In Search of Burned Buildings at the Berry Site:

Buried, Burned, Burke Buildings at Berry

by David G. Moore & Christopher B. Rodning

The Berry site represents a Mississippian town that dates to the 1400s and 1500s in the upper Catawba River valley of North Carolina's western Piedmont region. The results of excavations conducted in 1986 and subsequent analysis have led researchers to suggest that the site may be the location of the native town of Xualla or Joara, visited by Hernando de Soto in 1540 and by Juan Pardo during his expeditions from Santa Elena in 1566-1568. Joara was also the location of one of the forts, Fort San Juan, built by Pardo. As has just been described by Rob Beck and Tom Hargrove, a 1997 proton-magnetometer survey of a portion of the Berry site identified several anomalies that were believed to represent burned structures. Significantly, major anomalies were found in a portion of the site from which sixteenth-century Spanish artifacts had been recovered. Summer, 2001, excavations by Warren Wilson College have demonstrated that these anomalies are indeed burned buildings. In this paper we present the results of these excavations and consider the potential evidence for the presence of Fort San Juan. We begin with a brief introduction of the Burke phase and the Berry site.

The Burke phase is a Middle Lamar phase dating from AD 1400-1600. Burke phase archaeological sites are situated in the foothills of the Blue Ridge Mountains, along the upper reaches of the Catawba and Yadkin Rivers in western North Carolina (Moore in press). The earliest archaeological investigations in this region were conducted by the Bureau of American Ethnology's mound exploration program during the late nineteenth century, as a result of which Cyrus Thomas (1887, 1891, 1894) and William H. Holmes (1903) reported more than 20 earthen mounds along the upper Catawba and Yadkin rivers. Until recently, Thomas's (1887, 1894: 333-344) famous report of the Bureau's work at several Yadkin Valley mounds remained the only description of excavated sites in this region for more than 100 years.

Aside from the unusual distribution of earthen mounds, this area was best known for the presence of soapstone-tempered, Burke-series ceramics. William Henry Holmes (1903: 143-144) first described Burke pottery (though not by the term "Burke") in his descriptions of several vessels from mounds in the upper Yadkin Valley. These vessels exhibited what Holmes referred to as "southern traits," that today are considered characteristic of Lamar pottery. Holmes' illustrated vessels are excellent examples of Burke-type vessels (Moore 1999: 178-179).

Following Mark Williams and Gary Shapiro in their description of the Lamar ceramic tradition (1990:5), we define Burke pottery as a regional variant of Lamar pottery, characterized by complicated stamped jars, incised and burnished cazuela bowls, and the near exclusive use of crushed soapstone temper. This set of ceramic traits is largely restricted to the upper Catawba and Yadkin valleys and we have defined the Burke phase to explain this regionally distinctive distribution of pottery and mounds (Moore in press).

Based on survey data, we estimate that at least fifty Burke phase sites are located in the study area. The best-known Burke phase sites are the Berry site (31BK22), the T. F. Nelson Mound (31CW1) and Triangle, the Davenport Jones Mound, the Lenoir Indian Burial Pit, and the Broyhill-Dillard Mound (31CW8).

The Berry Site

The Berry site is located on Upper Creek, a tributary of the Catawba River. Cyrus Thomas first described the site in 1891 as a "Mound on the west Bank of Upper Creek 8 miles north of Morganton (about 15 feet high and unexplored)" (Thomas 1891: 151). Both the earthen mound and its 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. 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 surrounding field. The entire site covers at least five hectares (13 acres), based on the extent of surface artifacts as determined through controlled surface collections by Rob Beck (1997a).

Initial excavations at the Berry site covered an area of approximately 220 square meters and identified an abundance of postholes and features adjacent to the mound. Trench A was placed across the mound and revealed an undisturbed, basket-loaded mound fill beneath the plowed soil. Trench B, located adjacent to the mound, yielded a deeper and more complex stratigraphy related to the formation and destruction of the mound.

These limited excavations provided abundant artifactual information but, unfortunately, provided little information on site structure. Mound construction itself is believed to have occurred during the Burke phase.

The Berry site ceramic assemblage consists almost exclusively of Burke series pottery. The high frequencies of curvilinear complicated stamping, carinated bowls with Lamar incising, and medium-to-wide thickened and punctated jar rims is most similar to Tugalo phase (AD 1450-1600) pottery in the upper Savannah River (Anderson, et al. 1986; Hally 1990, 1994).

