largely accurate. Archaeological discoveries made in the 1990's, however, strongly suggest that the Berry site, twenty-one miles east of Marion, is the probable location of Joara; investigations at the Berry site have revealed an impressive assemblage of sixteenth-century Spanish ceramics and hardware, including 13 sherds Slide from at least four different Olive Jars Slide, a sherd of Caparra Blue majolica Slide, which is diagnostic of the early contact period, from 1492 to 1600, and has also been recovered at the site of De Soto's first winter encampment in Florida, a sherd of gray earthenware Slide which, as yet, has only been recovered from the Berry site and Santa Elena, nails Slide, lead shot and sprue Slide, a buckle fragment, and a brass ball button.

In contrast to these utilitarian artifacts, only a few non-diagnostic glass Slide and brass Slide beads have been recovered from the site. While the beads and an iron knife Slide recovered from a burial can probably be classified as trade goods, Worth's examination of sixteenth- and seventeenth-century Spanish trade lists has yielded no evidence of items such as iron nails and Olive Jars ever being traded to Indians. Most of the sixteenth-century Spanish artifacts recovered from the Berry site are thus believed to be utilitarian debris associated with Fort San Juan, the earliest European fort constructed in the interior Southeast.

The Berry site is located on Upper Creek Slide, a tributary of the Catawba River, about eight miles north of Morganton in Burke County. The site is situated on the extreme northeast margin of a 200-acre alluvial bottomland formed by the confluence of Upper Creek and Irish Creek. The Berry site was first recorded in Cyrus Thomas's 1891 report of the BAE's Mound Explorations, where it is described as a "Mound on the west Bank of Upper Creek 8 miles north of Morganton (about 15 feet high and unexplored)." Both the mound and the surrounding site were regularly plowed, and in 1964 the mound itself was bulldozed to provide fill for a low-lying area west of the mound that was often subject to flooding. Today, the remaining mound is approximately 200 feet in diameter and is visible as a slight rise about two feet above the level of the field. The entire site covers at least five hectares (13 acres), as determined by Beck's systematic surface collection, and David Moore's excavations in 1986 revealed numerous intact features (roasting pits, burials, and postmolds) that had survived the plow and the bulldozer.

In June 1997 Slide , Tom Hargrove, Rob Beck, David Moore, Chris Rodning, and volunteers from the Burke County Historical Society conducted a magnetometer survey of a 0.9 acre section of the Berry site. A magnetometer designed for archaeological fieldwork detects certain kinds of cultural features as magnetic anomalies-patterned, localized distortions in the background magnetic field. Substances that produce distinctive magnetic anomalies include feromagnetic artifacts (iron, for example), or iron oxides such as hematite and magnetite. When soils, clays, or stones with some magnetic mineral content are subjected to intense heat (as in a prehistoric hearth, a burned wattle and daub structure, bricks, or a pottery kiln), they become strongly magnetic and stand out as anomalies. Since topsoils are usually more magnetic than subsoils, a pit dug into subsoil and later filled with topsoil may also appear as a positive magnetic anomaly. (NO SLIDE)

The magnetometer we used at the Berry site was a Geoscan FM 18 fluxate gradiometer. This configuration uses a pair of magnetometers in one instrument: a lower, measurement detector for sensing localized magnetic disturbances on a very small scale; and an upper, reference detector for keeping track of background magnetic changes. The magnetometer took readings every 25 cm along north-south transects spaced 50 cm apart within nine separate grid squares, each square measuring 20 X 20 meters. This sampling interval meant that each grid square was surveyed by 3,200 separate data points. The site had been planted in corn, but we were able to run the north-south transects between corn rows. The corn stalks sometimes obstructed smooth movement of the very finicky magnetometer, which is reflected in some roughness in the collected data. These data were processed with Geoplot 2.01 software.

The Berry site's natural setting Slide favors the use of a magnetometer. The far western Piedmont's soils fall into the Felsic Crystalline System, where soils are largely derived from igneous rock. The famous red color of North Carolina's Piedmont soils is due to the presence of iron, often in the form of hematite or iron oxides. The natural magnetic susceptibility of this soil permits the creation of artificially magnetized features, especially when the soil is exposed to intense heat. The soil of the Berry site is a well-drained alluvium derived from parent materials with some magnetic susceptibility. The relative lack of modern metal in the site (apart from shotgun shells and other minor items) also helped in the collection and processing of the data.