Although two radiocarbon dates from Berry provide a solid fifteenth-century context for the site, specific Burke ceramic attributes suggest that the major component dates into the sixteenth century. A mid- to late sixteenth century occupation is also supported by the presence of a small and unusual assemblage of sixteenth-century Spanish artifacts. Rob Beck briefly reviewed these artifacts in the previous paper. We believe this assemblage results from direct Spanish - Native contact at the Berry site and may reflect the presence of Pardo's Fort San Juan.

2001 Excavations at Berry

Since the initial excavations in 1986 and the proto-magnetometer survey in 1997, a portion of the Berry site has been leased for a tree nursery. However, the property owner has reserved 12 acres for use by local gardeners and to allow continued archaeological investigations. The summer 2001 excavations were conducted under the auspices of the Warren Wilson College Archaeology Field School. The four-week course was directed by the authors and included students from Warren Wilson College as well as students from several other colleges and universities and members of the general public. The goal of this summer's investigation was to evaluate the status of the magnetometer anomalies identified by Tom Hargrove and Rob Beck in the 1997 survey. In the fall of 1999, we had discovered -- to our alarm -- that the Berry site recently had been deeply plowed and there were indications that the buried anomalies had been completely plowed away. Therefore, the field school was designed to investigate as many of the larger anomalies as possible. Our access to the site was restricted to the area of the magnetometer survey, one-acre north of the mound, the same area from which most of the Spanish artifacts have been recovered.

Our field methodology consisted of placing 2 by 2 meter excavation units adjacent to selected anomalies and then expanding these units as necessary. This slide shows the location of the magnetometer survey. This slide shows the ultimate layout of the excavation as we removed the plow zone from units totaling 117 square meters. We began with Anomalies 1, 2, and 3 as well as several other large anomalies that indicated the possible presence of metal. As Rob Beck just described, we believed that Anomalies 1, 2, and 3 were burned buildings and we hoped to locate and define the edges of each of them.

Our initial results were not promising. It immediately became clear that the recent plowing had gone far below the 30 cm depth of the previous plow zone identified in the 1986 excavations and in the 1997 coring regimen. In fact, the new plow zone ranged from 50 to nearly 60 cm in depth. As we reached the bottom of the plow zone in our first few units we found almost no sign of features or postholes and it appeared that most of the site had been plowed away. Fortunately, as we expanded the units we began to find a small number of postholes and a few features. Most importantly, our units placed next to Anomalies 1, 2, and 3 ultimately demonstrated that burned structural debris is still present in the areas of these major anomalies.

This is most clearly demonstrated with Anomaly 1. Here, we found the edge of one building that was located within 10 cm of the location indicated by the proton-magnetometer. This slide clearly shows the remains of a burned structure. Vertical burned posts are in evidence along the exterior walls of the building and fragments of many burned timbers lay horizontally in the fill of the building interior. A band of dark brown soil lines the structure on the exterior side of the upright posts. This suggests that the original construction may have involved an earthen embanked wall or a semi-subterranean floor. The structure appears to be square and is approximately 8 meters in diameter. We did not excavate any portion of the feature but based on the initial coring regimen we believe that approximately 5-10 centimeters of the burned structure remains intact.

The excavation units associated with Anomalies 2 and 3 also revealed large features with burned soils and though we are unable to determine the nature of either of these anomalies yet, we believe they also represent burned structures. These slides illustrate the excavation units associated with Anomaly 2. and Anomaly 3. We plan to expose larger portions of these features next summer.

With one exception, we were unsuccessful at locating other specific anomalies. It seems likely that the feature exposed here is associated with one or both of the recorded anomalies south of Anomaly 2. As you can see we were not able to verify the presence of several other anomalies and it is possible that they were plowed away. It is also interesting to note that our excavations identified two other large features and two burials, none of which registered on the proton-magnetometer survey. Feature 20 was a circular pit approximately one meter in diameter that contained a small quantity of pottery and a large mass of charred wood. Feature 23 is a massive feature measuring nearly 2 by 3 meters. It possibly represents two or more contiguous pit features. Feature 23 was not entirely exposed until the last few days of the field season and will be excavated during our 2002 field school. In general, it is difficult to accurately assess the effectiveness of the proton magnetometer survey since it is highly likely that the deep plowing destroyed features that originally were reflected in the survey. Nonetheless, though we can see that the survey did not identify all features, it was extremely accurate at locating large burned features. We hope to continue to use magnetometers and other remote-sensing tools at the Berry site in the future. If other burned structures are present, these techniques should be able to identify them.

Artifact Recovery

Most of our 2001 excavation efforts involved plow zone removal. Only two features were excavated. All plowzone soils were screened through ¼ inch hardware cloth and feature fill was water screened through 1/16 in hardware cloth. Artifacts consisted of primarily Burke pottery as well as far fewer numbers of miscellaneous lithic flakes, tools, and projectile points. Of course, we hoped to recover additional Spanish artifacts and though we have not yet identified any additional clearly diagnostic 16th century Spanish artifacts, a small number of historic materials were recovered. These include two glass beads (both split longitudinally). One bead is a small blue wire-wound bead the other is a twisted cane bead with a square profile that is similar to some of the Nueva Cadiz bead forms associated with sixteenth century sites, a metal button, 4 miscellaneous pieces of brass at least two of which appear to have cut edges, a brass tinkler, and a quartered lead shot. Interestingly, the quartered lead shot is identical to quartered lead shot recovered from Santa Elena. Furthermore, we must note that Fort San Juan was left well stocked by Pardo. Included among the supplies left at the fort were 100 pounds of lead shot, 135 pounds of lead, and 34 pounds of nails. According to the Bandera account, Fort San Juan was the only interior fort that pardo supplied with nails (Hudson 1990:147-152).

Despite the absence of clearly diagnostic 16th century Spanish materials, these artifacts, combined with the materials recovered in 1986 and 1997, continue to constitute an unusual assemblage of historic artifacts. We still have three times as many sixteenth-century olive jar fragments as we do glass beads and the assemblage as a whole still conforms with John Worth's characterization of it as consisting principally of items that are primarily not trade materials.

In fact, European artifacts are rarely found on Burke phase sites. Moore has discussed the iron knife and celts recovered during the Bureau of American Ethnology's excavations of the Nelson mound and Triangle sites in the upper Yadkin valley. These Burke phase sites are less than 15 miles away from the Berry site. We think it is also significant that among the more than 150 burials excavated from the Nelson Mound and Triangle and the nearby Davenport Jones Mound and the Lenoir Burial Pit, no other European artifacts were recovered.

In sum, among all of the historic artifacts found at the Berry site, more items can be attributed with some certainty to the sixteenth and early seventeenth centuries than can possibly be attributed to the eighteenth century.

What About Fort San Juan?

If the Berry site is indeed Joara, it is possible that the three large burned structures are directly associated with Fort San Juan. The Fort was built on Pardo's first expedition, probably in February of 1567 and Pardo left a detachment of thirty men when he returned to Santa Elena. Pardo returned to Joara on September 24, 1567. After a march into the mountains of east Tennessee where he became alarmed at the Native resistance, Pardo retreated to Joara on November 6. Once again returning to Santa Elena, Pardo left a garrison of 30 men and departed from Joara on November 24, 1567. Pardo never returned to Fort San Juan. Based on the accounts of Pardo's expeditions and documentary references to the destruction of the forts in the interior, it is certain that Fort San Juan stood for at least nine months and perhaps for as long as 15 months.

We do not know the size or configuration of Fort San Juan. According to Charles Hudson (1990), Pardo had the Indians build houses to store provisions. It is uncertain as to whether these houses also constituted a part of the fort. No other construction details are known. However, it appears that Fort San Juan was the largest and best provisioned of the forts that Pardo had built in the interior.

Finally, it must be noted that word reached Santa Elena in May, 1568, that all of the interior forts had been attacked and overrun by the Indians. Only one Spanish survivor, Juan Martin de Badajoz, is known to have escaped from the interior. It seems likely that if local communities like Berry wished to expel the Spaniards, ones means at their disposal would be to burn them out. Certainly at the Berry site, the proximity of three large burned structures suggests such an event.

Though it will be difficult to demonstrate that the Berry site is Joara or that these burned structures are associated with Fort San Juan, we expect that further excavations will provide more complete structural, material, and temporal evidence by which we may evaluate this designation.

Future Research

To conclude, we have begun a long-term research project on the Burke phase that is focused on the Berry site. Next summer we will continue excavation at Berry and we will also begin test excavations at five additional Burke phase sites located within 3 kilometers of Berry. Our major research questions concern the nature of the community situated at the Berry site. What is its relationship to smaller, nearby Burke sites? What is the source of the Spanish artifacts, and if they, in fact, result from occupation by the Spanish army, does the nature of this community change after this contact? We know there is an historic reference to Joara as late as 1605 but there is also archaeological evidence that suggests that most of the Catawba valley is depopulated by the end of the seventeenth century. We hope to gain a better understanding of these complex issues over the next few years.

  In the Swannanoa Valley of the Blue Ridge Mountains