Focusing on the area of the site where most of the Spanish artifacts have been found Slide, the magnetometer recorded several strong, well-defined anomalies indicating house floors and other subsurface features, including large, burned structures that may represent buildings associated with Fort San Juan. If the Berry site includes the remains of Fort San Juan, then we might also expect the magnetometer to detect some of the metal tools with which the outpost was scantily supplied when it was overrun in the spring of 1568. The records of the Pardo expedition list several types of iron objects left at the fort that might appear as magnetic anomalies: mattocks, picks, chisels, axes, and wedges. Some late sixteenth century nails have already been recovered from the site, and an iron knife was recovered in a burial excavated by Moore in 1986. Most of the Spanish metal tools, however, would probably not have remained in the fort, but would have been salvaged by the people of Joara after the fort was overrun. Many of these items probably ended up scattered throughout Joara, and then throughout western North Carolina Slide, perhaps including some of the famous iron artifacts recovered from nearby sites in the Yadkin River valley. The magnetometer did pick up several isolated metal signals, but most of these have not been confirmed as sixteenth century artifacts. Volunteers from the Burke County Historical Society, using metal detectors, did recover a sixteenth century rose-headed nail and lead shot during the magnetometer survey.

This slide Slide shows an overall shade plot of the nine grid squares, covering about 0.9 acres. The magnetometer detected a large number of positive anomalies, which appear as darker areas on the shade plots and the dot-density plots. At least five of these anomalies have a size, shape, and magnetic response that led us to suspect large cultural features, such as house floors, large hearths, wattle and daub structures, or fortification walls. Many of the smaller, less-defined anomalies are also probably cultural features. After we completed the magnetometer survey, Beck Slide returned to ground truth anomalies 1, 2, and 3. Each of these three anomalies was tested using a split spoon auger, with individual soil cores taken every meter along grids set up over the respective anomaly.

This slide Slide shows Anomaly 1, a square anomaly with a solid, positive core and positive linear outliers. The anomaly measures 8 meters across, and in form and size Slide it has a striking resemblance to the footprints of late prehistoric Pisgah phases houses of the Appalachian Summit area; these outlying positives seem to be structure walls, and the interior core is possibly either a central hearth or roof and wall collapse. Auger testing of this anomaly, and each of the others, confirmed our suspicion that these large features were the signatures of burned structures. The soil cores Slide from these areas revealed thick lenses of burned debris below the plow zone-black charcoal, red fired clay, and burned wood; one core may have gone through a burned timber. This slide Slide shows the distribution of positive soil cores from the area associated with Anomaly 1, and like the magnetometer test results, indicates the presence of a burned rectangular building.

Anomaly 3 Slide appeared as a smaller, 4 meter wide, square anomaly, and like Anomaly 1 had strong linear and corner outliers surrounding a postive interior core. Again, this combination of features suggests the four walls of a building around a central hearth. Auger testing Slide of this anomaly, illustrated here, revealed a probable burned structure, rectangular in shape, and somewhat smaller than Anomaly 1.

Anomaly 2 Slide has a different magnetic signature than anomalies 1 and 3. Unlike these previous anomalies, Anomaly 2 lacks a postive interior core, and appears as a circular feature rather than having a rectangular shape; also, this anomaly is less well-defined, exhibiting a more diffuse magnetic signature. Auger testing of this anomaly revealed a probable burned structure Slide--these soil cores also revealed layers of burned wood and heavily fired red and orange soils--but the pattern, seen in this slide Slide, is much less regular than that of anomalies 1 and 3, again following closely to the magnetic signature. Although it is impossible to know for certain without more fieldwork, it is possible that Anomaly 2 represents a different kind of structure than that represented by Anomaly 1 and Anomaly 3, lacking a central hearth, having a circular shape, and having a more complex pattern of walls and/or features around the main walls of the building.

Anomaly 4 (SLIDE) is a squarish anomaly, not auger-tested, about four to five meters across, with a very strong, almost metallic positive response. The twin peaks, however, look like responses created by heavily baked clay features with highly magnetized peripheries and less magnetic centers. This anomal may have been a feature within the mound, or possibly a part of a feature that preceded the mound's construction.

Anomaly 5 (SLIDE) is a more amorphous, positive anomaly of respectable strength, indicating an area of about eight meters by four meters, but not well-defined. It may be part of a house floor or another large feature possibly disturbed by the bulldozing of the mound. (NO SLIDE)

In conclusion, the magnetometer survey and subsequent auger testing at the Berry site has been highly effective at detecting certain types of cultural features, especially features associated with episodes of burning. Other, non-magnetic features are certain to be found in the surveyed area. It is intriguing that so many structure-sized features with evidence of destruction by fire were identified in such a relatively small area, in particular the area where most of the Spanish artifacts have been recovered. It is tempting to speculate that the controlled surface collections, magnetometer survey, and auger testing have picked up evidence of Fort San Juan, with evidence of a conflagration that may have occurred in the spring of 1568 when the Spanish outposts were overrun by local Native Americans. As the following paper by Dave Moore and Chris Rodning will show, excavations this spring at the Berry site have confirmed these test results, and have revealed additional European artifacts in this area of the site. Excavations are planned for next year, as well as additional geophysical surveys. Future work at the Berry site promises to offer us a fascinating glimpse of Native American and Spanish interactions at the point of Contact between the Old and New Worlds.

  In the Swannanoa Valley of the Blue Ridge Mountains